Professional Documents
Culture Documents
Classical Hopf Algebras and Their Applications Algebra and Applications 29 Pierre Cartier Frederic Patras Online Ebook Texxtbook Full Chapter PDF
Classical Hopf Algebras and Their Applications Algebra and Applications 29 Pierre Cartier Frederic Patras Online Ebook Texxtbook Full Chapter PDF
https://ebookmeta.com/product/algebra-and-
applications-2-combinatorial-algebra-and-hopf-algebras-1st-
edition-makhlouf-abdenacer/
https://ebookmeta.com/product/algebra-and-applications-1-non-
associative-algebras-and-categories-1st-edition-makhlouf/
https://ebookmeta.com/product/chess-explained-
the-c3-sicilian-1st-edition-sam-collins/
https://ebookmeta.com/product/starting-out-the-c3-sicilian-1st-
edition-john-emms/
The Paper Issue 83 1st Edition Origamiusa
https://ebookmeta.com/product/the-paper-issue-83-1st-edition-
origamiusa/
https://ebookmeta.com/product/algebra-4-lie-algebras-chevalley-
groups-and-their-representations-1st-edition-ramji-lal/
https://ebookmeta.com/product/the-american-
revolution-1774-83-2nd-edition-daniel-marston/
https://ebookmeta.com/product/bnf-83-british-national-formulary-
march-2022-joint-formulary-committee/
https://ebookmeta.com/product/english-grammar-exercises-with-
answers-part-4-your-quest-towards-c2-1st-edition-daniel-b-smith/
Algebra and Applications
Pierre Cartier
Frédéric Patras
Classical Hopf
Algebras
and Their
Applications
Algebra and Applications
Volume 29
Series Editors
Michel Broué, Université Paris Diderot, Paris, France
Alice Fialowski, Eötvös Loránd University, Budapest, Hungary
Eric Friedlander, University of Southern California, Los Angeles, CA, USA
Iain Gordon, University of Edinburgh, Edinburgh, UK
John Greenlees, Warwick Mathematics Institute, University of Warwick,
Coventry, UK
Gerhard Hiß, Aachen University, Aachen, Germany
Ieke Moerdijk, Utrecht University, Nijmegen, Utrecht, The Netherlands
Christoph Schweigert, Hamburg University, Hamburg, Germany
Mina Teicher, Bar-Ilan University, Ramat-Gan, Israel
Algebra and Applications aims to publish well-written and carefully refereed
monographs with up-to-date expositions of research in all fields of algebra,
including its classical impact on commutative and noncommutative algebraic and
differential geometry, K-theory and algebraic topology, and further applications in
related domains, such as number theory, homotopy and (co)homology theory
through to discrete mathematics and mathematical physics.
Particular emphasis will be put on state-of-the-art topics such as rings of
differential operators, Lie algebras and super-algebras, group rings and algebras,
Kac-Moody theory, arithmetic algebraic geometry, Hopf algebras and quantum
groups, as well as their applications within mathematics and beyond. Books
dedicated to computational aspects of these topics will also be welcome.
Announcement (30 November 2020)
Alain Verschoren (1954-2020), Professor of Mathematics and Honorary Rector
of the University of Antwerp, became an editor of the Algebra and Applications
series in 2000. His contribution to the development of the series over two decades
was pivotal. We, the Springer mathematics editorial staff and the editors of the
series, mourn his passing and bear him in fond and grateful remembrance.
123
Pierre Cartier Frédéric Patras
Institut des Hautes Études Scientifiques Laboratoire J.A.Dieudonné
Bures-sur-Yvette, France Université Côte d’Azur
Nice, France
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
The present volume is dedicated to classical Hopf algebras and their applications.
By classical Hopf algebras, we mean Hopf algebras as they first appeared in the
works of Borel, Cartier, Hopf, and others in the 1940s and 50s: commutative or
cocommutative Hopf algebras. The purpose of the book is twofold. It first of all
offers a modern and systematic treatment of the structure theory of Hopf algebras,
using the approach of natural operations. According to it, the best way to under-
stand the structure of Hopf algebras is by means of their endomorphisms and their
combinatorics. We therefore put weight on notions such as pseudo-coproducts,
characteristic endomorphisms, descent algebras, or Lie idempotents, to quote a few.
We have included in this treatment the case of enveloping algebras of pre-Lie
algebras, extremely important in the recent literature, many interesting Lie algebras
actually being the Lie algebras obtained by antisymmetrization of a pre-Lie product.
Second, the book surveys important application fields, explaining how Hopf
algebras arise there, what problems they allow to address, and presenting the cor-
responding fundamental results. Each application field would require a textbook on
its own, we have therefore limited our exposition to introducing the main ideas and
accounting for the most fundamental results on which the use of Hopf algebras in
the field are grounded.
The book should thus be useful as a general introduction and reference on
classical Hopf algebras, their structure and endomorphisms; as a textbook for
Master 2 or doctoral-level programs; and mostly and ultimately to scholars in
algebra and the main application fields of Hopf algebras.
As for the book itself, it is the result of a long-lasting project. It originates
ultimately in 1989, when one of the authors initiated a Ph.D. under the direction
of the other. One of the ideas that emerged then was that the combinatorics of
dilations underlying the theory of finite integration as appearing in Hilbert’s third
problem (on equidecomposability of polytopes under pasting and gluing operations)
had far-reaching applications and generalizations. It extends, for example, to
properties of the direct sum of the symmetric group algebras or the study of power
maps on H-spaces. This lead to a purely combinatorial proof of structure theorems
v
vi Preface
for graded connected cocommutative Hopf algebras around which the content of the
central chapters of the first part of the book is organized.
At the time, the interest of the mathematical community for classical Hopf
algebras was limited. A certain number of classical tools and structure results were
available and were for the most part enough for the needs of applications, for
example, in rational homotopy—the subdomain of algebraic topology where tor-
sion phenomena are ignored. The situation evolved progressively, leading to the
writing of the present book that brings together classical results, some of which go
back to the 1950s, and recent advances under the unifying point of view of com-
binatorial structure results and techniques.
Many developments have contributed to the renewal of interest for classical
Hopf algebras. In algebraic combinatorics, the works of Ch. Reutenauer, J.-Y.
Thibon, and others generated again interest for the combinatorial theory of free Lie
algebras, Lie idempotents, (noncommutative) representation theory of symmetric
groups, and related objects. From the mid-1990s onward, the theme of combina-
torial Hopf algebras, whose first idea can be traced back to Rota, gained momentum
and grew steadily up to becoming one of the leading arguments of contemporary
algebraic combinatorics.
Another line of development has several independent origins: deformation the-
ory, differential calculus and differential geometry, numerical analysis and control,
theoretical physics... It relates to the notion of pre-Lie algebras and to Hopf algebras
of trees, forests, and diagrams. The notion of pre-Lie algebra dates back from the
early 1960s (Gerstenhaber, Vinberg) and can even be found earlier in the work of
Lazard. From the group and Lie theoretic point of view, which is also one of the
Hopf algebras, a key step in the development of the theory of pre-Lie algebras is
due to Agrachev and Gamkrelidze in the beginning of the 1980s. In hindsight, their
work started to develop the extension to pre-Lie algebras of the combinatorial
theory of Lie algebras and their enveloping algebras. However, the systematic
development of the theory is much more recent. The work of Connes and Kreimer
on Hopf algebras in perturbative quantum field theory around 2000 played here a
particularly important role. They featured the role of pre-Lie algebras of trees and
Feynman diagrams and their enveloping algebras in renormalization. Brouder
rapidly connected their insights with methods and results in numerical analysis.
Pre-Lie algebras and their enveloping (Hopf) algebras came to the forefront of
researches on Hopf algebras and their applications. The recent surge of Hopf
algebra techniques in stochastics (with rough paths, regularity structures) connects
to this line of development.
In algebraic topology, homological algebra and related areas, where the very
notion of Hopf algebra was born, the use of Hopf algebra techniques was classical
since the 1940s. Besides in the study of topological groups, they appear, for
example, in the study of loop spaces, algebras of operations such as Steenrod’s or
homology of Eilenberg–MacLane spaces. Pre-Lie algebras first appear in this
context with the work of Gerstenhaber. They relate to the more general idea of
brace operations that was introduced in the mid-1990s by Getzler, Gerstenhaber,
and Voronov in the context of cochain complexes and the theory of operads. Here,
Preface vii
the Hopf algebras at play have a particular structure: they are free or cofree as (co)
associative (co)algebras (free or cofree (co)commutative when arising from pre-Lie
algebras). This idea of Hopf algebras with extra structures proved also important, as
those structures carry with themselves the existence of additional properties and
operations.
Algebraic combinatorics and combinatorial Hopf algebras; algebraic topology,
homological algebra, and operadic structures; pre-Lie algebras together with their
many applications: we can give only very fragmentary indications about the many
developments that occured during the last 30 years and have deeply reshaped the
subject of classical Hopf algebras. We mention specifically these three lines of
thought since they motivated various choices made in the writing of this book. We
also wanted to point out with these examples the high level of activity surrounding
the subject of Hopf algebras, which appears over and over as a central topic in
contemporary mathematics.
Overall, the subject is too vast to be covered by a single textbook, we therefore
had to make choices. The book is structured into two parts: general theory and
applications. In the first part, we give a systematic account of the structure theory of
commutative or cocommutative Hopf algebras with emphasis put on enveloping
algebras of graded or complete Lie algebras and the dual polynomial Hopf algebras,
mostly over a field of characteristic 0. The second part is dedicated to several key
applications of the theory, classical, and recent. These application chapters can be
read separately, but we advise the reader seriously interested in using Hopf algebras
to read all of them as they offer complementary insights. Many techniques and
intuitions can actually be carried over from an application field to another.
It is impossible to acknowledge here all those who contributed along the years
by discussion, collaborations, and joint works to the building of the picture of Hopf
algebras and their applications addressed hereafter.
Frédéric Patras would like to thank especially Kurusch Ebrahimi-Fard together
with those others with whom he developed long-lasting research projects on the
topics addressed in this book; many of these projects have run over the last 20 years
and are still ongoing: Christian Brouder, Patrick Cassam-Chenaï, Loïc Foissy,
Joachim Kock, Claudia Malvenuto, Simon Malham, Dominique Manchon, Frédéric
Menous, Christophe Reutenauer, Nikolas Tapia, Jean-Yves Thibon, Anke Wiese,
and Lorenzo Zambotti. A special thought to a late friend, Manfred Schocker: we
had started together a vast program on Hopf algebras in combinatorics that was
interrupted by his premature death, the chapter dedicated to combinatorial Hopf
algebras is a tribute to his memory.
All linear structures are defined over a ground field k, excepted otherwise specified.
Categories are written with bold symbols, for example, Alg stands for the cat-
egory of algebras with a unit over k.
Algebras are algebras with unit, coalgebras are coalgebras with counit, excepted
otherwise specified. Ideals are two-sided, that is, simultaneously left and right
ideals, excepted otherwise specified.
We tend to abbreviate notations. For example, we will often write A for an
algebra, instead of the triple ðA; mA ; gA Þ, where mA and gA stand for the product and
the unit.
Symbols
When a symbol (e.g., Cogk ) is followed by “resp., . . .” (e.g., resp., Cog), this means
that the first symbol is the complete symbol associated to a notion, whereas the
following ones stand for abbreviations used to alleviate the notation when no
confusion can arise.
A þ : augmentation ideal of an augmented algebra A.
Abe: category of abelian groups.
Algk (resp., Alg): category of associative unital algebras.
AutC ðXÞ (resp., AutðXÞ): automorphisms of X in the category C.
Algck (resp., Algc ): category of complete augmented algebras.
cð1Þ . . . cðnÞ : (abbreviated) Sweedler notation for Dn ðcÞ.
C: complex numbers.
CðA; BÞ: set of morphisms in the category C from A to B (also denoted
HomC ðA; BÞ).
CðMÞ: subcoalgebra of a coalgebra C associated to a C-comodule M.
Cmd (resp., CmdC ): category of comodules over a coalgebra C.
Cogk (resp., Cog): category of coassociative counital coalgebras.
ix
x Conventions
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Linearization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Coalgebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Gebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Natural Endomorphisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.5 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.6 Structure of the Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
xiii
xiv Contents
Part II Applications
7 Group Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
7.1 Compact Lie Groups are Algebraic . . . . . . . . . . . . . . . . . . . . . 157
7.2 Algebraic Envelopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
7.3 Free Groups and Free Lie Algebras . . . . . . . . . . . . . . . . . . . . . 165
7.4 Tannaka Duality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
7.5 Bibliographical Indications . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
8 Algebraic Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
8.1 Homology of Groups and H-Spaces . . . . . . . . . . . . . . . . . . . . . 177
8.2 Hopf Algebras with Divided Powers . . . . . . . . . . . . . . . . . . . . 182
8.3 Eilenberg–MacLane Spaces and the Bar Construction . . . . . . . . 186
8.4 The Steenrod Hopf Algebra and Its Dual . . . . . . . . . . . . . . . . . 189
8.5 Bibliographical Indications . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
9 Combinatorial Hopf Algebras, Twisted Structures,
and Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
9.1 Vector Species and S–Modules . . . . . . . . . . . . . . . . . . . . . . . . 200
9.2 Hopf Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
9.3 The Hopf Species of Decorated Forests . . . . . . . . . . . . . . . . . . 204
9.4 Twisted Hopf Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
9.5 The Tensor Gebra as a Twisted Hopf Algebra . . . . . . . . . . . . . 211
9.6 From Twisted to Classical Hopf Algebras . . . . . . . . . . . . . . . . 212
9.7 The Gebra of Permutations . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
9.8 The Structure of Twisted Hopf Algebras . . . . . . . . . . . . . . . . . 220
9.9 Bibliographical Indications . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
10 Renormalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
10.1 Wick Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
10.2 Diagrammatics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
10.3 The Hopf Algebra of Feynman Graphs . . . . . . . . . . . . . . . . . . 234
10.4 Exponential Renormalization . . . . . . . . . . . . . . . . . . . . . . . . . . 237
10.5 Bibliographical Indications . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
1.1 Linearization
One of the key ideas underlying the theory of Hopf algebras, closely related to the
standard properties of the exponential and the logarithm, is linearization : that is,
making (non-linear) group-theoretical problems into linear ones. The correspondence
between groups and Lie algebras that can often be understood through a common
© Springer Nature Switzerland AG 2021 1
P. Cartier and F. Patras, Classical Hopf Algebras and Their Applications,
Algebra and Applications 29, https://doi.org/10.1007/978-3-030-77845-3_1
2 1 Introduction
embedding of both objects into a completion of the enveloping algebra of the Lie
algebra is one of its most striking and successful illustrations.
When Hopf algebras have extra structures, interesting new group-theoretical phe-
nomena arise. A nice example is provided by graded pre-Lie algebras and their
enveloping algebras: two exponentials relating the pre-Lie algebra to the associated
group can be defined in this context, and the study of their interactions leads to deep
formulas and properties. This leads, for example, to a refinement, in this setting, of the
usual combinatorial and free Lie algebra analysis of the Baker–Campbell–Hausdorff
problem (the computation of the logarithm of a product of exponentials).
In practice, it is often the case that algebras of functions provide the simplest way
to linearize an object, and this is actually what happens in most correspondences
existing between groups and Hopf algebras. For example, consider a finite group G.
The vector space k G := Set(G, k) is an algebra for the pointwise product
m : kG ⊗ kG → kG
: kG → kG ⊗ kG ∼
= k G×G ,
(λ)(g, h) := λ(gh),
so that
(λ) = λ(x y)δx ⊗ δ y , (1.1)
x,y∈G
1.2 Coalgebras
A second key idea is that coalgebras, preferably to algebras, are often the natural
framework to encode linearly group-type properties. This is already clear in the
example we have just considered: on the algebra of functions k G , the product does
not carry any interesting information (any set of functions with values in an algebra
is an algebra for the pointwise product). The coproduct instead encodes the product
rule on G. There are many technical reasons underlying this broad statement, and
some of them, related to duality phenomena, will become clear later in this book.
1.2 Coalgebras 3
A wit, due to Serre, accounts for a related feature of coalgebras: “there is a general
principle: every calculation relative to coalgebras is trivial and incomprehensible” (Il
y a un principe général : tout calcul relatif aux cogèbres est trivial et incompréhen-
sible).
1.3 Gebras
It often occurs that a given Hopf algebra structure goes along with other algebraic
structures or, more generally, that a given vector space can be equipped with several
interacting algebraic structures. To mention a few, associative, commutative, Lie,
pre-Lie, shuffle, quasi-shuffle algebra structures, and the dual coalgebraic notions
can coexist on a given space. This phenomenon can be found in many parts of the
recent literature on Hopf algebras, some of which are accounted for in this book.
The space of tensors T (V ) over a given vector space V will be one of the central
objects in this volume and is a good illustration of this phenomenon: it carries several
products, coproducts, is the free algebra over V for several algebraic structures,
can be viewed as a twisted Hopf algebra, and so on. Similar comments apply to
other fundamental objects such as the descent algebra (also known as the algebra
of noncommutative symmetric functions), the direct sum of the symmetric group
algebras or the vector spaces generated by trees, forests, and other classical families
of combinatorial objects.
It would be tedious to devise specific names to describe all the possible combina-
tions of structures existing on such objects. On the other hand, referring, for example,
to T (V ) as “the tensor algebra” is misleading in that it favors implicitly its free asso-
ciative algebra structure, not always the most interesting one. It is also actually very
convenient to view many vector spaces such as T (V ) as equipped with a family of
structures that can be extended progressively. We propose to use the word gebra
that was introduced by Serre as unifying the worlds of algebras and coalgebras, in
such situations. For example, the tensor gebra (over V ) will refer to T (V ) equipped
simultaneously with the various structures alluded to above and other ones that could
be defined. This allows, among others, to make unambiguous statements involving
several structures. Say, for example, we will see that the tensor gebra can be given a
free associative, but also a free commutative and a free commutative shuffle algebra
structure.
4 1 Introduction
A fourth key idea is, technically, less classical. It will be a central theme in this
book. It can be understood as a form of Galois theory: understand the structure
of a mathematical object through its automorphisms or endomorphisms. In many
situations, groups and gebras arise naturally from endomorphisms of objects, of
classes of objects, of endofunctors, or, more generally, of functors from one category
to another.
The most classical example is provided by linear representations. There is a gen-
eral philosophy that goes back to the early foundations of group theory, according
to which a group is little more than, and essentially given by, the collection of its
representations. Tannakian duality is one of its outsprings: it encodes the way in
which groups can be constructed from linear categories “behaving as categories of
representations,” from the knowledge of the forgetful functor from these categories
to vector spaces.
There are other, less known, illustrations of this general idea. Some are very
interesting for the systematic study of Hopf algebras and generalizations thereof.
The fundamental example here, which admits various variants, is the descent gebra
of graded connected cocommutative Hopf algebras. Among other properties, it is
the subalgebra of the convolution algebra of (natural) linear endomorphisms of this
family of Hopf algebras generated by projections on the graded components. It
is stable by the composition product of endomorphisms, and a Hopf algebra. Its
properties allow to recover Cartier’s structure theorem and many essential results of
the theory of free Lie algebras.
1.5 Applications
Last but not least, Hopf algebras are useful. They are very present in many application
fields, with a strong expansion during the last two decades. Thinking to applications
allows to enrich the classical set of tools and results and provides new insights on
the theory. Choosing among all application fields those that would be developed
here was a difficult task. We considered including chapters on several topics we had
independently studied from the Hopf algebraic point of view: algebraic groups over
the integers; quantum groups and their applications in Galois theory; symmetric
functions and their generalizations; numerical analysis and geometric integration;
stochastic integration, rough paths, and regularity structures; and classical and free
probabilities. But choices had to be done; the other topics we have chosen to develop
here should however give a good idea of the range of possible applications of the
theory.
1.5 Applications 5
We should point out that, contrary to the first part of the volume, where the treat-
ment is systematic, the four applications chapters on algebraic groups, algebraic
topology, combinatorial Hopf algebras, and renormalization are meant as introduc-
tions to the Hopf algebraic point of view, as each topic would deserve a volume on
its own.
Chapter 10 concludes the book with the Hopf algebraic approach to renormal-
ization. Various aspects are treated: Wick products, the Hopf algebraic construction
of Feynman diagrams, and Hopf algebras of Feynman diagrams. Renormalization
proper is presented following the general group-theoretical approach of the expo-
nential method. When amplitudes (concretely, functions on Feynmann diagrams)
take values in a Rota–Baxter algebra, the Bogoliubov recursion applies. Its Hopf
algebraic Birkhoff–Wiener–Hopf interpretation terminates the chapter.
An appendix is dedicated to the language and elementary notions of the theories
of categories and operads.
Each chapter concludes with separate bibliographical indications. They are aimed
at orienting the reader and suggest further or complementary readings on the topics
covered in this volume. They do not claim for exhaustivity or completeness, neither
contentwise nor historically. We usually point out at some of the works that origi-
nated the subject, the references we used, found most useful, and hint at works that
complement directly the account given in the chapter.
Part I
General Theory
Chapter 2
Coalgebras, Duality
I d ⊗-
m ∼
= ∼
=-
A⊗ A⊗ A A⊗ A k⊗A A A⊗k
m ⊗ Id ? m? η ⊗ Id ? Id ? Id ⊗ η
?
A⊗ A - A, A⊗ A - A A ⊗ A.
m m m
An associative algebra is always equipped with a canonical Lie bracket defined as the
commutator of the product: [x, y] := x y − yx. This construction defines a functor
from associative algebras to Lie algebras; we will study later its left adjoint, the
enveloping algebra functor, one of the most natural ways to construct Hopf algebras.
The notion of coalgebra is dual to the one of algebra. Categorically, this means that
a coalgebra is simply an algebra in Linop , the opposite category of the category of
vector spaces. Concretely, this means that coalgebras are defined diagrammatically
by inverting all the arrows in the diagrammatic definition of algebras, see below.
To make things precise, the dual notion of a bilinear product on a vector space A,
that is, a linear map m : A ⊗ A → A is the one of a coproduct, on a vector space C: a
linear map from C to C ⊗ C. It is often written C , or simply when no confusion
can arise
: C → C ⊗ C = C ⊗2 .
One should take care that this notation is really a shortcut for (c) since most often
it is not true that an element x of C ⊗ C can be written x1 ⊗ x2 with x1 , x2 ∈ C:
in general, it can be written only asa linear combination of such elementary tensor
products, as for example (λ) = λ(x y)δx ⊗ δ y in eq. (1.1). It is the reason why,
x,y
(1)
instead of c(1) ⊗ c(2) , various authors use the original Sweedler notation, cα ⊗
α
cα(2) . However, this last notation is less economic and, as with Einstein’s summation
conventions with repeated indices, the use of the abbreviated Sweedler notation does
not lead to problems provided one reminds that upper indices (i) refer to the tensor
expansion of a coproduct.
The coproduct is coassociative if the following identity holds between maps from
C to C ⊗3
( ⊗ I d) ◦ = (I d ⊗ ) ◦ . (2.1)
- C ⊗C
C
? ?Id ⊗
C ⊗C - C ⊗ C ⊗ C.
⊗ Id
n := ( ⊗ I d ⊗n−2 ) ◦ n−1 ,
n = (I d ⊗i ⊗ ⊗ I d ⊗n−i−2 ) ◦ n−1 .
n 1 +···+n k = (n 1 ⊗ · · · ⊗ n k ) ◦ k
We will denote this element of C ⊗3 by c(1) ⊗ c(2) ⊗ c(3) and more generally will
write
n (c) := c(1) ⊗ · · · ⊗ c(n) .
Although the notation is slightly ambiguous since c(1) ⊗ c(2) in the last equation
could be confused with (c), it does not create ambiguities in practice provided
some caution is taken in its use.
The coproduct is cocommutative if and only if
T ◦ = ,
and, more generally, it holds that for any permutation σ in Sn , the symmetric group
of order n,
c(1) ⊗ · · · ⊗ c(n) = c(σ (1)) ⊗ · · · ⊗ c(σ (n)) .
The dual notion of unit for an algebra is the one of counit: it is a linear map εC
(written simply ε when no confusion can arise) from C to k such that
(I d ⊗ ε) ◦ = I d = (ε ⊗ I d) ◦ , (2.2)
C ⊗C C - C ⊗C
ε ⊗ Id ? Id ? ?Id ⊗ ε
k⊗C ∼- C C ⊗k
= ∼
=
In the Sweedler notation, c = c(1) ε(c(2) ) = ε(c(1) )c(2) . We set C̄ := ker (ε).
Example 2.2.1 The ground field k is equipped with the structure of a (cocommuta-
tive) coalgebra by the identity map: = ε = I d : k → k = k ⊗ k.
Exercise 2.2.1 (Poset coalgebras) Let (X, ) be a locally finite poset. That is, given
x, y ∈ X with x y, there are finitely many z with x z y. Let C be the vector
space generated by ordered pairs (x, y) with x ≤ y. Check that
(x, y) := (x, z) ⊗ (z, y); ε(x, y) = δxy
x≤z≤y
The category of coalgebras is written Cog (or Cogk when the choice of a ground
field has to be emphasized). Later on, we will omit expliciting the definition of
morphisms in an algebraic category when it is a straightforward consequence of the
very definition of its objects (that is, when morphisms are the structure-preserving
maps).
2.2 Coalgebras: Definition and First Properties 13
Algebras and coalgebras are dual objects. The dual of a coalgebra is always an
algebra and the dual of a noncounital coalgebra is an algebra without a unit. However
in infinite dimension the definition of the dual coalgebra of an algebra requires some
care, as we shall see later.
λ ∗ β := m k ◦ (λ ⊗ β) ◦ C ,
The coproduct
εC⊗D := εC ⊗ ε D
equip the tensor product C ⊗ D with a structure of coalgebra. In the Sweedler nota-
tion,
C⊗D (c ⊗ d) = (c(1) ⊗ d (1) ) ⊗ (c(2) ⊗ d (2) ).
Exercise 2.2.2 Show that an algebra A is commutative if and only if the map m : A ⊗
A → A is a morphism of algebras. Dually, show that a coalgebra C is cocommutative
if and only if : C → C ⊗ C is a morphism of coalgebras.
14 2 Coalgebras, Duality
Cog(C, D) × Cog(C, E) ∼
= Cog(C, D ⊗ E),
- f ⊗g
-
C C ⊗C D ⊗ E.
The situation is dual to the case of commutative algebras, where A ⊗ B, the tensor
product of A and B equipped with the product law
(a ⊗ b) · (a ⊗ b ) = aa ⊗ bb ,
Com(A, C) × Com(B, C) ∼
= Com(A ⊗ B, C),
λ ∗ β(d) = λ(d (1) )β(d (2) ) = 0 = β(d (1) )λ(d (2) ) = β ∗ λ(d)
As for ideals, our coideals are two-sided. A right (resp., left) coideal is defined
by the condition (I ) ⊂ I ⊗ C (resp., (I ) ⊂ C ⊗ I ).
A subalgebra of an associative algebra A with product m and unit 1 is a subvector
space B of A, not necessarily strict, such that 1 ∈ B and m(B ⊗ B) ⊂ B. The notion
of subalgebra is dual to the one of coideal in the following sense. Assume that A = C ∗
and let J ⊂ A be the annulator of a coideal I of C. Then ε ∈ J is the unit of A and
for λ, β ∈ J , and d ∈ I , λ ∗ β(d) = λ(d (1) )β(d (2) ) = 0 since (d) = d (1) ⊗ d (2) ∈
I ⊗ C + C ⊗ I . Hence λ ∗ β ∈ J .
The dual algebra of k G identifies, through the pairing < λ|g >:= λ(g), with the
group algebra kG, the set of linear combinations of elements of G equipped with the
bilinear product extending the product of G. The pointwise product of functions in
k G dualizes accordingly into the diagonal map on kG,
kG (g) := g ⊗ g
that, together with the counit εkG (g) := 1, defines on kG the structure of a coalgebra.
Summarizing, we have on k G
(λ · β)(g) = λ(g)β(g), (δg ) = δh ⊗ δh ,
hh =g
(g) = g ⊗ g, g · h = gh.
The coalgebraic part of this construction holds more generally for all sets:
k S (s) := s ⊗ s, εk S (s) := 1
for s ∈ S equip k S, the linear span of an arbitrary set S, with the structure of a
coalgebra.
(c) = c ⊗ c.
Exercise 2.3.1 Show that group-like elements of a coalgebra C are linearly inde-
pendent over the ground field. Deduce that, for an arbitrary set S, (k S) = S.
Cog(k X, C) ∼
= Set(X, (C))
and that
Cog(k X, kY ) ∼
= Set(X, Y ).
< a|c >< b|c >=< ab|c >=< a|c(1) >< b|c(2) >,
¯
(x) := (x) − x ⊗ 1 − 1 ⊗ x,
2.3 Primitive and Group-Like Elements 17
¯ n : C̄ → C̄ ⊗n
¯ 2 := ,
is defined inductively by ¯
¯ n+1 = (
¯ ⊗ I dC⊗n−1 ) ◦
¯ n.
Notice that the condition (c) = c ⊗ 1 + 1 ⊗ c implies c ∈ C̄ as, from the counit
condition, c = ε(c)1 + ε(1)c = c + ε(c)1.
2.4 Tensors
Important examples of coalgebras, that will be useful in later chapters of this book,
are given by coalgebra structures on tensor spaces. Let V be a vector space, not
necessarily finite dimensional. We denote T n (V ) := V ⊗n the tensor product of n
copies of V for n ≥ 0, with T 0 (V ) = k. We write T (V ) for the direct sum
T (V ) := T n (V ),
n≥0
and call T (V ) the tensor gebra over V . We use the word notation and denote v1 . . . vn
the tensor product of elements v1 , . . . , vn in V . Elements of T n (V ) are called tensors
of length n.
n
d (v1 . . . vn ) := v1 . . . v p ⊗ v p+1 . . . vn ,
p=0
18 2 Coalgebras, Duality
and the counit ε(1) = 1, ε(v1 . . . vn ) = 0 for n ≥ 1, where the sum runs over par-
titions of [n] with I = {i 1 , . . . , i p }, i 1 < · · · < i p , J = { j1 , . . . , jn− p }, j1 < · · · <
jn− p , and v I := vi1 . . . vi p , is a cocommutative coalgebra called the unshuffle tensor
coalgebra over V .
The proof that these coproducts d and u indeed define coalgebra structures is left
to the reader.
Let us assume from now in this section that the ground field k is of characteristic
0. We denote T S n (V ) the vector subspace of T n (V ) of symmetric tensors of length
n, that is, tensors in T n (V ) invariant under the natural action of the symmetric group
Sn : that is, the action defined by, for σ ∈ Sn ,
(−1)n xσ (1) . . . xσ (n) = (−1)Car d(H ) (x H )n .
σ ∈Sn H ⊂[n]
n n
n
d (γn (v)) = γ p (v) ⊗ γn− p (v), u (γn (v)) = γ p (v) ⊗ γn− p (v).
p=0 p=0
p
The space of covariants T n (V )/Sn identifies with k[V ], the space of poly-
n≥0
nomials over V . Commutative monomials will be also written using the notation
v1 . . . vn , this should not create confusion as it will always be clear from the context
whether we are referring to commutative monomials or to words (noncommutative
monomials). Covariants and invariants are related by the linear isomorphism from
T S(V ) to k[V ] induced by the canonical projection from T (V ) to k[V ], and the
symmetrization map, from k[V ] to T S(V ):
v1 . . . vn −→ vσ (1) . . . vσ (n) .
σ ∈Sn
This map is also an isomorphism of vector spaces (recall we have assumed char (k) =
0). We refer to k[V ] as the polynomial gebra over V .
The polynomial gebra is equipped with a coalgebra structure by the polynomial
coproduct p
p (v1 . . . vn ) := vI ⊗ v J
I J =[n]
and the counit ε(1) = 1, ε(v1 . . . vn ) = 0 for n ≥ 1, where notations are as in the def-
inition of the unshuffle coproduct. We call it the polynomial coalgebra or coalgebra
of polynomials over V .
Notice that the polarization argument applies and we have in k[V ]
(−1)n
v1 . . . vn = (−1)Car d(H ) (v H )n
n! H ⊂[n]
and
n
n p
p (v ) =n
v ⊗ vn− p .
p=0
p
Exercise 2.4.1 Use the linear isomorphisms between T S(V ) and k[V ] described
above to compare the three coalgebras (T S(V ), u ), (k[V ], p ), and (T S(V ), d ).
2.5 Endomorphisms
(V ⊗ V ∗ ) ⊗ (V ⊗ V ∗ ) = V ⊗ (V ∗ ⊗ V ) ⊗ V ∗ → V ⊗ k ⊗ V ∗ ∼
= V ⊗ V ∗.
n
v∗ ⊗ w → v∗ ⊗ ei ⊗ ei∗ ⊗ w,
i=1
where (ei )i≤n stands for an arbitrary basis of V and (ei∗ )i≤n for the dual basis. The
counit is defined by ε(v∗ ⊗ w) :=< v∗ |w >. In the basis ci j := ei∗ ⊗ e j , (ci j ) =
cik ⊗ ck j . Notice that End ∨ (V ) = End(V ∗ ).
k
n
Notice also that since δV = ei ⊗ ei∗ identifies with the identity map of V ,
i=1
(v∗ ⊗ w) = v∗ ⊗ δV ⊗ w = v∗ ⊗ I dV ⊗ w.
n
(v ⊗ w∗ ) := (ei ⊗ w∗ ) ⊗ (v ⊗ ei∗ ),
i=1
n−1
an ⊗ bn = λi ai ⊗ bi ,
i=1
n−1
we would have 3 (c) = ai ⊗ bi ⊗ (ci + λi cn ), a contradiction.
i=1
Let C denote the linear span of the ai , bi , ci . Then, since the coproduct is coas-
sociative,
4 (c) = (ai ) ⊗ bi ⊗ ci = ai ⊗ (bi ) ⊗ ci = ai ⊗ bi ⊗ (ci ),
i i i
k
k
3 (c) = (x j ) ⊗ y j = x j ⊗ (y j ) ∈ C ⊗ C ⊗ C
j=1 j=1
we get that the (x j ) and the (y j ) belong to C ⊗ C . Finally, the linear span of c,
the x j , y j and the ai , bi , ci is a finite-dimensional subcoalgebra of C.
The same statement holds for left, or left and right translates.
Proof The very definition of representative functions implies that R y f can be written
as a linear combination of the f i . The lemma follows.
A representation π of X is a monoid homomorphism from X to the monoid of
linear endomorphisms of a finite–dimensional vector space V , π : X −→ End(V ).
Given a basis B = (ei )1≤i≤n of V , we write πi, j (g), 1 ≤ i, j ≤ n for the entries of
the matrix of π(g) in B. Since π(x y) = π(x)π(y),
πi, j (x y) = πi,k (x)πk, j (y) (2.4)
1≤k≤n
The same statement holds for left, or left and right translates.
Proof Let V be the space of right translates of f . The map y −→ R y defines a
finite-dimensional representation of X on V . Define δ : V → k by δ(g) := g(1).
Then, f (y) =< δ, R y ( f ) >: f ∈ R(X ) and f is representative.
Corollary 2.7.1 The functions f i , f i in eq. (2.3) can be chosen to be representative.
The coproduct
(πi, j ) := πi,k ⊗ πk, j
1≤k≤n
and the counit (πi, j ) := πi, j (1) are defined on any space of coefficients C(π ),
hence on R(X ) = C(π ). They equip R(X ) with the structure of a coassociative
π
coalgebra.
Proof Only the coassociativity assertion requires a proof; it follows from the identity
πi, j (x yz) = πi,k (x)πk,l (y)πl, j (z).
k,l≤n
Example 2.7.1 Matrix elements of representations are clearly, in view of Prop. 2.7.1,
canonical examples of representative functions. Another canonical example is pro-
vided by finite monoids. In that case, in view of Formula (1.1), that applies to monoids,
all functions on X are representative and R(X ) = k X . This may be understood from
the representation theoretical point of view by considering the regular representation
of X , that is, the action by left translations on k X , the linear span of X . The delta
functions δx () generate linearly k X , and
δx (yz) = δa (y)δb (z).
ab=x
2.8 Comodules
δ : M → M ⊗C
such that
(I d M ⊗ ) ◦ δ = (δ ⊗ I dC ) ◦ δ
24 2 Coalgebras, Duality
and
I d M = (I d M ⊗ ε) ◦ δ.
Graphically, the first identity translates into the commutativity of the diagram:
δ - M ⊗C
M
δ? I dM ⊗ ?
M ⊗C - M ⊗ C ⊗ C.
δ ⊗ I dC
Left comodules are defined similarly, but since we will use only right ones, we
call the latter simply comodules. The notion of subcomodule is the canonical one.
The category of (resp., finite dimensional) comodules over C is denoted Cmd
(resp., Cmd f ) or when the underlying coalgebra has to be explicited CmdC (resp.,
f
CmdC ).
Recall from Def. 2.5.1 that for an arbitrary finite-dimensional vector space V ,
the endomorphism coalgebra Enc(V ) = V ⊗ V ∗ is equipped with the coproduct
n
(v ⊗ w∗ ) := ei ⊗ w∗ ⊗ v ⊗ ei∗ , where (ei )1≤i≤n stands for an arbitrary basis
i=1
of V and (ei∗ )1≤i≤n for the dual basis. The vector space V is then equipped with a
comodule structure over Enc(V ) by
V → V ⊗ Enc(V )
n
v → ∂(v) := ei ⊗ (v ⊗ ei∗ ).
i=1
Lemma 2.8.1 The coalgebra Enc(V ) is universal in the sense that it is equivalent
to define a C-comodule structure on V or a coalgebra map γC from Enc(V ) to C.
Given ∂C : V → V ⊗ C,
n
j
∂C (e j ) = ei ⊗ ci ,
i=1
the image of the basis element e j ⊗ ei∗ of Enc(V ) in C by this coalgebra map is ci .
j
∂C - V ⊗C
V
∂C ? I dV ⊗ C ?
V ⊗C - V ⊗C ⊗C
∂C ⊗ I dC
n
n
n
n
ei ⊗ (cij ) = ∂C (ek ) ⊗ ckj = ei ⊗ cki ⊗ ckj .
i=1 k=1 i=1 k=1
n
(cij ) = cki ⊗ ckj .
k=1
n
n
γC ⊗ γC (ek ⊗ ei∗ ⊗ e j ⊗ ek∗ ) = cki ⊗ ckj
k=1 k=1
γC - C
V ⊗ V∗
Enc(C) ? C ?
V ⊗ V∗ ⊗ V ⊗ V∗ - C ⊗ C.
γC ⊗ γC
We also have
εC (cij ) = δi = ε Enc(V ) (e j ⊗ ei∗ ).
j
It is interesting, and useful for various purposes, to understand how the notion of
subcoalgebras is reflected in terms of comodules. If D is a subcoalgebra of C, any
f
D-comodule is a C-comodule since M ⊗ D → M ⊗ C, and Cmd D is isomorphic
f
to a subcategory D̃ of CmdC .
26 2 Coalgebras, Duality
Proof For E as in the proposition, let us write C(E) for the coalgebra sum of the
coalgebras C(M) for M ∈ E. We have C( D̃) = D for any subcoalgebra D of C:
obviously C( D̃) ⊂ D, and the identity follows by noticing that D is the union of its
finite-dimensional subcoalgebras.
= E. Clearly, E ⊂ C(E).
Let us show that C(E) It is therefore enough to show that,
given a finite-dimensional C-comodule M in E and F a finite-dimensional C(M)-
comodule, then there exists a n ≥ 0 such that F is isomorphic to a subcomodule
of a quotient of M n . Since the structure map δ : F → F ⊗ C(M) is an injective
morphism of comodules and F ⊗ C(M) ∼ = C(M)dim(F) , it is enough to show the
property for F = C(M). Let us fix a basis (m i )i∈I of M. We get a sequence of
C(M)-comodule maps:
⊕i∈I m i∗ ⊗C(M)
M dim(M) ∼
⊕i∈I
-δ - C(M).
= M M ⊗ C(M)
i∈I i∈I
The total map from M dim(M) ∼= M ⊗ M ∗ to C(M) is surjective. This follows from
the fact that C(M) is the image of the map from M ⊗ M ∗ = Enc(M) to C(M)
induced by δ. The proposition follows.
Let now V be a vector space and G a group. We call V -valued representative functions
the elements of V G (maps from G to V ) whose left (or right) translates generate a
finite-dimensional vector space. The set of V -valued representative functions of G
is written R(G, V ) (so that R(G, k) = R(G)). Clearly, R(G, V ) ∼ = V ⊗ R(G).
If furthermore V is a locally finite G-module (every v ∈ V is contained in a finite-
dimensional sub-G-module of V ) with structure map π : G → End(V ), any v ∈ V
defines an element written δ(v) of R(G, V ) by
δ(v)(g) := π(g)(v).
In coordinates:
ek ⊗ (πk j ⊗ π ji ) = δ(e j ) ⊗ π ji .
j,k≤n j≤n
Equivalently,
(I dV ⊗ ) ◦ δ = (δ ⊗ I d R(G) ) ◦ δ.
Corollary 2.9.1 The map δ : V → V ⊗ R(G) just defined equips any locally finite
G-module V with the structure of a comodule over R(G).
Proof We have seen that any locally finite G-module is naturally equipped with the
structure of a comodule over R(G). Conversely, let δ : V → V ⊗ R(G) define a
R(G)-comodule structure on V . Then, the map
We leave the reader check that these constructions are inverse to each other.
Although we featured the point of view of groups, the same arguments and results
hold more generally for locally finite A-modules V over an associative algebra A,
see Sect. 2.12. Moreover, we have the following.
Lemma 2.9.1 Under the previous isomorphism the tensor product of two locally
finite G-modules, V ⊗ W (with structure map g(v ⊗ w) := g(v) ⊗ g(w)), is reflected
in the following comodule structure:
V ⊗ W → (V ⊗ R(G)) ⊗ (W ⊗ R(G)) ∼
= (V ⊗ W ) ⊗ (R(G) ⊗ R(G)) → (V ⊗ W ) ⊗ R(G),
where the first map is the tensor product of the comodule structure maps for V and
W , the second is induced by the switch map R(G) ⊗ W ∼ = W ⊗ R(G) and the last
by the product map on R(G).
We used and will use for pseudo-coproducts the shortcut notation F = f F(1) ⊗
f F(2) . As for the Sweedler notation, this notation does not mean that F can be written
as the tensor product of two endomorphisms in End(A) since, in general, F can
be written only as a linear combination of such tensor products (see the examples
below).
Notice that, in general, even when A is a free (associative or commutative) algebra,
an element of End(A) may admit several pseudo-coproducts. For example, when
A = C[x], equipped with the usual augmentation map (the evaluation at 0 of a
polynomial), let f be defined by f (x i ) := x i if i ≥ 2 and f (x) = f (1) = 0. Then,
for the maps g, h, k defined by g(x i ) := x i+1 for i = 0, g(1) := 0; h(x i ) := x i−1 for
i = 0, h(1) := 0, l(x i ) := x i if i ≥ 1 and l(1) = 0, the two tensor products g ⊗ h +
f ⊗ ν + ν ⊗ f and l ⊗ l + f ⊗ ν + ν ⊗ f are pseudo-coproducts for f . Another
example is provided by polynomial derivations. For example, x n ∂x admits various
coproducts such as x n ∂x ⊗ I d + I d ⊗ x n ∂x , ∂x ⊗ x n I d + x n I d ⊗ ∂x .
We will see that the choice of a particular pseudo-coproduct for a map f most
often does not matter in practice. In many applications, it will also happen that there
is a well-defined natural pseudo-coproduct for f . We will therefore often abbrevi-
ate f F(1) ⊗ f F(2) to f (1) ⊗ f (2) excepted when we want to emphasize explicitly the
dependency of the pseudo-coproduct of f on a particular choice for F.
Proof Let F = f (1) ⊗ f (2) and G = g (1) ⊗ g (2) . For x, y ∈ A, we have f ◦ g(x y) =
f (g (1) (x)g (2) (y)) = f (1) (g (1) (x)) f (2) (g (2) (y)), so that F ◦ G = f (1) ◦ g (1) ⊗ f (2) ◦
g (2) is a pseudo-coproduct for f ◦ g. The other assertions are straightforward.
Corollary 2.10.1 The set of group-like elements in End(A), written EndAlg (A), is
a monoid for the composition product.
Proof Indeed, the composition of algebra endomorphisms is an algebra endomor-
phism. Using pseudo-coproducts: let f, g ∈ EndAlg (A), then f ⊗ f and g ⊗ g are
pseudo-coproducts for f and g, so that ( f ◦ g) ⊗ ( f ◦ g) is a pseudo-coproduct for
f ◦ g. The unit of the monoid is the identity map of A.
Take care that in spite of its name, the monoid of group-like elements EndAlg (A)
is not a group for the composition product in general: for example, when A is aug-
mented, the map ν belongs to EndAlg (A) but, for any f ∈ EndAlg (A), f ◦ ν = ν.
Proof Recall that any associative algebra with product ∗ is equipped naturally with
a Lie algebra structure and a Lie bracket usually written [ , ] (or [ , ]∗ , to emphasize
what is the underlying associative product, defined by [x, y] := x ∗ y − y ∗ x.
The first part of the corollary simply states the property that derivations of an
algebra form a Lie algebra. The use of pseudo-coproducts allows to give element-
free proofs of this kind of structural statements.
Let f, g be derivations in End(A), a pseudo-coproduct for [ f, g] := f ◦ g − g ◦
f is given by
( f ⊗ I d + I d ⊗ f ) ◦ (g ⊗ I d + I d ⊗ g) − (g ⊗ I d + I d ⊗ g) ◦ ( f ⊗ I d + I d ⊗ f )
= f ◦ g ⊗ Id + Id ⊗ f ◦ g − g ◦ f ⊗ Id − Id ⊗ g ◦ f
= [ f, g] ⊗ I d + I d ⊗ [ f, g].
( f ⊗ ν + ν ⊗ f ) ◦ (g ⊗ ν + ν ⊗ g) − (g ⊗ ν + ν ⊗ g) ◦ ( f ⊗ ν + ν ⊗ f )
= f ◦g⊗ν+ν⊗ f ◦g−g◦ f ⊗ν−ν⊗g◦ f
= [ f, g] ⊗ ν + ν ⊗ [ f, g].
Example 2.10.3 For example, when A = C[x1 , ..., xn ], the derivations, which are
the polynomial vector fields, form a Lie algebra.
Theorem 2.10.2 The logarithm and the exponential series define inverse bijections
1
between EndAlg (A) and Der 0 (A).
(−1)n−1
( f ⊗ f − I d ⊗ I d)n = log( f ⊗ f ) = log(( f ⊗ I d) ◦ (I d ⊗ f ))
n≥1
n
Let us show now briefly how the previous constructions dualize. We omit some details
since most ideas are similar to the case of algebra endomorphisms. Let (C, , ε) be a
coalgebra over a field k of characteristic 0, possibly but not necessarily coaugmented.
When this is the case, we denote by η : k → C the coaugmentation map. We recall
that C̄ is the kernel of ε and we write 1 for η(1). As usual we set ν = η ◦ ε and we
have ε ◦ η = I dk .
C - C ⊗C
f ? F
?
C - C ⊗ C.
We will use for quasi-coproducts the shortcut notation F = f F(1) ⊗ f F(2) . An ele-
ment of End(C) may admit several quasi-coproducts.
A coderivation of a coalgebra is an endomorphism f such that ( f ⊗ I d + I d ⊗
f ) ◦ = ◦ f . Therefore, we have
Lemma 2.11.1 The map f is primitive if and only if f (1) = 0 and it maps the
coaugmented coalgebra C to Prim(C), the space of primitive elements in C.
Proof Indeed, for any c ∈ C̄, any such f , and since is counital:
Remark 2.11.2 The set of group-like elements in End(C), written EndCog (C), is
a monoid for the composition product: a composite of coalgebra endomorphisms is
a coalgebra endomorphism.
Van Vleuten had het in zijn betrekking hoe langer hoe drukker
gekregen. En met de drukte waren zijn verdiensten zoodanig
gestegen, dat zij de gewone behoeften van het huishouden
ruimschoots dekten. Mocht dit hem eenigszins nonchalant hebben
gemaakt ten aanzien van de door Wiechen te betalen rente, Betsy
vond, dat het niet aanging dien man maar geheel te laten betuilen,
en telkens later en later die rente te ontvangen.
Inderdaad was hij zoo langzamerhand ruim drie maanden ten achter
geraakt, wat het stipte vrouwtje in het geheel niet beviel. Zij kende
de praatjes waarmee haar man zich liet afschepen. Men had van
den beginne af een speling wat den tijd betrof, in het oog gevat, en
bij hooge rente als deze, moest men wat weten te geven en te
nemen. Opperbest, vond Betsy, doch die speling, waarop men
gerekend had, was gefixeerd op vijf dagen, dus de vijfde dag
behoorde in elk geval de dag te zijn, waarop men recht had over de
rente te beschikken. Wat onder „geven en nemen” te verstaan was,
verklaarde zij niet te begrijpen, althans niet in verband met de hooge
rente. In geen geval mocht dat beteekenen, dat men zich een
vermindering door systematisch veroorzaakten achterstal behoefde
te laten welgevallen. [137]
Van Vleuten kon daartegen weinig aanvoeren. Van zijn standpunt uit
maakte hij nog een mooie rente, zelfs al werd daar eens wat op
beknibbeld. Maar hiermee behoefde hij bij Betsy niet aan te komen.
Zij toch had een anderen kijk op de zaak. Van al hetgeen men boven
normale rente kreeg, moest het kapitaal weer bijeengespaard
worden, dat zij verloren waande. Men mocht zich dus door Wiechen
niet laten bedotten. En of hij haar inzicht nu deelde of niet, dat kwam
er nu niet meer op aan, sedert zij zijn belofte had.
Op zekeren dag nam zij een kloek besluit. Als van Vleuten er niet
voor zorgen wilde, of geen tijd had, zou zij er zich eens krachtig mee
gaan bemoeien. En toen hij in den ochtend op reis was gegaan,
stapte zij moedig op de tram, en begaf zich naar Wiechens kantoor.
Hem thuis vindende, vroeg zij kortweg om betaling der achterstallige
rente.
Als een kikker danste Wiechen om haar stoel heen, nu hier, dan
daar een boek of een papier grijpende, tot hij eindelijk zijn besluit
genomen had.
„Ja, ziet u, mevrouwtje,” zeide hij; „ik zou die zaak liever met meneer
behandelen.”
„Absoluut onnoodig,” meende zij. „Hier valt niets anders te doen dan
te betalen …”
„Hm.… Ik blijf erbij, dat ik de zaak liever met meneer behandel. Het
is.… och, mannen willen niet altijd voor hun vrouwen weten, wat zij
uitgeven, en waaraan.… U begrijpt, mevrouwtje.…”
Bij de laatste woorden had hij zijn hand als sussend op haar arm
gelegd. Betsy vloog van haar stoel op, rillend van walging en
verontwaardiging. [138]
Betsy was hevig geschrikt. Zich omdraaiend, verliet zij haastig het
kantoor, in de tusschenkamer den bediende naar binnen wijzende.
Doch deze begreep haar niet, en de deur openende, liet hij haar uit.
En met opgericht hoofd week zij terzijde, en liep door. Een eind
verder had ze er spijt van. Dat was de tweede domheid vandaag! Als
zij nu Marie niet had afgestooten, zou deze haar wellicht hebben
kunnen helpen.… maar neen, liever zonder, dan met die hulp. Toch
was het dom. Ze dacht erover, welke daad zij nu nog kon doen, om
het getal van drie vol te maken. Bah, wat was ze onhandig geweest!
[139]
Maar één gevolg moest haar gang van heden toch hebben. Het was
nu guerre déclarée, wat allicht beter was, dan dat gezeur. En dan
moest ze maar ineens zorgen, dat de knoop werd doorgehakt. Nu in
godsnaam niet halverwege blijven staan!
Wacht, ze wist het! Bij de halte stapte zij op de tram, moe van het
loopen zoo ver, en reed, den omweg over het Plein nemende, naar
huis. Zonder zich van haar hoed en mantel te ontdoen, snelde zij
naar de slaapkamer, waar het trommeltje stond, waarin het bewijs
van Wiechen. Gehaast, als vreesde zij in haar voornemen gehinderd
te zullen worden, kreeg zij het stuk, en verliet het huis weer.
„De acte van deposito. Kijkt u nu zelf nog eens na, of u daarin iets
vinden kunt over de verschuldigde rente.”
„Juist mevrouw, u begrijpt den toeleg, zie ik. Konden wij aantoonen,
dat hij zich tot een zoodanige rentebetaling verbonden had, als wel
het geval zal zijn, doch dat niet te bewijzen is, dan zouden we hem
bij exploit in gebreke kunnen stellen, en alles, hoofdsom en rente,
opvorderen. Doen wij dat nu, dan ontkent hij eenvoudig, of begroet
ons met een rekening-courant, waarop heel netjes zijn betalingen in
mindering van het deposito gebracht zijn, terwijl hij een rente-
vergoeding van drie percent heeft bijgeschreven. Dat is zoowat de
usantieele rente, zooals u in haast elke courant kunt zien, bij
kapitaal, dat met een halfjaar opzegbaar is.”
„Begint u daar nooit mee!” viel de advocaat in. „Het geval heeft zich
een paar jaar geleden eens voorgedaan, dat iemand een eed niet
wilde afleggen. En ik herinner mij, dat het geval eenige sensatie
maakte. De oudste rechter in onze Rechtbank zeide mij toen, dat dit
de eerste maal in zijn leven was, dat een opgedragen eed niet werd
afgelegd. Dat wil nog al wat zeggen dunkt me.”
„Ik zal mij wel wachten dat te beweren.” zeide hij. „De les, die er uit
te trekken valt, is, dat men beter doet niet te procedeeren, als men
tevoren weet, dat de zaak van een eed afhangt. En nu wat uw zaak
betreft, meen ik geen ander advies te mogen geven, dan het
deposito te laten opzeggen, tegen heden over zes maanden.”
„Maar wat moeten wij dan al dien tijd doen? Hij zal natuurlijk
weigeren iets te betalen, als we hebben opgezegd.”
„De rechter krijgt de zaak niet onder de oogen, dan even vóór het
pleidooi, en dan nog maar gedeeltelijk. Maar het zou te lang voeren
u dat nu uit te leggen. Ik bedoelde met mijn vraag, of meneer van
Vleuten om het geld verlegen is. Ik meende laatst gehoord te
hebben, dat hij nog andere ressources had.”
„We hebben weinig omgang,” zeide Betsy. „Maar … hij kent hier een
zekeren Boom, dien hij met geld heeft bijgestaan.”
„Tweeduizend gulden.”
„Zoo zeker als iets. Van de geheele bende acht ik dien Boom de
gevaarlijkste. Weet u wat, laat meneer van Vleuten eens een dag
vrijmaken en bij mij komen. Ik zal dan eens onderzoeken hoever hij
de dupe is van dat volk, en zien op de voordeeligste manier voor
hem te liquideeren.” [142]
„Zeker, mevrouw, dat doet hij ook. Maar oordeelt u eens zelf. Hij
heeft enkele jaren geleden een reusachtig faillissement geslagen,
dat nog niet is afgeloopen. Gewerkt heeft hij sedert nooit. Zijn familie
kan weinig of niets aan hem doen. Hij is altijd even netjes gekleed,
en als hij niet slaapt, zit hij in een koffiehuis. Zuinig is hij niet eens.
En waarvan moet die man nu leven, als het niet is van oplichterijtjes,
meent u?”
„Zijn het dat allemaal smeerlappen, in dit land?” riep zij eindelijk uit.
„O meneer, helpt u ons uit hun handen! Daar zijn wij niet tegen
opgewassen. Ik zal mijn man dadelijk bij u sturen. Wanneer wilt u dat
hij komt?”
„Niets,” zeide hij, het boek dichtslaande. „Ik bladerde een beetje in
een bijbel, dien ik toevallig inhanden kreeg. Waar ben jij heen
geweest?”
„Je hebt ergens veel plezier van,” zeide Betsy, hem oplettend
aanziend. „Wat is het?”
„Dat zal ik je bij gelegenheid wel eens vertellen,” beloofde hij. „Ik
geloof, dat ik wat gevonden heb, maar.… enfin, eerst moet ik zeker
zijn. Waar ben jij intusschen heen geweest.”
„Ben je dol?”
„Nog niet, maar ik ben het bijna geworden. Jan, ik heb beloofd, dat je
vanavond bij den advocaat zou zijn. Beloof je me er heen te zullen
gaan? Als je niet dadelijk handelt, zijn we alles kwijt. O, ik heb het
altijd wel gevoeld.”
Het was erg genoeg, maar had nog erger kunnen zijn. Want, daar
zijn solvabiliteit althans tot nog toe onbesproken was, zouden zij er
af kunnen komen met opoffering van de rente, die zij toch eigenlijk
bovenmatig hadden genoten, en een beetje leergeld toe.
Zoo troostten zij zich, en onmiddellijk na het eten begaf zich Van
Vleuten op weg zijn advocaat.
Meer zakelijk dan hij tegenover Betsy had willen doen, zette deze
voor Van Vleuten zijn zienswijze uiteen. Vooropstellende, dat men
enkel gissen kon, meende hij echter als een bijzonder veeg teeken
te moeten aanmerken, dat Wiechen in den laatsten tijd geen verweer
meer deed voeren, wanneer men hem in rechte betrok. In geen
zijner bij de Rechtbank aanhangige zaken verscheen voor hem een
procureur. De tegenpartij had dus vrij spel. In het eerst had dit niet
zoozeer de aandacht getrokken, daar men wist, dat de advocaat die
vroeger voor zijn zaken opkwam, hem [144]had bedankt, en een
ander, die sedert zijn zaken had overgenomen, nu juist niet als de
handigste in procureurswerkzaamheden bekend stond, zoodat men
aan vergissingen of verzuimen geloofde, die zich wel zouden
herstellen. Doch toen kort daarna was gebleken, dat ook deze zich
aan de zaken van Wiechen had onttrokken, en voor dezen in een
nieuwe zaak zich niemand procureur stelde, werd het duidelijk dat
daar meer achter school, en Wiechen zich eenvoudig alles liet
aanleunen, wat men in rechte tegen hem op touw wilde zetten.
Daarvoor viel geen andere verklaring te vinden, dan dat er bij hem
niets meer te halen viel. En hoewel deze indruk vrij plotseling kwam,
vertoonde zich toch het gewone verschijnsel, dat de eene crediteur
vóór, de andere na, aarzelde de kat den bel aan binden, hetzij door
een executie te ondernemen, hetzij door ineens faillissement aan te
vragen.
„Jawel, u spreekt als een rechtsgeleerde, die zich tevreden stelt met
een straf. Wij, menschen van zaken, verlangen ons geld, en doen u
de straf gaarne cadeau.”
Uit den eeuwigen strijd van belangen kwam toch een norm voort, te
alle tijden, en wie zich daarbuiten waagde, moest er op rekenen òf
een buitengewoon succes te zullen hebben, òf een buitengewone
klap te zullen krijgen. Dat had hij toch in Indië ook geweten, zij het
dat zich de dingen daar langzamer ontwikkelden.
Met waren galgenhumor bekeek hem van Vleuten een poos, zooals
hij daar zat. Werkelijk, op een afstand, wanneer hij zich niet bewoog,
en men de kleine incorrectheden in zijn spreken niet hoorde, was
hem een zekere distinctie niet te ontzeggen. Een heldere teint, en
dat licht sensueele in de oogen, dat men anders slechts bij buiten
echt geborenen te vinden pleegt.
„Ja,” meende Boom, „een mensch kan maar één ding tegelijk doen.
Ik heb het met de zaken zóó druk gehad …”
Van Vleuten keek den ander aan. Was hij dronken? Dat scheen
overigens toch niet.
„Nu ja, maar zoolang ik daar niet aan beginnen kon, zouden we het
geld gezamenlijk uitzetten.”
„Zeker. Dat was, naar ik begrepen heb, mijn helft van de winst. U
mag uw boeking inrichten, zooals u wilt; maar wat u daareven zeide,
dat u uzelf voor kapitaalinbreng te mijnen koste heeft geboekt, is
onzuiver, U kunt uw deel van de winst nemen, zoo goed als ik, maar
het kapitaal blijft van mij, en heeft zijn bestemming.”
„Dat was toch de bedoeling niet. Enfin, laat dan die geldzaken
rusten, en werk uw machine af.”
„Waarvan?”
„Neen, maar ik heb nog accepten. Ziet u eens,” zeide Boom, een
dunne portefeuille opende, „hier zijn er verscheiden. Excuseer mij
nu, men heeft mij al verscheiden malen gewenkt om in te vallen.”
In elk geval, hij begreep nu, dat Boom het geld nog had, zij het in
den vorm van accepten, en dus was de zaak zoover tenminste in
orde. [149]
[Inhoud]
HOOFDSTUK XIV.
„Dat zult u toch niet bij mij zoeken?” vroeg de oude man treurig.
„Ja zeker!” riep dr. Arnolds uit. „En u verkeert in dat geval? Hé, daar
wist niets van.”
„Mijn meisje komt zelden hier, en dan nog maar voor korten tijd. Zij
woont in Berlijn. Uw zoon kent haar heel goed.”
„Neen, dat begrijp ik niet,” zeide Boom. „Ik zou eer denken, dat als
men het maar meer algemeen wist, de aanvragen niet zouden
ontbreken.”
„Er zullen toch verscheidenen zijn, die zouden willen trouwen, als ze
geld hadden om hun eerste inrichting te bekostigen. Zoo in het
algemeen, bedoel ik.”
„O, maar daar bemoeit zich de vereeniging niet mee. Enkel gevallen
meisjes trekt zij zich aan, en tracht degenen die er de oorzaak van
zijn, op deze wijze te bewegen hun plicht te doen.”
„Welnu, dan laat men ’t meisje even vallen, als dat per se een
voorwaarde is,” lachte Boom.
„Sst! Laat u toch nooit meer zoo uit, of de dames zouden.… Van u is
het toch heusch waar?”
„En of! Hoeveel krijg ik er voor? Uw zoon sprak toen van duizend
gulden.”
„Ja, dat kan, voor iemand als u. En trouwens, ze zullen blij zijn weer
eens een geval te hebben. Wanneer moet u het geld hebben?”
„Het is jammer, dat er juist een paar dames op reis zijn. Want ze
weigeren wel nooit, maar willen er allen in gekend zijn. Een beetje
uitpluizen.…”
„Dan zal ik moeten wachten. Dat spijt me. Ik had er alles voor in
orde; tot zelfs de acte van huwelijksche voorwaarden is besteld. Hoe
lang zou het duren, denkt u?”
„Een kleine zes weken. Maar ik weet wat. Als ik u eens [151]een
accept van mij gaf, op twee maanden, en u maakte daar geld op?
Binnen dien tijd heb in het in orde, en los mijn accept in.”
„Een accept van u? Dat zal moeielijk te plaatsen zijn, zoo vlak na dat
geval. Maar om ’t even. Ik wil het wel hebben, en kan ik het niet
kwijt, dan moet ik maar zoolang wachten. Een zegeltje heb ik wel bij
me. Mag ik een pen van u hebben?”
„Ja, dat weet ik,” zeide Boom, het papiertje dichtvouwend. „Dank u.
Ik hoop u spoedig mijn huwelijk te annonceeren.—Hoe staat het met
de zaak van uw zoon?”
„Neen, dat is het niet. Viehof heeft nog wat tegoed van vroeger. Hij is
een tijd geleden ook vervolgd, en toen hebben ze hem moeten