1024 BOOK REVIEWS

Modules over endomorphism rings

(Encyclopedia of Mathematics and Its Applications 130)
By Theodore G. Faticoni: 372 pp., £70.00 (US$120.00), isbn 978-0-521-19960-5
(Cambridge University Press, Cambridge, UK, 2009).

Ce 2011 London Mathematical Society

doi:10.1112/blms/bdr064
Published online 7 July 2011

Theodore Faticoni enriches the mathematical literature with a comprehensive and up-to-date
account of the study of a large class of abelian groups (see [1]), presenting all the relevant
results in a very well organized fashion, most of which have been obtained by the author since
the 1970s [3, 5]. The so-called reduced torsion-free finite rank abelian groups, or rtffr groups
for short, have been intensively studied for the last half-century, and Faticoni’s contribution
is remarkable. His key technique consists of exploiting in a systematic way the relationship
between a group G, its endomorphism ring End(G), and the quotient ring of End(G) by its
nilradical. The book is aimed at an expert readership, or a very keen algebraist with a notable
background in the theory of abelian groups as presented in [1] for instance.
Hence, the mathematical content of the book is an in-depth analysis of torsion-free finitely
generated abelian groups, focusing on the rtffr groups. A torsion-free abelian group G is reduced
if Hom(Q, G) = {0}; that is, if G has no nontrivial divisible subgroup. In what follows, we let
R be a unital associative ring and G an rtffr group with endomorphism ring End(G), and such
that G is a right R-module. Write E(G) for the quotient of End(G) by its nilradical.
The author splits his approach into three strands: number theory of algebraic number
fields, module and ideal theory, and point set topology, with homological algebra firmly
set in the background. There are eighteen chapters in which Faticoni provides a panoply of
results.
The first chapter briefly introduces the main concepts used thereafter, and which the reader
needs to add to the notable assumed background on the theory of abelian groups. It includes an
Azumaya–Krull–Schmidt-type theorem about direct sum decomposition of G as an R-module
with some additional assumption. The motivation comes from a result by Arnold and Lady [2]
which gives an equivalence between projective right End(G)-modules and direct summands of
direct sums of G.
Next, Faticoni introduces algebraic number fields and links the class number of such a
field k to the power cancellation property of the rtffr groups G such that E(G) is the ring
of algebraic integers in k. The ring theoretic approach to rtffr groups continues with the
definition of the conductor of an rtffr group. The conductor of G is the ideal τ in the ring
E(G) defined as follows: let E be an integrally closed ring such that E(G) ⊆ E ⊆ QE(G) and
with E/E(G) finite. Then τ is the largest ideal in the centre of E(G) such that τ E ⊆ E(G).
The localisation E(G)τ of E(G) at τ leads to a bijection between P0 (G), the category of
finitely generated direct summands of direct sums of copies of the right R-module G, and
P0 (Eτ (G)), the category of finitely generated projective right Eτ (G)-modules, where Eτ (G) =
QE(G)(τ ) × E(G)τ and QE(G)(τ ) = {x ∈ QE(G) | xc = 0 for some c ∈ E(G)τ }. Further, G
splits into the direct sum G(τ ) ⊕ Gτ , where the direct sum decompositions of G(τ ) are
equivalent to those of finitely generated semi-simple QE(G)(τ )-modules, while direct sum
decompositions of Gτ are equivalent to those of finitely generated E(G)τ -modules. Other results
about the decomposability of G and E(G) are shown; in particular, G has a locally unique
decomposition if and only if the right E(G)τ -module E(G)τ also does. In Chapter 7, Faticoni
links his findings back to the class number of algebraic number fields as above. Chapter 8 is
an interlude, introducing and comparing certain families of groups: L-, J - and Eichler groups,
in particular.
The third viewpoint taken by the author starts in Chapter 9, a compilation of the results
taken from three publications including his AMS Memoir [4]. Here, various categories are
considered using arguments from homological algebra. The objective is to study and compare
these categories built up from an rtffr group, or another sort of abelian group, and prove that,
 BOOK REVIEWS 1025

under certain additional assumptions, some of these categories are equivalent to each other.
For instance, there is an equivalence between the categories of finitely generated projective
right EndR (G)-modules and P0 (G), the category defined above. Special attention is given to
the category of G-plexes. A G-plex is a complex of modules in P(G), the category of direct
summands of direct sums of copies of the right R-module G, such that the R-module G has a
certain ‘lifting property’.
The dual setting is handled in a similar fashion directly afterwards; in fact, in each of
the following chapters, a dual question naturally arises, which the author systematically
examines. In Chapter 10, quasi-projective modules and the tensor functor TG = − ⊗EndR (G) G
are introduced and used in the analysis of the chain of inclusion of categories

P(G) ⊆ G − Sol ⊆ G − Pres ⊆ G − Gen ⊆ G − Gen ⊆ Mod − R.

The objective is to obtain Morita-type equivalences that use G as bimodule. In Chapter 11, the
aim is to find relationships between G and EndR (G) using invariant theory, that is, homological
dimension. A neat glossary of terms pertaining to R-modules ends this chapter. In the following
two chapters, the author methodically investigates projective and finitely generated modules.
The sharper results are obtained for rtffr groups. Faticoni adds to his numerous ingredients
the centre of the quotient ring of R modulo its nilradical, and hence relates his latest findings
to results about direct sum decompositions of G. Chapter 14 attempts to classify the so-called
E-projective rtffr groups and looks at the minimal number of generators for these. Chapter 15
is concerned with the study of injective right EndR (G)-modules. From the analogous question
about projective modules, Faticoni reaches the expected conclusions, which he then uses to
improve his previous results on projective EndR (G)-modules.
Chapters 16 and 17 consist of an impressive diagrammatic treatment of the matter, with
the intention of determining as many connections as possible between the various categories
introduced so far. Techniques from homological algebra are exploited, and the outcome
comprises several correspondences and numerous statements of the form ‘the following are
equivalent’.
The last chapter takes a more traditional ring-theoretical viewpoint. Faticoni introduces the
concept of marginal isomorphisms, which he says ‘must be viewed as a very coarse measure
of a module’. Using this, he investigates marginal properties of G and obtains an Azumaya–
Krull–Schmidt-type theorem for those G which have an endomorphism ring EndR (G) with a
semi-primary classical right ring of quotients. The final part of the book is an extensive list of
results obtained by a minute and methodical analysis of margimorphisms.
An extensive bibliography will help fill possible theoretical gaps.
The terminology appears to be standard in this very specific area, and hence may lead to
some confusion with similar terms bearing other meanings in neighbouring fields which are
explored by the author in his investigations. For instance, the notion of quasi-isomorphism
defined in Chapter 18, though already used in Chapters 13 and 16, is not the same as that
used in the context of homological algebra.
This last observation raises one of the only two drawbacks of the book, in the reviewer’s
opinion: first, the potentially confusing notation and terminology, especially in the chapters
on categorical algebra and topology; secondly, a few inconsistencies in the presentation of the
material. Indeed, a concept may be defined long after it has been used, if defined at all, whereas
other notions are repeated where one least expects them (for instance, the omnipresent acronym
rtffr is amplified in Chapter 16).
These two features may perturb and somewhat hamper a perfectly smooth reading of the
text, but they would have been difficult to avoid in view of Faticoni’s objective: to present
his remarkable contribution to the study of a wide class of groups in a single publication. The
overall structure of the text slightly suffers from these irregularities; otherwise, the book is a
succession of very sharp results, which are the fruits of methodical and thorough investigations
started about 40 years ago by the author. Page after page, Faticoni engages the reader in a
friendly mathematical discussion. As incentive, he provides some historical motivations for the
questions he considers, for example, the classification of algebraic number fields, which goes
back to Gauss and Kummer, while numerous examples, exercises, and problems for future
1026 BOOK REVIEWS

research at the end of each chapter will certainly provide the reader with some stimulating
challenges.

References
1. D. Arnold, Finite rank torsion free Abelian groups and rings, Lecture Notes in Mathematics 931 (Springer,
New York, 1982).
2. D. Arnold and L. Lady, ‘Endomorphism rings and direct sum of torsion free abelian groups’, Trans. AMS
211 (1975) 225–237.
3. T. Faticoni, ‘On the lattice of right ideals of the endomorphism ring of an abelian group’, Bull. Aust. Math.
Soc. 38 (1988) 273–291.
4. T. Faticoni, ‘Categories of modules over endomorphism rings’, Memoirs AMS 492 (1993).
5. T. Faticoni, Direct sums decompositions of torsion-free finite rank groups (Chapman and Hall/CRC, Boca-
Raton, 2007).

Nadia Mazza
Lancaster University
Lancaster
LA1 4YF
United Kingdom
n·mazza@lancaster·ac·uk

Global aspects of ergodic group actions

(Mathematical Surveys and Monographs 160)
By Alexander S. Kechris: 237 pp., US$77.00, isbn 978-0-8218-4894-4
(American Mathematical Society, Providence, RI, 2010).

Ce 2011 London Mathematical Society

doi:10.1112/blms/bdr065
Published online 18 July 2011

In the last two decades, there has been a confluence of interests of ergodic and descriptive set
theorists, which has led to a widespread and highly active area of research, of which Kechris
is one of the leading proponents. Several different topics have contributed to this confluence;
on the one hand, originating in work of Mackey, the study of Borel transformations of Polish,
that is, separable completely metrisable, spaces in the absence of invariant measures has led to
substantial activity, as exposed in the delightful book ‘Basic ergodic theory’ by Nadkarni [6].
Though many of the basic concepts of this theory take their cues from classical ergodic theory,
for example, conjugacy, orbit and Kakutani equivalence, colouring properties, the subject is at
the same time distinctively different as it requires constructions to be done everywhere and not
just off of a set of measure zero. On the other hand, the theory of Borel reducibility of definable
equivalence relations on Polish spaces, stemming from a descriptive formulation of classification
problems originating in model theory and operator algebra, has been an extremely fertile area
of research equally oriented towards applications and internal theory development. Since, by
a fundamental theorem due to Feldman and Moore [2], the Borel equivalence relations with
countable classes coincide with those induced by Borel measurable actions of countable groups,
the study of this subclass is somewhat more tractable than the general theory and allows for
stronger tools coming from ergodic and geometric group theory. Again this has led many of
the techniques of orbit equivalence rigidity to be introduced into descriptive set theory; see,
for example [4].