You are on page 1of 53

Algebras, Rings and Modules, Volume

2: Non-commutative Algebras and


Rings 1st Edition Michiel Hazewinkel
Visit to download the full and correct content document:
https://textbookfull.com/product/algebras-rings-and-modules-volume-2-non-commutati
ve-algebras-and-rings-1st-edition-michiel-hazewinkel/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Foundations of Commutative Rings and Their Modules 1st


Edition Fanggui Wang

https://textbookfull.com/product/foundations-of-commutative-
rings-and-their-modules-1st-edition-fanggui-wang/

Biota Grow 2C gather 2C cook Loucas

https://textbookfull.com/product/biota-grow-2c-gather-2c-cook-
loucas/

Non Associative Normed Algebras Volume 2 Representation


Theory and the Zel manov Approach 1st Edition Miguel
Cabrera García

https://textbookfull.com/product/non-associative-normed-algebras-
volume-2-representation-theory-and-the-zel-manov-approach-1st-
edition-miguel-cabrera-garcia/

Chapter 1 Burnside rings Serge Bouc

https://textbookfull.com/product/chapter-1-burnside-rings-serge-
bouc/
Algebra 1 Groups Rings Fields and Arithmetic 1st
Edition Ramji Lal

https://textbookfull.com/product/algebra-1-groups-rings-fields-
and-arithmetic-1st-edition-ramji-lal/

Universal Algebras Joseph Muscat

https://textbookfull.com/product/universal-algebras-joseph-
muscat/

Affine Vertex and W algebras Dražen Adamovi■

https://textbookfull.com/product/affine-vertex-and-w-algebras-
drazen-adamovic/

Simple Relation Algebras 1st Edition Steven Givant

https://textbookfull.com/product/simple-relation-algebras-1st-
edition-steven-givant/

Algebras and Representations MATH 3193 2016 Alison


Parker

https://textbookfull.com/product/algebras-and-representations-
math-3193-2016-alison-parker/
Algebras, Rings and Modules
Non-commutative Algebras
and Rings

Volume 2
Algebras, Rings and Modules
Non-commutative Algebras
and Rings

Volume 2

Michiel Hazewinkel
Dept. of Pure and Applied Mathematics
Centrum Wiskunde & Informatica
Amsterdam, Netherlands
and
Nadiya Gubareni
Institute of Mathematics
¸
Czestochowa University of Technology
¸
Czestochowa, Poland

R
$6&,(1&(38%/,6+(56%22.
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2017 by Taylor & Francis Group, LLC


CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Printed on acid-free paper


Version Date: 20161017

International Standard Book Number-13: 978-1-138-03582-9 (Hardback)

This book contains information obtained from authentic and highly regarded sources. While all reasonable
efforts have been made to publish reliable data and information, neither the author[s] nor the publisher
can accept any legal responsibility or liability for any errors or omissions that may be made. The publish-
ers wish to make clear that any views or opinions expressed in this book by individual editors, authors
or contributors are personal to them and do not necessarily reflect the views/opinions of the publishers.
The information or guidance contained in this book is intended for use by medical, scientific or health-
care professionals and is provided strictly as a supplement to the medical or other professional’s own
judgement, their knowledge of the patient’s medical history, relevant manufacturer’s instructions and the
appropriate best practice guidelines. Because of the rapid advances in medical science, any information
or advice on dosages, procedures or diagnoses should be independently verified. The reader is strongly
urged to consult the relevant national drug formulary and the drug companies’ and device or material
manufacturers’ printed instructions, and their websites, before administering or utilizing any of the drugs,
devices or materials mentioned in this book. This book does not indicate whether a particular treatment
is appropriate or suitable for a particular individual. Ultimately it is the sole responsibility of the medi-
cal professional to make his or her own professional judgements, so as to advise and treat patients ap-
propriately. The authors and publishers have also attempted to trace the copyright holders of all material
reproduced in this publication and apologize to copyright holders if permission to publish in this form has
not been obtained. If any copyright material has not been acknowledged please write and let us know so
we may rectify in any future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, trans-
mitted, 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.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
Preface

This the second part of a two-part treatise on (selected parts of) non-commutative
algebra and ring theory. The contents of the first part [103] are given below.
Representation theory is a fundamental tool for studying groups, algebras and
rings (and many other things). In this volume we consider representation theory for
finite posets, finite dimensional algebras and semiperfect rings. The description of
modules over some classes of semiperfect rings is reduced to mixed matrix problems
over discrete valuation rings and division rings, which are considered in this book.
One of the main goals in ring theory is to reduce in a certain sense the description of
large classes of rings to simpler classes by the use of some ring theoretic constructions.
The best classical example is the Wedderburn-Artin theorem, however, there are a
number of other general results that can be mentioned.
Section 1.1 represents the definition and main properties of the basic main
construction of rings which are incidence rings of posets over associative rings.
In section 1.2 we consider some properties of a special class of incidence rings of
the form T(S) = I(S, D), where S is a finite partially ordered set and D is a division
ring. This class of rings properly contains the class of hereditary serial rings and all
Artinian rings with quivers that are trees.
A special class of right hereditary rings A(S, O) of a finite poset S over a family of
discrete valuation rings {Oi }i ∈I with a common skew field of fractions is introduced
and studied in section 1.3.
In section 1.4 we introduce and study a special class of incidence rings modulo
the radical of the form I(S, Λ, M) related to a finite poset S over a local Noetherian
domain Λ with Jacobson radical M.
Serial and semidistributive rings of the form I(S, Λ, M), where Λ is a discrete
valuation ring with Jacobson radical M are considered in section 1.5.
The main aim of chapter 2 is to study the properties and structure of different
classes of rings whose lattices of submodules are distributive (or semidistributive).
Such rings are called distributive (or semidistributive) and they can be considered as
a non-commutative generalization of Prüfer domains. The class of distributive rings
is very wide and includes, for example the ring of integers and the ring of polynomials
K[x] over a field K, rings of integral algebraic numbers, and commutative principal
ideal rings. More generally, all commutative Dedekind rings and Prüfer domains are
examples of distributive rings. The first papers which are devoted to distributive rings
appeared in the fifties-sixties of the 20-th century in [26], [27], [29], [155], [156].
vi Algebras, Rings and Modules

The systematic study of these rings began with papers of H. Achkar [2], V. Camillo
[39], H.H. Brungs [33] and W. Stephenson [203].
In section 2.1 we consider the main properties of distributive modules and rings.
Section 2.2 is devoted to study of semidistributive modules and rings. The structure
of Noetherian distributive and semidistributive rings is discussed in section 2.3.
The properties of right hereditary SPSD-rings are studied in section 2.4. It is
shown that the structure of all such rings is closely connected with right hereditary
rings of the form A(S, O) as described in section 1.3.
Section 2.5 is devoted to the study of the main properties and structure of
semihereditary SPSD-rings.
Some of the more fundamental notions and results of the theory of homological
algebra were studied in different sections of [100], [101]. Chapter 3 gives some
additional facts from homological algebra.
The notions of direct limits and inverse limits for a set of modules were considered
in [103, section 1.4] (see also [100, section 4.7]). In the particular case, when
this set has only two modules these constructions have their own names, pullback
and pushout. They are very important and useful in studying rings and modules.
Therefore for the convenience of the reader these module constructions and their
main properties are considered in more detail in section 3.1.
In the theory of homological algebra and its applications there is very important
the statement which is often known as the “snake lemma” (it is also called “zig-zag
lemma”, or “serpent lemma”). It is valid in every Abelian category and it is an
important tool in the construction of long exact sequences. This lemma consists
of two claims: 1) the construction of an exact sequence, which is often called
“kernel-cokernel sequence”, for any commutative diagram of a special type; and 2)
the construction of long exact sequences of homology groups for any given short
exact sequence of complexes. The second part of this lemma was proved in [100,
theorem 6.1.1]. The proof of the first part of this lemma is given in section 3.2.
The functors Extn as right derived functors for the contravariant left exact functor
Hom using projective resolutions were introduced in [100, section 6.4]. This functor
is closely related to module extensions.
Let A be a ring, and let X, Y ∈ Modr A. In this chapter the interpretation of the
group Ext1A(Y, X) is given in terms of short exact sequences. Section 3.3 is devoted
to studying extensions of modules in terms of short exact sequences.
Following R. Baer the addition of extensions of modules, which makes the set of
equivalence classes of all extensions an Abelian group, is introduced in section 3.4.
Some main properties of the group Ext1A(Y, X) are considered in section 3.5. The
results of section 3.6 show that there is an isomorphism between equivalence classes
of the group of extensions of X by Y and elements of Ext1A(Y, X) as considered in
[100].
In chapter 4 the main results about modules over semiperfect rings are given.
Some basic properties of semiperfect rings and modules over them were considered
in [100, chapter 10]. There are a number of equivalent definitions of semiperfect
rings. One of them is given by H. Bass in terms of projective covers. He proved
Preface vii

that a ring is semiperfect if and only if any finitely generated module has a projective
cover (see [103, theorem 1.8.2]).
In [100, section 10.4] the structure of finitely generated projective modules
over semiperfect rings was discussed. It was proved that any such right module
can be uniquely decomposed into a direct sum of principal right modules. The
generalization of this statement, the important theorem [103, theorem 7.2.7], states
that any projective module over a semiperfect ring is a direct sum of principal modules.
In section 4.1 this result is used to study the structure of finitely generated modules
over semiperfect rings. The main result of this section, which was proved by R.B.
Warfield, Jr. concerns the decompositions of any finitely generated module over a
semiperfect ring. This theorem gives a possibility to introduce the notion of stably
isomorphic modules.
In section 4.2 we prove that all modules over a semiperfect ring can be divided
into the equivalence classes of stably isomorphic modules. Moreover, each stable
isomorphism class of finitely generated modules over a semiperfect ring contains a
unique (up to isomorphism) minimal element. Most of the results of this section were
obtained by R.B. Warfield, Jr. (see [220], [222]).
In [101, section 4.10] we considered the duality in Noetherian rings, which is
given by the covariant functor ∗ = Hom A(−, A). For an arbitrary ring A this functor
induces a duality between the full subcategories of finitely generated projective right
A-modules and left A-modules. In section 4.3 the main properties of this functor and
torsionless modules are studied for the case of modules over semiperfect rings.
Section 4.4 presents an introduction to the duality theory of Auslander and
Bridger [13], and yields a connection between finitely presented right modules and
finitely presented left modules over semiperfect rings. Some main properties of the
Auslander-Bridger transpose, which is closely connected with almost split sequences,
are discussed. These sequences were first introduced and studied by M. Auslander
and I. Reiten in [15] and [20], and they play an important role in the representation
theory of rings and finite dimensional algebras. Section 4.5 is devoted to the study
of some main properties of these sequences. In section 4.6 we study almost split
sequences over semiperfect rings and prove the existence of these sequences for
strongly indecomposable modules. Section 4.7 presents an introduction to the theory
of linkage of modules over semiperfect Noetherian rings using two types of functors:
syzygy and transpose.
Chapter 5 is devoted to finite partially ordered sets (posets) and their
representations, which play an important role in representation theory. They were first
introduced and studied by L.A. Nazarova and A.V. Roiter [163] in 1972 in connection
with problems of representations of finite dimensional algebras. M.M. Kleiner
characterized posets of finite type [129] and described their pairwise non-isomorphic
indecomposable representations [130]. He proved a theorem which gives a criterium
for posets to be of finite representation type.
Recall that a finite poset P is called primitive if it is a cardinal sum of linearly
ordered sets L1, . . . , Lm . It is then denoted by P = L1 ⊔ · · · ⊔ Lm . This chapter
viii Algebras, Rings and Modules

gives the proof of the criterium for primitive posets to be of finite representation type
following [30].
To prove this criterium there are used only the trichotomy lemma which was proved
by P. Gabriel and A.V. Roiter in [82], the Kleiner lemma about the representations
of a pair of finite posets proved by M. Kleiner in [129] and the main construction,
considered in section 5.5. Note that this construction (in some form) was introduced
by L.A. Nazarova and A.V. Roiter in [163].
An important problem in the theory of representations of finite dimensional
algebras (or f.d. algebras, in short) is to obtain the full list of different kinds of
algebras which are of finite representation type (or finite type, or f.r.t., in short). The
first classes of associative f.d. algebras of f.r.t which have been described were the
classes of algebras with zero square radical and hereditary algebras over algebraically
closed fields.
There are different approaches to study the representations of f.d. algebras. One
of them is the approach of P. Gabriel [79], which reduces the study of representations
of algebras to the study of representations of quivers. Another approach was first
considered by L.A. Nazarova and A.V. Roiter [163]. This approach is to solve “matrix
problems”, that is, the reducing of some classes of matrices by means of admissible
transformations to their simplest form. A third approach is due to M. Auslander and
it is connected with the technique of almost split sequences.
Chapter 6 can be considered as an introduction to the theory of representations of
quivers and finite dimensional algebras. This chapter gives some main notions and
some fundamental results of these representations, most of which are given without
proof. In section 6.1 we consider the notions of finite quivers and their representations
and give the main results of this theory. Section 6.2 is devoted to species and their
representations. In section 6.3 we consider some main notions and results of the
representation theory of finite dimensional algebras.
As it turns out the category of representations of finite dimensional algebras is
equivalent to the category of representations of special classes of quivers, which are
called bound quivers. That is why the quivers play a central role in the theory of
finite dimensional associative algebras and their modules.
For right Artinian rings one can also introduce the notion of a ring of finite
representation type. As has been shown by D. Eisenbud and P. Griffith [71] this
notion is left-right symmetric. They proved this fact using the duality theory of
Auslander and Bridger. This result is proved in section 7.1.
For finite dimensional algebras along with the notion of finite representation type
there is also considered the notion of bounded representation type. Recall that a
finite dimensional algebra A is called of bounded representation type if there is a
bound on the length of the indecomposable finite dimensional A-modules. The first
Brauer-Thrall conjecture says that these notions are the same in the case of a finite
dimensional algebra A (as was proved by A.V. Roiter [101, theorem 3.5.1]) and in the
case of Artinian algebras (as was proved by M. Auslander [15]).
A ring A, not necessarily Artinian, is said to be of finite representation type, if it
has a finite number of non-isomorphic indecomposable finitely presented A-modules.
Preface ix

For Artinian rings this definition coincides with the earlier one, because in this
case each finitely generated A-module is finitely presented as well. The main
results concerning Artinian hereditary rings of finite representation type are given in
section 7.4.
For Artinian rings along with the notion of finite representation type there is
considered the notion of bounded representation type. Recall that a right Artinian
ring A is said to be of bounded representation type if there is a bound on the
length of finitely generated indecomposable right A-modules. The first Brauer-Thrall
conjecture asserts that these notions are the same. M. Auslander proved that this
conjecture is true for right Artinian rings (see theorem 7.2.11).
Following R.B. Warfield, Jr. a ring has right bounded representation type if
there is an upper bound on the number of generators required for indecomposable
finitely presented right A-modules. In his paper [222] R.B. Warfield, Jr. puts the
following question:
Question 4. For what semiperfect rings is there an upper bound on the number of
generators required for the indecomposable finitely presented modules?
Chapter 8 describes special classes of semiperfect rings of bounded representation
type, which constitute some sort of answer to this question.
As shown by R.B. Warfield, Jr. there is a serious restriction on the structure
of rings of bounded representation type connected with modules of finite Goldie
dimension. In section 8.1 it is proved that a semiprimary ring of finite right bounded
representation type is right Artinian (see [222]).
O-species and tensor algebras, which are generalizations of k-species as introduced
by P. Gabriel, are considered in section 8.2. The connection between right hereditary
SPSD-rings and special kinds of (D, O)-species is considered in section 8.3.
In section 8.4 we discuss the reduction of representations of (D, O)-species to
mixed matrix problems over discrete valuation rings and their common skew field of
fractions. Some important mixed matrix problems are considered in section 8.5.
Sections 8.6 and 8.7 are devoted to the study of (D, O)-species of bounded
representation type. There is a theorem which gives the structure of these species
in the terms of diagrams which can be considered as generalizations of Dynkin
diagrams.
The right hereditary SPSD-rings of finite right bounded representation type are
described in section 8.8.
The book is written on a level accessible to advanced students who have some
experience with modern algebra. It will be useful for those new to the subject as well
for researchers and serves as a reference volume.
While writing this book the second author was in particular supported by FAPEST
of Brazil in 2010. The author would like to express cordial thanks to the Institute
of Mathematics and Statistics of the University of São Paulo, and especially prof.
M. Dokuchaev and prof. V.M. Futorny, for their warm hospitality during her visit in
2010. The author is also grateful to prof. V.V. Kirichenko for useful discussions and
fruitful remarks.
CONTENTS

Contents of Volume 1
Preface v

1. Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Basic Concepts Concerning rings and Modules . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Categories and Functors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3 Tensor Product of Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.4 Direct and Inverse Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5 Projective, Injective, and Flat Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.6 The Functor Tor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.7 The Functor Ext . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.8 Hereditary and Semihereditary Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.9 Semiperfect and Perfect Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.10 Serial and Semidistributive Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.11 Classical Rings of Fractions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
1.12 Quivers of Algebras and Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

2. Basic General Constructions of Groups and Rings . . . . . . . . . . . . . . . . . . . . . 37


2.1 Direct and Semidirect Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.2 Group Rings, Skew Group Rings and Twisted Group Rings . . . . . . . . . . . . 44
2.3 Polynomial and Skew Polynomial Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.4 Formal Power and Skew Formal Power Series Rings . . . . . . . . . . . . . . . . . . . 55
2.5 Laurent Polynomial Rings and Laurent Power Series Rings . . . . . . . . . . . . 62
2.6 Generalized Matrix Rings. Generalized Triangular Matrix Rings . . . . . . . 65
2.7 G-graded Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
2.8 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

3. Valuation Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.1 Valuation Domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.2 Discrete Valuation Domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
3.3 Invariant Valuation Rings of Division Rings . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.4 Examples of Non-commutative Non-discretely-valued Valuation Rings . . 93
3.5 Discrete Valuation Rings of Division Rings . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.6 Total Valuation Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
xii Algebras, Rings and Modules

3.7 Other Kinds of Valuation Rings with Zero Divisors . . . . . . . . . . . . . . . . . . 105


3.8 Approximation Theorem for Valuation Rings . . . . . . . . . . . . . . . . . . . . . . . . 113
3.9 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

4. Homological Dimensions of Rings and Modules . . . . . . . . . . . . . . . . . . . . . . . 119


4.1 Projective and Injective Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.2 Flat and Weak Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.3 Some Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
4.4 Homological Characterization of Some Classes of Rings . . . . . . . . . . . . . . 130
4.5 Torsion-less Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
4.6 Flat Modules and Coherent Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
4.7 Modules over Formal Triangular Matrix Rings . . . . . . . . . . . . . . . . . . . . . . . 152
4.8 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

5. Goldie and Krull Dimensions of Rings and Modules. . . . . . . . . . . . . . . . . . . 160


5.1 Uniform Modules and Uniform Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . 161
5.2 Injective Uniform Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
5.3 Nonsingular Modules and Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
5.4 Nonsingular Rings and Goldie Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
5.5 Reduced Rank and Artinian Classical Ring of Fractions . . . . . . . . . . . . . . . 191
5.6 Classical Krull Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
5.7 Krull Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
5.8 Relationship between Classical Krull Dimension and Krull Dimension . 218
5.9 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225

6. Rings with Finiteness Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227


6.1 Some Noetherian Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
6.2 Dedekind-finite Rings and Stably Finite Rings . . . . . . . . . . . . . . . . . . . . . . . 232
6.3 FDI-Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
6.4 Semiprime FDI-Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
6.5 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

7. Modules with Change Property and Change Rings . . . . . . . . . . . . . . . . . . . 260


7.1 The Exchange Property . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
7.2 The Azumaya Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
7.3 Cancelation Property . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
7.4 Exchange Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
7.5 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291

8. Hereditary and Semihereditary Rings. Piecewise Domains . . . . . . . . . . . . 294


8.1 Rickart Rings and Small’s Theorems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
8.2 Dimensions of Hereditary and Semihereditary Rings. . . . . . . . . . . . . . . . . . 301
8.3 Right Hereditary and Right Semihereditary Prime Rings . . . . . . . . . . . . . . 305
8.4 Piecewise Domains. Right Hereditary Perfect Rings . . . . . . . . . . . . . . . . . . 313
8.5 Primely Triangular Rings. Structure of Piecewise Domains . . . . . . . . . . . . 316
8.6 Right Hereditary Triangular Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
Contents xiii

8.7 Noetherian Hereditary Primely Triangular Rings . . . . . . . . . . . . . . . . . . . . . 326


8.8 Right Hereditary Species and Tensor Algebras . . . . . . . . . . . . . . . . . . . . . . . 328
8.9 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
9. Serial Nonsingular Rings. Jacobsons’s Conjecture . . . . . . . . . . . . . . . . . . . 334
9.1 Structure of Serial Right Noetherian Piecewise Domains . . . . . . . . . . . . . 335
9.2 Structure of Serial Nonsingular Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
9.3 Serial Rings with Noetherian Diagonals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
9.4 The Krull Intersection Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
9.5 Jacobson’s Conjecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
9.6 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369

Contents of Volume 2
Preface v
1. Rings Related to Finite Posets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Incidence Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Incidence Rings I(S, D) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3 Right Hereditary Rings A(S, O) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.4 Incidence Rings Modulo the Radical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.5 Serial and Semidistributive Rings I(S, Λ, M) . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.6 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2. Distributive and Semidistributive Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.1 Distributive Modules and Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2 Semiprime Semidistributive Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3 Semiperfect Semidistributive Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.4 Right Hereditary SPSD-Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.5 Semihereditary SPSD-Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.6 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3. The Group of Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.1 Module Constructions Pushout and Pullback . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.2 The Snake Lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.3 Extensions of Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.4 Baer Sum of Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.5 Properties of Ext1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.6 Ext1 and Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
3.7 Additive and Abelian Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.8 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
4. Modules Over Semiperfect Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.1 Finitely Generated Modules over Semiperfect Rings . . . . . . . . . . . . . . . . . . . 99
xiv Algebras, Rings and Modules

4.2 Stable Equivalence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101


4.3 Auslander-Bridger Duality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
4.4 Almost Split Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
4.5 Cogenerators and Some Identities for Finitely Presented Modules . . . . . . 119
4.6 Pure Submodules and Pure Exact Sequences . . . . . . . . . . . . . . . . . . . . . . . . 125
4.7 Almost Split Sequences over Semiperfect Rings . . . . . . . . . . . . . . . . . . . . . . 131
4.8 Linkage and Duality of Modules over Semiperfect Rings . . . . . . . . . . . . . . 140
4.9 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
5. Representations of Primitive Posets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
5.1 Representations of Finite Posets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
5.2 Main Canonical Forms of Matrix Problems . . . . . . . . . . . . . . . . . . . . . . . . . . 165
5.3 Trichotomy Lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
5.4 The Kleiner Lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
5.5 The Main Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
5.6 Primitive Posets of Infinite Representation Type . . . . . . . . . . . . . . . . . . . . . . 180
5.7 Primitive Posets of Finite Representation Type . . . . . . . . . . . . . . . . . . . . . . . 183
5.8 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
6. Representations of Quivers, Species and Finite Dimensional Algebras . . 191
6.1 Finite Quivers and Their Representations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
6.2 Species, Valued Quivers and Valued Graphs . . . . . . . . . . . . . . . . . . . . . . . . . 209
6.3 Finite Dimensional Algebras and Their Representations . . . . . . . . . . . . . . 223
6.4 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
7. Artinian Rings of Finite Representation Type . . . . . . . . . . . . . . . . . . . . . . . . . 234
7.1 Eisenbud-Griffith’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
7.2 Auslander’s Theorem for Right Artinian Rings . . . . . . . . . . . . . . . . . . . . . . . 238
7.3 Artinian Semidistributive Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
7.4 Artinian Hereditary Semidistributive Rings of Finite . . . . . . . . . . . . . . . . . . . . .
Representation Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
7.5 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
8. O-Species and SPSD-Rings of Bounded Representation Type . . . . . . . . . . 255
8.1 Semiperfect Rings of Bounded Representation type . . . . . . . . . . . . . . . . . . . 256
8.2 O-Species and Tensor Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
8.3 (D, O)-Species and Right Hereditary SPSD-Rings . . . . . . . . . . . . . . . . . . . 263
8.4 Representations of (D, O)-Species and Mixed Matrix Problems . . . . . . . 267
8.5 Some Mixed Matrix Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
8.6 (D, O)-Species of Unbounded Representation Type . . . . . . . . . . . . . . . . . . 278
8.7 (D, O)-Species of Bounded Representation Type . . . . . . . . . . . . . . . . . . . . . 287
8.8 Right Hereditary SPSD-Rings of Bounded Representation Type . . . . . . . 295
8.9 Notes and References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
Glossary of Ring Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
CHAPTER 1

Rings Related to Finite Posets

One of the main goals in the ring theory is to reduce in a certain sense the description
of large classes of rings to simpler classes using some ring theoretic constructions.
The most classical example is the Wedderburn-Artin theorem, however one can
mention a number of other general results.
Section 1.1 represents the definition and main properties of a basic construction
of rings. These are called incidence rings of posets over associative rings.
In section 1.2 we consider some properties of a special class of incidence rings of
the form T(S) = I(S, D), where S is a finite partially ordered set and D is a division
ring. This class of rings properly contains the class of hereditary serial rings and all
Artinian rings with quivers that are trees.
A special class of right hereditary rings A(S, O) of a finite poset S over a family of
discrete valuation rings {Oi }i ∈I with a common skew field of fractions is introduced
and studied in section 1.3.
In section 1.4 we introduce and study a special class of incidence rings. These
are called incidence rings modulo the radical. They are denoted by I(S, Λ, M) and
involve a finite poset S over a local Noetherian domain Λ with Jacobson radical M.
Serial and semidistributive rings of the form I(S, Λ, M), where Λ is a discrete
valuation ring with Jacobson radical M, are considered in section 1.5.

1.1 Incidence Rings

This section covers the definition and some properties of incidence rings of partially
ordered sets over associative rings.
Let S be a set with a binary relation denoted by ≼. The relation ≼ is called a
preorder if the following properties satisfy:

1. a ≼ a for any a ∈ S (reflexivity).


2. (a ≼ b) ∧ (b ≼ c) implies a ≼ c for any a, b, c ∈ S (transitivity).
A set which is equipped with a preorder is called a preordered set.
Note that if a preorder ≼ is antisymmetric, that is,
2 Algebras, Rings and Modules

3. (a ≼ b) ∧ (b ≼ a) implies a = b for each a, b ∈ S (antisymmetry)


then ≼ is a partial order, and a set which is equipped with a partial order is
called a partially ordered set (or poset, for short).
If a preorder ≼ is symmetric, that is,
4. a ≼ b implies b ≼ a for each a, b ∈ S (symmetry)
then ≼ is called an equivalence relation.

Let S be a poset with partial ordering relation ≼ (which is transitive, reflexive and
antisymmetric). Denote by x ≺ y the strict order, i.e. the relation "x ≼ y and x , y".
An element x ∈ S is called maximal if there is no element y ∈ S satisfying x ≺ y.
Dually, x is minimal if there is no element y ∈ S satisfying y ≺ x. An element x ∈ S
is said to be a least element if x ≼ y for all y ∈ S, and x ∈ S is said to be a greatest
element if y ≼ x for all y ∈ S.
A subset C of a poset S is called a chain if it is totally ordered, i.e. for any x, y ∈ C,
either x ≼ y or y ≼ x. Denote a chain of cardinality n by Cn , or (n). In this case
the number n is called the length of the chain Cn . A subset A of a poset S is called
an antichain if all its elements are pairwise incomparable in S, i.e. for any pair of
distinct elements x, y ∈ A, both x  y and y  x.

In order to visualize a poset S one uses the so called diagram of S. Let x and y
be distinct elements of S. It is said that y covers x if x ≺ y but there is no element z
such that x ≺ z ≺ y. Recall that the diagram H(S) of a poset S is the directed graph
whose vertex set is S and whose set of edges is given by the set of covering pairs
(x, y) of S, moreover, there is a (directed) edge from a vertex x up to a vertex y if and
only if y covers x.
For example, the diagram below

•O a3

• a1 • a2
represents the poset (S, ≼) with 3 elements {a1, a2, a3 } and the one relation a2 ≺ a3 .

Remark 1.1.1. The diagram H(S) of a poset S is often called its Hasse diagram.
Usually it is drawn in the plane in such a way that if y covers x then the point
representing y is drawn higher than the point representing x. In this case the Hasse
diagram is drawn without arrows. For example, the Hasse diagram below

• a3

• a1 • a2
represents the same poset (S, ≼) as above, i.e. S = {a1, a2, a3 } with the one relation
a2 ≺ a3 .
Rings Related to Finite Posets 3

Definition 1.1.2. An interval [x, y] of a preordered set S with preorder ≼ is defined


as the set [x, y] = {z ∈ S : x ≼ z ≼ y}. A preordered set S is locally finite if each
its interval is finite.

The idea of incidence algebras had its beginning in the works of R. Dedekind
and E.T. Bell. The incidence algebra of a locally finite poset over a field was first
introduced by G.-C. Rota in [185]. This notion can be extended to the case of a
non-commutative ring A.

Definition 1.1.3. Let S be a locally finite poset with a partial ordering relation ≼,
and A an associative (not necessarily commutative) ring with identity. The incidence
ring I(S, A) of S over A is the set of functions f : S × S → A such that f (x, y) = 0 if
x  y with operations:
( f + g)(x, y) = f (x, y) + g(x, y),

( f g)(x, y) = f (x, z)g(z, y)
x ≼z ≼y
(a f )(x, y) = a( f (x, y))
for all x, y, z ∈ S and a ∈ A.

An important set of elements of I(S, A) is formed by the characteristic functions


which are defined in the following way. If X ⊆ S then
{
1 if x = z ∈ X
δX (x, z) =
0 otherwise.

Let us restrict our attention (for the moment) to the case of finite posets. Let
S = {α1, . . . , αn } be a finite poset with a partial ordering relation ≼, and let Mn (A) be
the generalized matrix ring of all n × n-matrices with elements from an associative
ring A. In this case the incidence ring of S over A is the subring I(S, A) of Mn (A)
such that the (i, j)-entry of I(S, A) is equal to 0 if αi  α j in S. So, letting correspond
to any element αi ∈ S the matrix unit eii ∈ Mn (A), and to any pairs of elements
αi, α j ∈ S such that αi ≼ α j the matrix unit ei j ∈ Mn (A) one obtain a basis (over A)
of I(S, A).

Example 1.1.4. The poset R = {a, b, c, d : a ≺ b ≺ d; a ≺ c ≺ d} has a Hasse


diagram of the following form:

d
•8
  888
88

b •8 •c (1.1.5)
88 
88 
8 

a
4 Algebras, Rings and Modules

It is called the rhombus.


Then
A A A A
© ª
­0 A 0 A®
I(S, A) = ­ ®
­0 0 A A®
«0 0 0 A¬

Example 1.1.6. Let S be the poset with Hasse diagram of the following form:
2 •3 • 3
33 
33 
3 
• 1

• 4
then
A A A 0
© ª
­0 A 0 0®
I(S, A) = ­ ®
­0 0 A 0®
« A A A A¬
As these examples illustrate, transitivity of the ordering relation has much to do
with "I(S, A) is a subring of Mn (A)".
The following statement is obvious:

Lemma 1.1.7. Every interval of a finite poset S is a chain if and only if S contains
no subposets whose diagrams are rhombuses.

Denote by S the non-oriented graph obtained from the diagram of S by deleting


the orientation of all its arrows. Then it is easy to show

Lemma 1.1.8. The non-oriented graph S is a disconnected union of trees if and only
if S contains no subposets whose diagrams are rhombuses.

Note also the well-known fact:

Lemma 1.1.9. A finite poset S with partial ordering relation ≼ can be labeled as
S = {α1, . . . , αn } where αi ≼ α j implies i ≤ j.

Using this lemma it is easy to show that in the case that S has property just
mentioned the incidence ring I(S, A) is isomorphic to a subring of the ring of the
upper triangular matrix ring Tn (A). In particular, if a poset S is linearly ordered then
I(S, A)  Tn (A).

Proposition 1.1.10. Let S = {α1, . . . , αn } be a finite poset with partial ordering


relation ≼, and A an associative ring with identity and Jacobson radical R. Then
Rings Related to Finite Posets 5

1. I(S, A) is an FDI-ring.
2. I(S, A) is a semiperfect ring if and only if A is semiperfect.
3. The Jacobson radical of I(S, A) is the each set of elements of I(S, A) for which
(i, i)-entry is in R.
4. I(S, A) is a right (left) Noetherian ring if and only if A is right (left) Noetherian.
5. I(S, A) is a right (left) Artinian ring if and only if A is right (left) Artinian.

Proof.

1. This is obvious because 1 = e11 + · · · + enn ∈ I(S, A) ⊂ Mn (A).


2. This follows from [103, Theorem 1.9.3].
3. Thanks to statement 1 B = I(S, A) is an FDI-ring. Therefore B has a two-sided
Peirce decomposition B = (Bi j ). Since for any i , j αi  α j or α j  αi in S,
Bi j B ji = 0. So the statement follows now from [103, Proposition 2.6.10].

Statements 4 and 5 follows from [103, Theorem 1.1.23]. 

Remark 1.1.11. The notion of an incidence ring has been generalized to the case of
preodered sets in [75], [204].
Let S be a finite preordered set with a preorder ≼, and A an associative (not
necessarily commutative) ring with identity. The incidence ring I(S, A) of S over A is
a ring with the additive structure of a free A-module with basis { fxy : x ≼ y; x, y ∈ S},
where multiplication is given by the linear extension of:
{
fxu if y = z
fxy · fzu = (1.1.12)
0 otherwise,

for all x, y, u, v ∈ V.

Example 1.1.13. Let ℜ be a finite preordered set {1, 2, . . . , n} with preordering


relation ≼. One can consider the corresponding reflexive and transitive Boolean
matrix B = [bi j ] defined by bi j = 1 if and only if i ≼ j, otherwise bi j = 0. To the
finite preordered set ℜ and an associative ring A with identity one can associate the
structural matrix ring M(B, A) associated with B and defined by the following way:

M(B, A) = {X = [xi j ] ∈ Mn (A) : bi j = 0 =⇒ xi j = 0}. (1.1.14)

In fact, M(B, A) is simply the incidence ring of the preordered set ℜ over the ring A.

In this section we have (so far) considered incidence rings of posets and preordered
sets. They are special cases of a general construction - the incidence rings of an
arbitrary relation which is not required to be a preorder. Such rings were introduced
by G. Abrams in [1] and were called generalized incidence rings.

Definition 1.1.15. Let A be an associative ring with unity. Let ≼ be a binary


relation on a locally finite set S. The generalized incidence ring I(S, ≼, A) of S with
coefficients in A is the free left A-module with basis { fxy : x ≼ y in S; x, y ∈ S},
6 Algebras, Rings and Modules

where multiplication is given by the linear extension of:


{
fxu if y = z and x ≼ u in S
fxy · fzu = (1.1.16)
0 otherwise,

for all x, y, u, v ∈ S.

Remark 1.1.17. When ≼ is a preorder on a set S the definition of a generalized


incidence ring given above coincides with the definition of an incidence ring as given
in Remark 1.1.11.

For a ring I(S, ≼, A) to be associative it is necessary that the following special


property be satisfied.

Definition 1.1.18. ([1]). A reflexive relation ≼ on a set S is said to be balanced if


for all x, y, u, v ∈ S with x ≼ y ≼ u ≼ v and x ≼ v ∈ S

x ≼ u ∈ S ⇐⇒ y ≼ v ∈ S.

Remark 1.1.19. Note that a relation which is both reflexive and transitive is always
balanced, but there are balanced relations which are not transitive. Some of these are
shown in Example 1.1.23 below.

Proposition 1.1.20. [1].

1. The multiplication in I(S, ≼, A) is associative if and only if the relation ≼ on


S is balanced.
2. If a relation ≼ is reflexive then I(S, ≼, A) has multiplicative identity.

Proof.

1. If x, y, z, w, u, v ∈ S then the expressions ( fxy fzw ) fuv and fxy ( fzw fuv ) are both
zero unless y = z, w = u and x ≼ v. In this case, the first product is equal to
fxv precisely when x ≼ w, while the second product is equal to fxv precisely
when z ≼ v, i.e. this is the case if and only if ≼ is a balanced relation.
2. From (1.1.16) it immediately follows that the elements fxx = ex for all x ∈ S
are idempotents in I(S, ≼, A) and
{ {
fyz if x = y fyz if x = z
ex fyz = , fyz ex =
0 otherwise 0 otherwise

for all x, y, z ∈ S. This implies that the element ex is a multiplicative
x ∈S
identity of I(S, ≼, A). 

The definition of a generalized incidence ring can be given in terms of graphs (see
[4]).
Rings Related to Finite Posets 7

Definition 1.1.21. Let S = (V, E) be a graph with set of vertices V = {1, 2, . . . , n} and
set of edges E ⊆ V × V. Then the associated generalized incidence ring I(S, A) of
S over an associative ring A is defined as the free left A-module with basis consisting
of all edges in E where multiplication is defined by the linear extension of
{
(x, v) if y = u and (x, v) ∈ E
(x, y)(u, v) =
0 otherwise,

for all x, y, u, v ∈ V.

Definition 1.1.22. A directed graph S contained all loops is said to be balanced if


for all different x, y, u, v ∈ V with (x, y), (y, u), (u, v), (x, v) ∈ E

(x, u) ∈ E ⇐⇒ (y, v) ∈ E.

Example 1.1.23. We omit loops in all diagrams represented in these examples. The
following directed graphs are balanced and nontransitive.

a) ; u O • FFF b) ; u • FFF
xxxx FF xxxx FF
x # x #
y • cF /•v y • cF
FF x; FF x;• v
F xxxx F xxxx
•x •x
/• /•
c) ; u • cFFF d) •
xxxx FF x y u
x
y • cF •v
FF xxx;
F xx
•x

Proposition 1.1.20 and these examples show, in particular, that there are associative
generalized incidence rings which are not incidence rings of preordered sets. Note
that Example 1.1.23 is in terms of directed graphs but not Hasse diagrams.

Remarks 1.1.24.

1. The examples 1.1.23(a,b) may suggest why the terminology ’balanced’ is


used.
2. In definitions 1.1.18, 1.1.22 we assume that a relation is reflexive (and so its
corresponding directed graph has all loops). But some authors (see e.g. [4])
define balanced relation (and balanced graphs) without this assumption.
3. There is also a different notion in graph theory that goes by the name ’balanced
directed graph’. The requirement is that the in and out degrees at each vertex
are the same. See [50]. Reversing the arrow (x, v) in the picture above turns
it into a balanced digraph in this sense.
8 Algebras, Rings and Modules

In [190] V.D. Shmatkov introduced another generalization of an incidence ring.

Definition 1.1.25. A set of pairwise orthogonal idempotents E of a ring A is called


maximal if r.ann A(E) = l.ann A(E) = {0}.

Definition 1.1.26. Let S be a set with a binary relation ≼. A ring I = I(S) with identity
is called a generalized incidence ring if there exists a maximal set of idempotents
E = {eu : u ∈ S, eu ∈ I} such that:

1. Every idempotent eu is local, i.e. eIe/rad(eIe) is a division ring.


2. The set S is locally finite, i.e. each interval [x, y] = {z ∈ S : x ≼ z ≼ y} is a
finite set.
3. Multiplication and addition in I are defined as follows:
for any x, y ∈ S and f , g ∈ I

( f + g)(x, y) = f (x, y) + g(x, y),

if x  y then ( f g)(x, y) = 0, and if x ≼ y then



( f g)(x, y) = f (x, z)g(z, y)
x ≼z ≼y

where f (x, y) = ex f ey and g(x, y) = ex gey .


4. For any local idempotent f ∈ I there exist x, y ∈ S such that f (x, y) < rad(I).
5. For any set {c(u, v) : u, v ∈ S, c(u, v) ∈ eu Iev } there exists f ∈ I such that
f (u, v) = eu f ev = c(u, v) for any u, v ∈ S.

Example 1.1.27. Let A be a semiperfect ring. Then there exists a finite set E =
{e1, e2, . . . , en } of pairwise orthogonal local idempotents of A such that 1 = e1 + e2 +
· · · + en . Define the binary relation on E such that ei ≼ e j if and only if ei Ae j , 0.
Then A with respect to this binary relation is a generalized incidence ring in sense of
definition 1.1.26.

Remark 1.1.28. Another generalization of these classes of rings is the class of


incidence rings of a group automata which is at present studied actively in the
literature (see, in particular [120], [121], [122]).

Remark 1.1.29. If S and S ′ are isomorphic then of course the (various) incidence
rings I(S, A) and I(S ′, A) are isomorphic. One may wonder to what extent the reverse
is true. Some results and references pertaining to this are given in section 1.6 below.

1.2 Incidence Rings I(S, D)

In this section we consider the particular example of incidence rings I(S, D) of a finite
poset S over a division ring D.
Rings Related to Finite Posets 9

Let S = {α1, α2, . . . , αn } be a finite poset with a partial order relation ≼, and
D a division ring. Denote by Mn (D) the ring of all (n × n)-matrices over D with
matrix units ei j , where dei j = ei j d for all d ∈ D. Consider the subring T(S, D) ⊂
Mn (D) which is generated by those ei j ∈ Mn (D) for which αi ≼ α j in S. This is
equivalent to the following: eii T(S, D)e j j = D if and only if αi ≼ α j in S, otherwise
eii T(S, D)e j j = 0.
Obviously, T(S, D) = I(S, D). We will often denote this ring T(S) if a division
ring D is fixed and known from the context.

Example 1.2.1. Let S = {a, b, c, d : a ≺ b ≺ d; a ≺ c ≺ d} be a rhombus, which has


a diagram of the form (1.1.5), and let D be a division ring. Then

D D D D
© ª
­ 0 D 0 D®
T(S) = ­ ®. (1.2.2)
­ 0 0 D D®
« 0 0 0 D¬
As was shown in [103, Example 9.1.10], the ring T(S) is a piecewise domain but it is
neither right nor left hereditary.

Example 1.2.3. Let S = {α1, . . . , αn } be a linearly ordered set, i.e. αi ≼ α j if and


only if i ≤ j, and let D be a division ring. In this case T(S) = Tn (D), where Tn (D) is
a ring of all upper triangular (n × n)-matrices over D. This ring is Artinian serial and
hereditary. Moreover, by [100, theorem 13.5.2], any serial Artinian hereditary ring
A is Morita equivalent to a finite direct product of rings isomorphic to rings of upper
triangular matrices over division rings.

Proposition 1.2.4. There is a labeling of a poset S such that the ring T(S) is a
triangular ring, and so is isomorphic to a subring of Tn (D), where n is the number
of elements of the poset S.

Proof. Since S = {α1, . . . , αn } is a finite poset, S can be labeled in such a way that
αi ≼ α j if i ≤ j. Then for this numbering T(S) is obviously a triangular ring and
T(S) ⊆ Tn (D). Moreover, one has a strong equality if (and only if) S is a linearly
ordered set. 

Remark 1.2.5. In what follows, T(S, D) will be always assumed to be a triangular


ring.

From proposition 1.1.10 we immediately obtain the following statement.

Proposition 1.2.6.

1. A ring T(S, D) is a primely triangular Artinian semiperfect ring.


2. The Jacobson radical R of T = T(S, D) is equal to the prime radical N of
T(S, D), and the two-sided Peirce decomposition of R has the following form:
eii Reii = 0 and eii Re j j = eii T e j j f or i, j = 1, 2, . . . , n; i , j.
10 Algebras, Rings and Modules

Recall that a semiperfect ring A with Jacobson radical is a basic ring if A/R is a
direct sum of division rings.

Proposition 1.2.7. A ring T(S, D) is a basic ring.

Proof. This follows immediately from proposition 1.2.6. 

Proposition 1.2.8. A ring T(S, D) is Artinian and possesses a classical ring of


e which coincides with T(S, D).
fractions T

Proof. By proposition 1.2.6, T(S, D) is a primely triangular Artinian semiperfect


ring with Jacobson radical R which coincides with the prime radical N. Therefore
e which is an
by [103, theorem 8.5.31] T(S) possesses a classical ring of fractions T
Artinian ring.
Now let’s show that T(S) = T. e From [103, lemma 8.5.26] it follows that any
regular element r ∈ T(S, D) has the form

r = diag(r1, r2, . . . , rn ) + x,

where each ri is a non-zero regular element of eii T(S, D)eii = D (i = 1, 2, . . . , n),


and x ∈ R. Therefore any regular element of T(S, D) is invertible in T(S, D), and so
e 
T(S, D) = T.

A finite poset S is said to be connected if the Hasse diagram of S is connected. It is


obvious that a ring T(S, D) is indecomposable if and only if the poset S is connected.
Since T = T(S, D) is an Artinian semiperfect ring it is possible to construct the quiver
Q(T) of this ring. Recall that Q(T) = Q(T/R2 ). Let Pi = eii T be a right principal
module of T(S, D). If R is the Jacobson radical of T(S, D) then the right quiver of
T(S, D) can be constructed in the following way. If
⊕ t
P(eii R)  P ji j
j

or ⊕ t
eii R/eii R2  U ji j ,
j

where the U j are right simple T/R-modules, then in the quiver Q(T) the vertex i is
connected with the vertex j by ti j arrows.

Proposition 1.2.9. The quiver Q(T) of the ring T = T(S, D) coincides with the Hasse
diagram H(S) of the poset S.

Proof. One can assume that T = T(S, D) is an indecomposable ring. Then by [100,
theorem 11.1.9], the quiver Q(T) is connected. Let {ei j } be the set of all matrix units
of Mn (D) and 1 = e11 + e22 + · · · + enn the natural decomposition of the identity of
T(S) into a sum of pairwise orthogonal idempotents.
Rings Related to Finite Posets 11

Consider the Hasse diagram H(S) of the poset S and the following cases.

1. Assume that there is an arrow i → j in the diagram H(S), which means that
eii Re j j = D and eii Rekk = 0 or ekk Re j j = 0 for any integer 1 ≤ k ≤ n. Then
∑n
eii R2 e j j = k=1 eii Rekk ekk Re j j = 0. Therefore there is exactly one arrow
i → j in the quiver Q(T).
2. Assume that αi ≼ α j and there is no arrow of the form i → j in H(S).
This means that there is an integer 1 ≤ k ≤ n such that eii Rekk = D or
∑n
ekk Re j j = D. Then eii R2 e j j = k=1 eii Rekk ekk Re j j = D and therefore
eii Re j j /eii R2 e j j = 0, i.e. there is no arrow i → j in the quiver Q(T).

Conversely, suppose that there is an arrow i → j in the quiver Q(T). This


means that there is an exact sequence

P j −→ eii R −→ 0

Therefore eii Re j j , 0 and so αi ≼ α j , and there is an arrow i → j in the


diagram H(S). 

Recall that the endomorphism ring End A(P) of a finitely generated projective
A-module P is called a minor of the ring A (see [100, Section 13.3]). If P can
be decomposed into a sum of n indecomposable modules then End A(P) is called a
minor of order n. So that the ring End(eA) = eAe is a minor for any idempotent e
of a ring A. If e is a sum of n primitive idempotents then End(eA) = eAe is a minor
of order n. It was shown that the minors of right (left) Noetherian rings are right
(left) Noetherian, the minors of right (left) Artinian rings are right (left) Artinian (see
[100, Theorem 3.6.1]). From [100, Theorem 10.3.8] and [100, Corollary 10.3.11] it
follows that any minor of a semiperfect ring is semiperfect. If e is an idempotent of
a hereditary (semihereditary) ring then eAe is also hereditary (semihereditary) (see
[103, proposition 4.4.6(4)] and [103, theorem 8.2.16]).
Let
A11 A12 · · · A1n
© ª
­ A21 A22 · · · A2n ®
­
A=­ . .. .. .. ®®
­ .. . . . ®
« A n1 A n2 · · · A nn ¬

be a generalized matrix ring, where Aii are rings with identity 1i and Ai j are
Aii -A j j -bimodules. Then the identity of the ring A can be decomposed into the
sum 1 = e1 + e2 + · · · + en , where ei is an element of A with entry equal to 1i at (i, i)
place and 0 otherwise. So that ei Ae j = Ai j . Obviously, each ei is an idempotent of
A. Therefore eAe is a minor of A if e is a sum of some distinct idempotents ei whose
index is in a subset X ⊆ {1, 2, . . . , n}.

Proposition 1.2.10. A ring T(S, D) is hereditary if and only if the Hasse diagram
H(S) of S is a disconnected union of trees, i.e. the poset S contains no subposet
whose diagram is a rhombus.
12 Algebras, Rings and Modules

Proof. Let T = T(S, D) be a hereditary ring. Assume that the Hasse diagram of S
contains a subposet whose Hasse diagram is a rhombus. This means that T contains
an idempotent e such that eT(S)e = B, where

D D D D
© ª
­0 D®®
­ D 0
B=­ ®.
­0 0 D D®®
­
«0 0 0 D¬

This is impossible, since T(S) is a hereditary ring and for any non-zero idempotent
e ∈ T(S) the ring B = eT(S)e must be also a hereditary ring by [103, proposition
4.4.6(4)], but the ring B is not hereditary.
Conversely, suppose the two-sided Peirce decomposition of T(S, D) does not
contain minors of the form eT(S)e = B. Then the Hasse diagram of S is a
disconnected union of acyclic simply laced quivers with no extra arrows such that its
underlying graph S (obtained from S by deleting the orientation of the arrows) is a
disconnected union of trees. From proposition 1.2.9 it follows that the ring T(S, D)
can be considered as the path algebra corresponding to the graph S over a division
ring D. Therefore T(S, D) is a hereditary ring, by [101, theorem 2.3.4]. 

Proposition 1.2.11. A ring T = T(S, D) is an Artinian semidistributive piecewise


domain.

Proof. Since for any primitive pairwise orthogonal ( idempotents


) ( e,) f ∈ T(S, D) = T
D D D 0
the ring (e + f )T(e + f ) is either of the form or , the ring T(S, D)
0 D 0 D
is semidistributive, by [103, theorem 1.10.9].
Write Pi = eii T for i = 1, . . . , n. Let φ : Pi → P j be a non-zero homomorphism.
Then φ(eii a) = φ(eii )a = e j j a0 eii a, where a0, a ∈ T and e j j a0 eii is a non-zero
element from e j j T eii = D. Thus d0 = e j j a0 eii defines a monomorphism. Therefore
the ring T(S, D) is a piecewise domain. 

Proposition 1.2.12. If the Hasse diagram H(S) of a poset S contains no rhombuses


then T(S, D) is an Artinian hereditary semidistributive ring.

Proof. This follows immediately from propositions 1.2.10 and 1.2.11. 

Proposition 1.2.13. A ring T(S, D) is serial if and only if S is a disconnected union


of linearly ordered sets.

Proof. Since a ring T(S, D) is indecomposable if and only if the poset S is connected,
one can assume that T(S, D) is indecomposable. If S is a chain, then the statement
follows from example 1.2.3.
Rings Related to Finite Posets 13

Conversely, suppose that T(S, D) is a serial indecomposable ring. Then by


proposition 1.2.9 the quiver of the ring T(S, D) coincides with the diagram of the
poset S and by [103, theorem 1.12.2] it is a chain. Thus S is a linearly ordered poset.


1.3 Right Hereditary Rings A(S, O)

Let O be a discrete valuation ring (not necessarily commutative) with division ring
of fractions D and Jacobson radical M = radO = πO, where π ∈ O is a generator
of M. By [100, corollary 10.2.2] O is a local Noetherian hereditary ring which is a
right and left principal ideal domain (PID) and M is its unique maximal right and left
ideal. Recall (see [100, section 8.4]) that each non-zero element x ∈ D is uniquely
representable in the form x = επ k , where ε is an invertible element of the ring O and
k ∈ Z. Each proper O-module X ⊂ D is cyclic and can be represented in the form
X = π k O = Oπ k , where k ∈ Z. Write
O O ··· O
© ª
­M O ··· O®
Hn (O) = ­­ . .. .. .. ®® (1.3.1)
­ .. . . .®
«M M ··· O¬

which is a subring of the matrix ring Mn (D). Clearly, Hn (O) is a Noetherian serial
prime hereditary ring. And so, by the Goldie theorem, it has a classical ring of
fractions, which is Mn (D).
Let {Oi }i=1,...,k be a family of discrete valuation rings (not necessarily
commutative) with Jacobson radical Mi and a common skew field of fractions
D. Let S = {α1, α2, . . . , αn } be a finite poset with a partial order ≼, and
S0 = {α1, α2, . . . , αk } the subset of minimal points of S (k 6 n). Then S = S0 ∪ S1 ,
where S1 = {αk+1, αk+2, . . . , αn }. Corresponding to this partition of S consider a
poset S with weights so that the point i has the weight Hni (Oi ) for i = 1, 2, . . . , k;
ni ∈ N; and all other points have the weight D.
Construct a ring A = A(S, S0, S1 ; O1, . . . , Ok ; D; n1, n2, . . . , nk ) (or shortly,
A(S, O)) as a generalized matrix ring of the form

A11 A12 ··· A1n


© ª
­ A21 A22 ··· A2n ®
­ . .. .. .. ®®
­ . .
­ . . . ®
« An1 An2 ··· Ann ¬

where Aii = Hni (Oi ) for i = 1, 2, . . . , k; A j j = D for j = k + 1, . . . , n and Ai j is an


(Aii, A j j )-bimodule for i, j = 1, 2, . . . , n. Moreover, if{ αi  α j in S then Ai j = 0,
ni if i ∈ S0
otherwise Ai j = [D D . . . D]T = [D(ki ) ]T , where ki = .
1 otherwise
14 Algebras, Rings and Modules

So this generalized matrix ring A has the following form

Hn1 (O1 ) ... 0 M1


© ª
­ .. .. .. .. ®
­ . . . . ®
A=­ ®, (1.3.2)
­ ... Hn k (Ok ) ®
­ 0 Mk ®
« 0 ... 0 T (S1, D) ¬
where T(S1, D) is the incidence ring of the poset S1 over the division ring D, and Mi
is an (Hni (Oi ), T(S1, D))-bimodule for i = 1, 2, . . . , k. Obviously, A is a subring of
Ms (D) where s = n1 + n2 + · · · + nk + (n − k).
The identity of the ring A can be decomposed into the sum 1 = f1 + f2 + · · · + fn ,
where fi is an element of A with entry equal to identity 1i of the ring Aii at (i, i)-th
place and 0 otherwise. Obviously, each fi is an idempotent of A; fi A fi = Hni (Oi ) for
i = 1, 2, . . . , k and fi A fi = D otherwise; f A f = T(S1, D) for f = fk+1 + fk+2 +· · ·+ fn ;
and fi A f = Mi for i = 1, 2, . . . , k.

Proposition 1.3.3. Let {Oi }i=1,...,k be a family of discrete valuation rings with
common division ring of fractions D. Then A(S, O) is a right Noetherian semiperfect
ring.

Proof. By [103, theorem 1.1.23] the ring A is right Noetherian. Since the identity of
A = A(S, O) decomposes into a finite sum of pairwise orthogonal local idempotents,
A is semiperfect by [103, theorem 1.9.3]. Since all the Hni (Oi ) are Noetherian rings,
T(S1, D) is an Artinian ring, and all Mi are finite dimensional vector spaces over D,
A is a right Noetherian ring by [103, theorem 1.1.23]. 

Let N be the prime radical of a semiperfect ring A = A(S, O). Then N has the
following form
I1 ... 0 M1
© ª
­ .. . . .. .. ®
­ . . . . ®
N=­ ®
­ O ... I Mk ®
­ k ®
« 0 ... 0 N1 ¬

where Ii is the prime radical of Hni (Oi ) for (i = 1, 2, . . . , k), and N1 is the prime
radical of T(S1, D). Since all the Hni (Oi ) are prime rings, Ii = 0. By [103, proposition
1.1.25] the prime radical of the Artinian ring T(S1, D) coincides with its Jacobson
radical R1 = rad T(S1, D). Therefore

0 ... 0 M1
© ª
­ ... .. .. .. ®
­ . . . ®
N = ­­ ®
®
­ 0 ... 0 Mk ®
­ ®
0 ... 0 R1
« ¬
Another random document with
no related content on Scribd:
ähnlichen Verwüstungen, wie man sie hier täglich zu Dutzenden
sieht, vorzubeugen. Beim Europäer mit seiner weißen Haut ist das
Auffinden des Sandflohes übrigens weit leichter, als es den
Schwarzen gemacht wird, von deren Haut sich der dunkle Punkt
kaum abhebt. Die vier oder fünf Sandflöhe, die mich trotz steten
Tragens hoher, geschlossener Schnürschuhe bisher zu ihrem Sitz
auserkoren haben, hat mir der vielgewandte Knudsen
herausgehoben; ein Auswaschen der Höhlung mit Sublimat
erscheint mir dabei immer ganz angebracht. Die Neger haben ein
anderes Desinficiens, sie füllen die Öffnungen mit
Wurzelgeschabsel; in einem winzigen Makuadorf am Steilabhang
des Plateaus südlich von Newala sah ich eine Frau, die den Raum
unter den Nägeln prophylaktisch mit Wurzelpulver ausstopfte. Ob es
der Alten etwas nützen wird, wer weiß es.
Der Rest der vielen kleinen Hindernisse, die uns hier das Dasein
erschweren, wirkt mehr komisch als ernsthaft. In Ermangelung von
etwas anderem Rauchbaren greifen Knudsen und ich jetzt zu dem
Inhalt einer vom Inder in Lindi bezogenen Zigarrenkiste. Diese ist
sehr schön beklebt und aufgemacht, aber wehe dem Unglücklichen,
der sich, wie wir, mit ihrem Inhalt befaßt! Ob diese schwelenden
Giftnudeln Opium oder ein anderes Narkotikum enthalten, von uns
beiden weiß es niemand zu sagen, denn nach dem zehnten Zuge
sind wir beide „matt“; dreiviertel betäubt und hundeelend liegen dann
Wiking und Deutscher in sich zusammengesunken da. Langsam
erholt man sich — was geschieht? Nach einer halben Stunde greift
man doch wieder zu dem scheußlichen Kraut; so unstillbar ist hier in
den Tropen der Drang zum Rauchen!
Auch meine jetzigen Fieberanfälle sind kaum geeignet, noch
ernst genommen zu werden. Ich habe ihrer hier in Newala nicht
weniger als drei gehabt, aber alle mit unglaublich kurzem Verlauf.
Emsig fragend, schreibend und notierend quäle ich mich mit meinen
„Gelehrten“ herum, der starke Mittagskaffee hat die Lebensgeister
mächtig angeregt; das Gehirn arbeitet außerordentlich intensiv, so
daß die Arbeit rasch vorwärtsschreitet. Eine wohltuende Wärme
durchrieselt den ganzen Körper, macht jedoch mit einem Male einem
heftigen Kältegefühl Platz, das mich jetzt, beim wärmsten
Sonnenschein, nachmittags 3½ Uhr, bereits zwingt, den Überzieher
anzulegen. Jetzt arbeitet auch das Gehirn nicht mehr so scharf und
logisch, besonders bei syntaktischen Feststellungen des schwierigen
Imakuāni, der Sprache der Makua, an die ich mich zum Überfluß
auch noch herangewagt habe. Da halte ich es denn doch allmählich
für angezeigt, meine Temperatur zu messen, der Einfachheit halber
gleich im Sitzen und ruhig weiterarbeitend; 38,6° ist das Ergebnis!
Nun aber hinaus, meine Herren, heißt es im gleichen Augenblick!
Wenige Minuten später steht mein Bett in der Barasa; unmittelbar
darauf liege ich auch schon darin und beginne mich mit heißem
Zitronenwasser innerlich zu behandeln. Drei Stunden später zeigt
das Thermometer gegen 40°; ich lasse mich jetzt, beim Einsetzen
des Abendwindes, mitsamt meinem Bett ins Zelt zurücktragen —
würde ich meinen furchtbar schwitzenden Körper der eisigen
Abendtemperatur aussetzen, so könnte das meinen Tod bedeuten
—, liege dort noch eine kleine Weile und finde dann zu meiner
Beruhigung, daß das Fieber nicht mehr steigt, sondern anfängt
zurückzugehen. Das ist ungefähr 7½ Uhr; als ich kurz nach 8 Uhr
noch einmal messe, ist die Kurve zu meinem maßlosen Erstaunen
auf unter 37° heruntergegangen; mir ist absolut wohl; ich lese noch
ein paar Stunden und könnte sehr wohl rauchen, wenn ich etwas
Ordentliches hätte. Aber Inderzigarren? Pfui Teufel!
Wie ist so etwas denkbar? muß ich mich selbst als Laie fragen.
Das kann doch unmöglich Malaria sein; näher liegt die Vermutung,
daß diese rasch verlaufenden, hohen Fieberanfälle die Folge einer
zu intensiven Sonnenbestrahlung sind, eine Art Insolationsfieber
oder Sonnenstich. Wenn ich mein Fiebernotizbuch nachsehe, wird
mir dies immer wahrscheinlicher, denn regelmäßig treten diese
Anfälle im Anschluß an größere Strapazen und langen Aufenthalt in
praller Sonne ein. Für mich haben diese kurzen Unpäßlichkeiten
wenigstens das Gute, daß sie mich nur stundenweise von der Arbeit
abhalten, denn am nächsten Morgen bin ich regelmäßig wieder
vollkommen frisch und gesund.
Nicht so gut geht es leider meiner Perle von Koch und dem
Knaben Moritz; jener leidet an einer ungeheuren Hydrozele, die ihm
kaum erlaubt aufzustehen, Moritz aber hat Dunkelarrest wegen
seiner entzündeten Augen. Leider versteht Knudsens Koch, ein bis
vor wenig Wochen gänzlich unbeleckter Wilder von irgendwo aus
dem Busch, noch weniger als mein Omari. Folge: Nils Knudsen ist
selbst zum Koch avanciert. Er hat diese seine neue Tätigkeit
sogleich mit einer großen Tat begonnen; da wir nichts Ordentliches
mehr zu essen haben, hat er die vier von Matola erstandenen
Ferkel, hübsch säuberlich in einen großen Tragkorb gepackt, von
Chingulungulu heraufholen lassen und kaltblütig das größte von
ihnen gemordet. Den ersten Schweinebraten haben wir
leichtsinnigerweise doch Knudsens wildem Koch anvertraut; er war
infolgedessen ungenießbar; den Rest des Tieres haben dann wir zu
einem Gelee verarbeitet, das uns nach den langen Wochen der
Unterernährung herrlich mundet und von dem wir mittags und
abends geradezu fabelhafte Portionen vertilgen. Wenn nur nicht die
ewigen Teltower Rübchen dabei wären! O du gesegnete Stadt auf
märkischem Sande, wer hätte je geahnt, daß du so nachhaltig in die
Ernährung eines stillen, deutschen Gelehrten eingreifen würdest!
Dieser boshafte Dr. Jaeger! Er war ein Mann von Zeit und Muße; ihm
halste daher die Landeskundliche Kommission die Besorgung aller
Nahrungsmittel für seine und meine Expedition auf. Feierlich
überweist mir eines schönen Tages in Daressalam der mit der
Verpackung dieser Sachen betraute Handlungsbeflissene meinen
Anteil. Seitdem leide ich unter einer ständigen Rübenfurcht; ich habe
das Gericht an sich ganz gern, aber nur einmal im Jahre, ungern
häufiger. Doch wie ergeht es mir hier? Ich trete an die Kiste heran,
die gerade leergegessen werden muß; der Deckel fliegt hoch; ein
Griff hinein, eine Konservenbüchse kommt zum Vorschein; ein Blick
auf die Etikette: Teltower Rübchen. Puh! Die Dose verschwindet; ein
zweiter Griff; dasselbe Ergebnis; ein dritter, nichts anderes. Nach
langem Suchen erst kommt dann ein anderes Gemüse zutage; oder
auch nicht, denn diese anderen sind allmählich zu Ende gegangen,
nur die Teltower sind geblieben! „Denn helpt dat nich“, sage ich mit
Fritz Reuter; aber zehn Jahre lang esse ich zu Hause keine Teltower
mehr!
Bei all diesem kleinen Leid, das aber nun einmal dazu gehört, um
Afrika schmackhaft zu machen, gibt es wenigstens e i n erfreuliches
Moment: Nils Knudsen hat mit der Geschicklichkeit eines
Feinmechanikers meinen 9 × 12-Apparat wieder in Ordnung
gebracht oder ihn doch wenigstens so weit wieder hergestellt, daß
ich ihn mit einiger List gebrauchen kann. Wie der Mann ohne
Fingernägel mit dieser kniffligen Arbeit hat fertig werden können, bei
der er den ungemein komplizierten Momentverschluß nur mit Hilfe
eines plumpen Schraubenziehers auseinandernehmen und wieder
zusammensetzen mußte, ist mir noch heute schleierhaft, aber er hat
es geschafft. Der Mangel an Fingernägeln hingegen zeigt den guten
Nils von einer Seite, die mit seiner bei der Apparatreparatur
bewiesenen Intelligenz merkwürdig kontrastiert, die andererseits
allerdings auch aufs innigste mit seinem zehnjährigen
Hinterwäldlertum zusammenhängt. Wäscht er da eines Tages in
Lindi irgendeinen Köter. Dieser muß wohl eines schärferen
Reinigungsmittels bedürftig gewesen sein, denn Nils hat ein Gefäß
mitbekommen, dessen Inhalt stark und kräftig riecht. Gewissenhaft
nimmt unser Freund die Reinigung vor, wundert sich ein wenig, daß
sie dem Hunde sehr schlecht bekommt, ist dann aber sehr erstaunt
darüber, daß ihm seine eigenen zehn Fingernägel im Laufe weniger
Tage wegeitern. „Wie kann ich aber auch wissen, daß man
Karbolineum verdünnen muß“, knurrt er oftmals noch jetzt entrüstet,
wenn er seine schrecklich zugerichteten Fingerenden sorgenvoll
mustert!
Weit und breit haben wir die Umgegend durchschweift, seitdem
wir in Newala hausen; zunächst alter Gewohnheit gemäß, sodann
aber, weil der Akide Sefu mit der Zusammenstellung seines
Gelehrtenkollegiums durchaus nicht so rasch fertig geworden ist, wie
er sich zuerst anheischig gemacht hatte. Aber das schadet weiter
nicht, denn auch bloß von außen gesehen, sind Land und Leute
interessant genug.
Das Makondeplateau gleicht einer großen, rechtwinkligen, an
den Ecken abgerundeten Tafel; es ist, vom Indischen Ozean bis
Newala gemessen, etwa 120 Kilometer lang und im Mittel zwischen
dem Lukuledi und dem Rovuma gegen 80 Kilometer breit; es umfaßt
also gegen zwei Drittel der Fläche des Königreichs Sachsen. Nun ist
diese Fläche nicht horizontal, sondern von ihrem Südwestrande
flach, aber ganz gleichmäßig gegen den Ozean hin geneigt. Von der
Schwelle, auf der Newala liegt, kann man viele Meilen über den
Makondebusch nach Osten und Nordosten schauen, ohne einem
Hindernis zu begegnen; es ist ein grünes Meer, aus dem nur hie und
da dichte Rauchwolken in langer Erstreckung emporwirbeln und -
wallen, zum Zeichen dafür, daß auch hier Menschen wohnen und
daß sie ihre Feldkultur ganz nach der Weise so vieler anderer
Naturvölker vorwaltend auf die Verbrennung des
niedergeschlagenen Holzbestandes gründen. Dessen Asche ist
zugleich die einzige Düngung. Selbst am strahlend hellen Tropentag
ist so ein Brand ein großartiges Schauspiel.
Ungleich weniger wirkungsvoll ist der Eindruck, den gegenwärtig
die große Ebene vom Plateaurand aus erweckt. Sooft es mir meine
Zeit gestattet, unternehme ich den kleinen Ausflug an diesen Rand,
bald hierhin, bald dahin, stets in der stillen Hoffnung, endlich einmal
eine klare Luft mit weiter Aussicht vorzufinden; immer aber
vergebens: wohin man dort unten schaut, allerorten steigen
Rauchwolken hoch, der lebhafteste Beweis für die unausgesetzte
Tätigkeit des Waldbrennens; rauchig und dunstig ist auch die ganze
Luft. Schade drum, das Panorama von hier bis weit hinten an die
Madjedjeberge muß unter günstigeren Umständen wirklich großartig
sein. Jetzt haben photographische Aufnahmen eigentlich kaum einen
Zweck, die Profilzeichnung aber gibt nur einen sehr schwachen
Begriff der ganzen Szenerie.
Bei einem dieser Ausflüge habe ich mich absichtlich selbst
einmal am Makondebusch versucht. Der Plateaurand von heute ist
das Ergebnis einer ungeheuer tiefgreifenden Zerstörung durch
Erosion und Abrutschung; überall greifen kurze, aber Hunderte von
Metern tiefe Täler in die Makondeschichten ein. Eine Folge des
lockeren Gefüges dieser Formation ist es, daß nicht nur die
Seitenwände dieser Täler fast senkrecht abstürzen, sondern daß die
Täler auch mit einer ebenso steilen Rückwand enden; dergestalt ist
der Westrand des Makondeplateaus von lauter Talkesseln umsäumt.
Um von einer Seite eines solchen Kessels auf die andere zu
gelangen, habe ich mich eines Tages mit einem Dutzend meiner
Leute durch den Busch geschlagen. Es war eine sehr lichte Stelle,
mit mehr Gras als Buschwuchs; aber welche Mühe hat dieser Weg
von ein paar hundert Metern gekostet, und wie sahen wir alle
nachher aus! Die dünnen Kattunstoffe meiner Leute in Fetzen, sie
selbst aus hundert kleinen Wunden blutend; sogar unsere derben
Khakistoffe hatten den Dornen dieser Vegetationsformation nicht
standgehalten.

Negerpfad im Makondebusch. Gegend von Mahuta.

Meine seit langem gehegte Ansicht über die Entstehung dieses


Makondebusches hat sich immer mehr befestigt: er ist ohne Zweifel
kein Naturprodukt, sondern erst die Folge der menschlichen Kultur.
Wohin der Mensch hier oben auf dem Hochland noch nicht mit
Hacke und Axt gedrungen ist — ein halbwegs geübtes Auge sieht
dies ohne weiteres —, da steht auch heute noch ein wirklicher,
wunderschöner Hochwald, der den Vergleich mit unserem deutschen
Mischwald sehr wohl aufzunehmen vermag. Wo der Mensch aber
jemals seine Hütte gebaut und sein Feld beackert hat, da entsteht
hinterher dieser gräßliche Busch. Geht man auch nur ein paar
Stunden irgendwo auf dem Hochland die Barrabarra entlang, so hat
man vollauf Gelegenheit, diese Metamorphose in jeder Phase ihrer
Entwicklung zu verfolgen. Seitwärts tönt hallender Axthieb herüber,
nicht bloß von einer Stelle, sondern über einen ganzen Komplex
verteilt. Wenige Schritte weiter sieht der Wanderer, was vorgeht;
wohl meterhoch und höher liegt das niedergeschlagene Unterholz
geschichtet; zwischen ihm aber ragen als letzte Säulen alter Pracht
die Stämme des Hochwaldes. Doch auch sie gewähren ein Bild des
Jammers; der böse Makonde hat sie geringelt, d. h. er hat sie
ringsherum in breitem Bande der Rinde beraubt, so daß sie dem
Absterben verfallen sind; zudem hat er noch eine Reisigpyramide
um sie aufgebaut. Unverdrossen hacken Vater und Sohn, Mutter und
Schwiegersohn im Hintergrunde weiter; kaum daß das sonst so
neugierige Volk nach dem weißen Fremdling aufschaut. Und kommt
dieser Fremdling eine Woche später desselben Weges gezogen,
verschwunden ist das Reisig, verschwunden sind die Pyramiden;
eine dicke Aschenschicht lagert, wo vor kurzem noch grünender
Wald sich breitete. Die starken Bäume aber recken ihre noch immer
glimmenden, schwelenden Stämme und Äste in stummer Anklage
zum Himmel, oder aber sie sind bereits niedergebrochen, mehr oder
minder zu Asche verglüht und zeichnen sich dann als weißer
Streifen auf dunklem Grunde ab.
Das ist der Zerstörungsprozeß, den der Makonde in gleicher
Weise am jungfräulichen Urwalde wie auch an den Stellen seines
Heimatlandes vornimmt, wo er vor Jahren schon einmal geackert
hat, nur daß er im letztern Fall des Verbrennens der großen Bäume
überhoben ist. Diese gibt es in der sekundären Buschformation nicht
mehr.
In das gebrannte und mit der Hacke gelockerte Stück Waldland
sät der Eingeborene sein Getreide, pflanzt er sein Gemüse. Im
ganzen Lande hat er Beetkultur. Diese erfordert eine sorgsame
Pflege, die ihr der Neger auch zuteil werden läßt; Unkraut wird im
Süden Deutsch-Ostafrikas nicht geduldet. Mißernten kommen wohl
im trockneren Tiefland vor, auf dem niederschlagreicheren,
allmorgendlich taufeuchten Hochland sind sie ganz unbekannt.
Dessen glückliche Bewohner sind sogar in der angenehmen Lage,
die sonst so stolzen Yao und Makua von unten bei sich als Diener
und Knechte zu sehen. Hunger tut weh, und so ziehen es die
Angehörigen jener beiden Völkerschaften vor, einmal eine Zeitlang
da den Diener zu spielen, wo sie sonst zu herrschen gewohnt sind.
Jedoch der leichte sandige Boden ist bald erschöpft, er würde bei
einer nochmaligen Bestellung keine Ernte mehr ergeben. Dies weiß
der Eingeborene seit Jahrtausenden; längst hat er vorgearbeitet und
den Komplex nebenan mit Axt und Feuerbrand urbar gemacht. Auf
ihn siedelt er nunmehr mit seinen mannigfachen Kulturen über; das
alte Feld wird zur Brache. Doch nur ganz kurze Zeit liegt es wüst und
greulich anzusehen da, dann kommt Allmutter Natur und nimmt ihr
mißhandeltes Kind liebevoll in ihre Obhut; tausendfältig sprießt es
allerorten aus dem ausgesogenen Boden hervor, selbst die alten
Baumstrünke schlagen von neuem aus. Im nächsten Jahr ist der
Neuwuchs bereits mehr als kniehoch; rasch wuchert er in die Höhe;
nach wenigen Jahren schon ist er jener undurchdringliche,
schreckliche Busch, der erst wieder fällt, wenn der schwarze Herr
des Landes seinen Turnus beendigt hat und an die alte Stelle
zurückkehrt.
Mit diesem Busch sind die Makonde mit Leib und Seele
verwachsen, ja nach meinen Yaogewährsleuten bedeutet sogar ihr
Name nichts anderes als Buschvolk. Nach ihrer eigenen Tradition
sitzen die Makonde zwar schon seit langen, langen Zeiten hier oben,
aber zu meiner Überraschung legten sie doch eine sehr starke
Betonung auf eine ursprüngliche Einwanderung. Diese sei von
Südosten, von der Rovumamündung und von Mikindani her erfolgt;
der Anlaß dazu sei die ewige Beunruhigung ihrer friedlichen
Vorfahren durch die kriegerischen Schirasi der Küste und die
fortgesetzten Raubzüge der Sakalaven von Madagaskar herüber
gewesen; vor diesen hätten sich die Ur-Makonde auf das
unzugängliche Plateau zurückgezogen. Ich bin in der Völkerkunde
Afrikas auf Grund einer 20jährigen Beschäftigung mit ihr sehr wohl
bewandert, aber daß Bevölkerungsvorgänge in diesem so friedlich
und ruhig erscheinenden Erdteile sogar durch von außen kommende
Hochsee-Unternehmungen bedingt und veranlaßt worden seien, war
mir doch im ersten Augenblick etwas vollkommen Neues. Es wird
indessen schon seine Richtigkeit haben. Warum jedoch die
Makonde gerade im dicksten Busch und weit vom Plateaurand ab
wohnen müssen, und warum sie nicht an die rieselnden Quellen der
Niederung selbst zu dauerndem Wohnsitz herniedersteigen dürfen,
das lehrt aufs klarste ihre wunderhübsche Stammessage. Auch noch
manch anderes Lehrreiche steht darin.
„Die Geburtslandschaft des Stammes, mit Namen Mahuta, ist auf
der Südseite des Plateaus zum Rovuma hin gelegen; dort aber
stand nur dichter Busch. Aus diesem Busch hervor ging ein Mensch,
der sich niemals wusch und schor, der nur wenig aß und trank. Der
ging aus und machte ein Menschenbildnis aus dem Holze eines
Savannenbaumes, nahm es mit sich in seine Buschwohnung und
stellte es dort aufrecht hin. Während der Nacht erwachte das Bildnis
zum Leben, und es war ein Weib. Daraufhin gingen sie zusammen
hinunter zu den Wassern des Rovuma, um sich zu waschen. Hier
gebar das Weib ein Kind, welches jedoch nicht lebend zur Welt kam.
Sie verließen das Land und zogen über die Hochländer bis in das Tal
des Mbemkuru, wo sie sich niederließen. Dort gebar das Weib
abermals ein Kind, das wiederum tot zur Welt kam. Daraufhin
kehrten sie in die hochgelegene Buschlandschaft Mahuta zurück,
und dort wurde das dritte Kind geboren, welches nach der Geburt
am Leben und gesund blieb. Mit der Zeit zeugten sie noch viele,
viele Kinder und hießen sich Wamatanda. Diese bildeten die
Stammfamilie der Makonde, auch Wamakonde genannt, d. h.
Urbewohner. Der Stammvater, der Buschmensch, aber gab seinen
Kindern das Gesetz, daß sie ihre Toten aufrecht begraben sollen
zum Andenken an die erste Mutter, die aus Holz geschnitzt und
aufrecht stehend zum Leben erwacht sei; ferner warnte er seine
Kinder, in die Täler und an die großen Wässer zu ziehen, denn dort
wohne die Krankheit und der Tod. Als Regel solle gelten, daß
mindestens eine Stunde Weges sei von der Hütte bis zum
Wasserplatze; dann würden ihre Kinder gedeihen und von
Krankheiten verschont bleiben.“
Die Urmutter.
Holzskulptur eines
Makondekünstlers.
Die Erklärung des Namens Makonde lautet bei meinen
Gewährsleuten etwas anders als bei Pater Adams, dessen kleinem,
aber inhaltreichem Büchlein: „Lindi und sein Hinterland“ ich diese
Stammessage der Makonde entnehme. Aber sonst stimmt mein
Befund genau mit dem sachlichen Inhalt dieser Stammessage
überein. Waschen? Hapana, gibt es nicht. Wozu auch? Zudem ist
das Wasser spärlich und reicht kaum zum Kochen und Trinken; von
den anderen tut es ja auch keiner; warum soll also gerade ich so
unangenehm auffallen? Scheren aber ist bei dem kurzen,
krauswolligen Haarwuchs kaum vonnöten; also auch dieser
Vorschrift des Urahnen ist leicht zu folgen. Damit aber hört das, was
uns lächerlich dünkt, auf. Von einer Reihe hiesiger Künstler habe ich
eine ziemlich große Anzahl stattlicher, 40 bis 60 Zentimeter hoher
Holzskulpturen erworben, die allesamt Frauen aus der großen
Völkergruppe der Mavia, Makonde und Wamatambwe darstellen; die
Figuren sind merkwürdig gut gearbeitet und geben den Frauentypus
vortrefflich wieder, vor allem auch die später noch zu schildernde
Verschönerung des Körpers mit Ziernarben. Über Zweck und
Bedeutung ihrer Werke befragt, wußten die Künstler nichts
Plausibles anzugeben, oder aber, was ich heute für wahrscheinlicher
halte, sie wollten nicht. Also:

Was man sich nicht erklären kann,


Das sieht man als ’nen Fetisch an.

Ehrlich gestanden, ich habe diesen ins Ethnographische


variierten alten Vers beim Empfang jener Figuren laut in den
sonnendurchglühten Tropentag hinausgesprochen, aber ich habe
wohlweislich nicht danach gehandelt. Einstweilen mußte ich mich mit
der kargen Angabe eines der Künstler begnügen, die Figuren gäben
lediglich das „nembo“ wieder, die Körperverunstaltung durch Lippen-
und Ohrscheiben und Ziernarben. Mit der Adamsschen Sage jedoch
rücken diese Figuren ohne weiteres in ein anderes Licht, sie sind
doch mehr als zwecklose Kostümpuppen, ja man darf dreist
annehmen, daß sie, wenn auch der Mehrzahl der heutigen Makonde
unbewußt, Darstellungen jener Urmutter sind. Diesmal wäre der alte
Vers also doch angebracht gewesen, denn die Urmutter gehört
ebenso in den Kreis des hiesigen Eingeborenenkultus wie die Ahnen
überhaupt.
Auf die Vorgeschichte des Volkes bezieht sich in der
Stammessage unzweifelhaft zunächst der Hinweis auf den Abstieg
von Mahuta hinunter zu den Wassern des Rovuma, sodann der
andere auf die Wanderung über die Hochländer bis in das Tal des
Mbemkuru; beide Wanderungen des Urelternpaares bedeuten in
Wirklichkeit wohl Wanderungen des Volkes selbst. Der Abstieg in
das nahe, an seinen Rändern außerordentlich fruchtbare und
wildreiche Rovumatal ist ohne weiteres verständlich; doch auch das
Überschreiten der Lukuledisenke, der Aufstieg zum Rondoplateau
und der erneute Abstieg zum Mbemkuru liegen durchaus innerhalb
des Bereichs der Wahrscheinlichkeit, denn alle diese Gebiete weisen
genau dieselben Naturbedingungen auf wie der äußerste Süden.
Nun kommt aber etwas gerade für unser „bakterielles“ Zeitalter
höchst Interessantes. Die Ur-Makonde sind in den sumpfigen
Flußniederungen ihres Lebens nicht froh geworden, Krankheiten
waren bei ihnen an der Tagesordnung, und viele starben; erst
nachdem sie wieder nach Mahuta in die Heimat zurückgekehrt
waren, besserte sich der Gesundheitszustand des Volkes. Wir sehen
im Neger gern und mit Vorliebe den naturfremden und
naturfürchtigen Dümmling, dem alles Ungemach von bösen Geistern
und Naturgewalten herrührt. Viel richtiger wird es sein, hier
anzunehmen, daß die Leute malariadurchseuchte und malariafreie
Gebiete sehr bald haben unterscheiden lernen. Diese Erkenntnis
schlägt sich dann nieder in der Warnung des Urvaters, nicht wieder
in die Täler und an die großen Wässer zu ziehen, denn dort wohne
die Krankheit und der Tod. Um aber auch gleichzeitig vor den bösen
Mavia auf der Südseite des Rovuma gesichert zu sein, wird noch
bestimmt, daß jede Siedelung um einen Minimalabstand von jenem
Steilrand abliegen solle. So wohnen sie heute noch.
Sie wohnen so auch ganz gut, jedenfalls besser und geschützter
als die Makua, die modernen Eindringlinge des Südens, die hier auf
dem Westrande des Plateaus in ziemlich breiter Zone Fuß gefaßt
haben. Von der Stattlichkeit der Yaohäuser unten in der Ebene und
besonders in Massassi, Susa und Chingulungulu hat weder die
Behausung der Makua, noch die der Makonde etwas an sich. Jumbe
Chauro, ein an der Barrabarra nach Mahuta unweit Newala
gelegener Makondeweiler, ist von allen mir bisher bekannten
Siedelungen des Stammes noch der bei weitem stattlichste; seine
Hütten sind auch recht geräumig. Doch wie ruppig ist ihre bauliche
Ausführung gegenüber den fast eleganten Palästen der
Elefantenjäger in der Ebene! Das Dach ist noch verwahrloster, als es
in der Trockenzeit hier überall Usus ist; an den Wänden nur hie und
da die kümmerlichen Anfänge oder die kläglichen Reste eines
Lehmbewurfs; das Innere aber eine wahre Hundehütte; Schmutz,
Staub und Unordnung überall; von Zimmereinteilung ist nur in
wenigen Hütten etwas zu merken, und dann ist sie auch nur durch
ganz liederlich zusammengeflickte Bambuswände hergestellt.
Nur in einem habe ich hier einen Fortschritt feststellen können, in
der Methode des Hausverschlusses. Diese ist im ganzen Süden bei
aller Einfachheit sinnreich; die Tür besteht stets aus derben Stangen
von Bambus oder Holz, die mit dem uns schon bekannten
Bindemittel des Baumbastes an zwei Querriegel gebunden werden;
auch die Drehung um den einen Türpfosten erfolgt in zwei Schleifen,
und zwar nach innen. Will der Bewohner sein Haus verlassen, so
nimmt er zwei derbe Stangen her, oberarmdick und etwa 1,5 Meter
lang. Die eine lehnt er von innen schräg gegen die Mitte der Tür, so
daß sie einen Winkel von 60 bis 75 Grad mit dem Erdboden bildet;
dann nimmt er die andere Stange, dreht sie horizontal und drückt sie
mit aller Kraft auf die erste Stange hernieder. Dabei helfen ihm zwei
andere kräftige Pfeiler, die in einigem Abstand einwärts von den
Türpfosten stehen; sie sind das Widerlager für die Horizontalstange.
Der Verschluß ist absolut sicher; nur läßt er sich natürlich nicht bei
beiden Haustüren, der Vorder- und Hintertür, ausführen. Denn wie
sollte der Besitzer sonst in das Haus hineingelangen? Ich bin aber
einstweilen noch nicht über den Hintertürverschluß unterrichtet.
Der allgemein übliche Türverschluß.
Das ist also der Universalverschluß. Derjenige der Makonde von
Jumbe Chauro ist viel feiner, gediegener und origineller. Auch hier ist
in bezug auf die Tür alles wie sonst, nur steht an ihrer Innenseite ein
einzelner Pfahl etwa 15 Zentimeter von der Türkante ab frei im
Hüttenraum; in der Hüttenwand aber befindet sich an dieser Stelle
ein Loch, gerade groß genug, um den Arm hindurch zu stecken. Der
Tag ist heiß gewesen, nun aber will es Abend werden; still und
verlassen liegt das Makondedörfchen im dichten Busch verborgen
da. Es ist vollständig menschenleer, denn die ganze
Einwohnerschaft ist zu Schmaus und scharfem Umtrunk ins
Nachbardorf geladen. Jetzt nahen sich schlürfende Schritte, der
Hausvater und die Seinen kehren zurück; sie sind zum nicht
geringen Trost für alle diejenigen Blaßgesichter, die aller
Abstinenzbewegung zum Trotz noch immer gern ihr Gläschen Bier
genehmigen, in keiner anderen Verfassung, als man sich nach einer
schweren Sitzung auch bei uns befindet. Nur singt der Makonde
nicht; Africa non cantat; in dieser Beziehung wird er also nie ein
guter Deutscher werden! Jetzt kommt das Knifflige. Der Negerpapa
hat natürlich die Pflicht, das Haus zu öffnen. O ihr Göttinger
Semester, wie klar steht ihr wieder vor meiner Seele mit eurem
unermeßlichen Hausschlüssel, für den kein Schneider die Taschen
groß genug machen konnte, für dessen Last selbst die hintere
Hosenschnalle eigens verstärkt werden mußte! Größer als du,
unerläßliches Requisit froher Scholarenzeit, ist das Format auch
kaum, das der alte Neger aus irgendeinem Versteck zum Vorschein
bringt, nur Gestalt und Material sind anders als am Leinestrand.
Eisen ist eine plebejische Erfindung fremder, hergelaufener Völker,
der Makonde bleibt nach wie vor beim Holz; und wirklich geschickt
findet er sich damit ab: der Schlüssel ist ein etwa 30 Zentimeter
langer, etwas geschweift gearbeiteter Stab mit abgesetztem Griff;
der Bart hat drei derbe Zapfen, die gleich lang sind und in ein und
derselben Ebene liegen. Etwas unsicher — auch wieder à la
Göttingen — sucht Papa das Schlüsselloch, pardon, das Loch in der
Wand. Das hat er vermöge seiner Größe bald entdeckt. Mit nicht
unberechtigtem Stolz wagt er einen schüchternen Blick nach hinten,
wo Mama, den unvermeidlichen Sprößling im Rückentuch, geduldig
— sie ist ja heute mitschuldig — der Lösung des Problems harrt.
Diese ist wirklich nicht ganz leicht, die Pombe war gut und der Tag
war heiß; merkbar zittert die sonst so ruhige schwarze Hand, als der
alte Herr beginnt, die Öffnung für den Schlüssel zu suchen. Schon
hebt sich, einem Entenschnabel gleich, die siebenzentimetrige
Lippenscheibe der holden Gattin zu scharfer Aufmunterung, da ist
das große Werk endlich gelungen, das feine, rechteckige Loch in
dem freistehenden Pfeiler ist gefunden; rack, rack, rack, ein
dreimaliges rasches Heben und Senken des Schlüssels, schon zieht
er einen langen Riegel hinter sich her und quer durch den Pfeiler
hindurch. Dieser Riegel hatte mit seinem kolbenförmigen freien Ende
sich fest gegen die Innenseite der Tür gestemmt, sie dadurch
hermetisch verschließend; jetzt ist der kluge Neger gekommen, hat
mit seinem Schlüssel im Innern jenes Pfeilers in senkrechten Nuten
laufende Klötzchen gehoben und hat damit den Verschluß gelöst;
leicht und frei gleitet der Riegel zurück.

Türverschluß bei den Makonde von Jumbe Chauro.


Mit nicht geringem Selbstbewußtsein hat mir erst ein Hausvater
diese größte Erfindung des Hochlandes an Ort und Stelle
vordemonstriert, und dann ein anderer; beide Male habe ich ein
bewunderndes „Msuri sana, sehr schön!“ ausgerufen und den
Wunsch geäußert, diese Wunderdinge mit nach dem fernen Uleia zu
nehmen, um dort den Wasungu zu zeigen, was für tüchtige Kerle die
Makonde seien. Noch bin ich keine fünf Minuten in meinen Windfang
von Newala zurückgekehrt, da keucht es auch schon heran; im
selben Augenblick senken sich zwei stattliche Bäume vor meinen
Augen nieder, und feierlich, wie nach siegreicher Belagerung,
überreichen mir zwei stark schwitzende Gestalten die Schlüssel zum
Tor der gefallenen Feste. Zum Schlüssel gehört auch das Schloß,
hatten die beiden Kommandanten ganz logisch gedacht; ein Griff
nach der Axt, krachend fliegt das scharfe, dem trocknen, zähen
Tropenholz gegenüber jedoch zu weiche Eisen in die Basis des
schloßtragenden Pfeilers hinein. Den Pfahl aus dem Boden
auszugraben und ihn dergestalt intakt herbeizuschaffen, das war
den beiden Intelligenzen nicht eingefallen. So liegen die Stücke halb
zertrümmert vor mir, und statt einer Belobigung bekommen die
beiden Besitzer noch Schelte.
Auf der Suche nach dem Schlüsselloch.
Die Makuahütten sind in der Umgebung von Newala besonders
kümmerlich; in ihrer mehr als liederlichen Bauart erinnern sie mich
lebhaft an die Interimsbauten der Makua von Hatia; und dabei haben
die hiesigen Vertreter des Stammes durchaus keinen Krieg
mitgemacht. Es muß also wohl angeborene Faulheit sein, oder aber
das Fehlen einer straffen Häuptlingshand. Selbst die Barasa von
Mlipa, eine kleine Stunde südöstlich von Newala, nimmt an dieser
allgemeinen Verwahrlosung teil; während sonst die öffentlichen
Bauten hierzulande stets der Gegenstand einiger Sorgfalt sind, läuft
sie sichtlich Gefahr, vom ersten besten kräftigen Oststurm
umgewirbelt zu werden. Von einigem Reiz in dem ganzen weiten
Siedelungsdistrikt ist lediglich das Grab des verstorbenen Häuptlings
Mlipa selbst. Ich habe es in den ersten Vormittagsstunden einmal
besucht, wo noch brauende Nebel mit der durchbrechenden Sonne
kämpften; da sah der kreisrund angelegte Hain haushoher
Euphorbien, die nebst einem zerbrochenen Tongefäß allein noch von
der Ruhestätte des alten Negerkönigs zeugen, fast weihevoll aus.
Auch meine sonst so materiell und realistisch veranlagten Träger
mochten so etwas fühlen, denn sie sangen heute nicht ihre
gewohnten Schelmenlieder, sondern feierlich klang es, als wir von
dannen zogen, in den dichten, grünen Makondebusch hinein und
über ihn hinaus weithin den Steilabhang hinunter:
Lied anhören
MusicXML-Datei herunterladen
„Wir werden schon ankommen mit dem großen Herrn; wir stehen
in der Reihe und haben keine Angst, unser Essen und unser Geld
vom Serkal, der Regierung, zu bekommen. Wir sind nicht ängstlich;
wir gehen zusammen mit dem großen Herrn, dem Löwen, zur Küste
und kehren zurück.“
In bezug auf den Habitus der verschiedenen Stämme hier auf
dem Westrande des Plateaus komme ich zu keinem anderen
Ergebnis als dem bereits in der Ebene gewonnenen: es ist für einen
Nichtanthropologen unmöglich, es dem einzelnen direkt anzusehen,
welches Stammes er sei. Ich glaube aber, auch für den
Anthropologen von Fach möchte diese Unterscheidung schwer sein,
selbst auf Grund der peinlichsten Untersuchung; die ganze große
Völkergesellschaft hier im Osten des Erdteils, zwischen dem großen
zentralafrikanischen Graben, dem Tanganyika und dem Nyassa im
Westen und dem Indischen Ozean im Osten, ist nun einmal eng
miteinander verwandt; manche ihrer Sprachen unterscheiden sich
nur dialektisch; die Stämme werden zweifellos dieselbe
Schädelbildung und denselben Knochenbau besitzen; von
auffallenden äußeren Stammesunterschieden kann da unmöglich die
Rede sein.
Und selbst wenn sie beständen, hätte ich keine Zeit und Muße,
mich mit ihnen zu befassen, denn welch ungeheure Fülle von
ethnographischen Erscheinungen allein ist es, die Tag für Tag auf
mich einstürmt, die gesehen oder gehört, in beiden Fällen aber
begriffen, aufgezeichnet und niedergeschrieben sein will. Fast
könnte ich es als ein Glück bezeichnen, daß wenigstens einzelne
Forschungsgebiete durch äußere Umstände brachgelegt worden
sind. Da ist vor allem das Gebiet der Eisentechnik. Afrika gilt sonst
als ein Erdteil, wo der Eisenstein sozusagen auf der Straße liegt und
wo es verwunderlich erscheinen möchte, wenn seine Bewohner
nicht zur Verhüttung des überall anstehenden Materials gelangt
wären; tatsächlich reicht ja auch die Kenntnis des Eisenschmelzens
vom Nordrand bis zu den Kaffern.
Hier zwischen Rovuma und Lukuledi liegen die Verhältnisse nicht
so günstig. Raseneisenstein oder eine andere Eisenverbindung ist,
wie die Makonde erzählen, ihnen nicht bekannt; sie und ihre Vettern,
die Wamatambwe, sind demgemäß nicht bis zur Technik des
Eisenschmelzens fortgeschritten, sondern haben seit jeher ihre
Eisengeräte von den Nachbarstämmen kaufen müssen. Aber auch
den Bewohnern des Tieflandes ist es nicht leicht gemacht worden.
Nur ein einziger Fundi, ein alter Mann am Huwe, jenem
steilwandigen Granitklotz, der sich einsam mitten aus der weiten,
grünen Einöde zwischen Massassi und Chingulungulu erhebt und
dessen zackiges, wild zerklüftetes Haupt den Wanderer überall
grüßt, steht im Ruf, als letzter der Lebenden noch die Kunst des
Eisenschmelzens zu bewahren. Schon von Massassi aus wollte ich
den Mann in seiner Tätigkeit studieren; doch da hieß es: er ist aus
Angst vor dem Aufstand über den Rovuma gegangen; er kommt
indes bald wieder. Seitdem habe ich immer wieder gefragt: Ist er
denn nun endlich da, der Fundi? „Bado“, hieß es dann urecht
afrikanisch.

Gelbgießer beim Schmelzen des Messings.


Einen gelinden Trost hat mir dafür ein Gelbgießer gewährt, den
ich im Walde von Akundonde erwischte. Der Mann ist der Liebling

You might also like