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 Frédéric Patras
Visit to download the full and correct content document:
https://ebookmeta.com/product/classical-hopf-algebras-and-their-applications-algebra
-and-applications-29-pierre-cartier-frederic-patras/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Algebra and Applications 2 Combinatorial Algebra and

Hopf Algebras 1st Edition Makhlouf Abdenacer
https://ebookmeta.com/product/algebra-and-
applications-2-combinatorial-algebra-and-hopf-algebras-1st-
edition-makhlouf-abdenacer/
Algebra and Applications 1: Non-associative Algebras

and Categories 1st Edition Makhlouf
https://ebookmeta.com/product/algebra-and-applications-1-non-
associative-algebras-and-categories-1st-edition-makhlouf/
Chess Explained The c3 Sicilian 1st Edition Sam Collins
https://ebookmeta.com/product/chess-explained-
the-c3-sicilian-1st-edition-sam-collins/
Starting Out The c3 Sicilian 1st Edition John Emms
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/
Algebra 4: Lie Algebras, Chevalley Groups, and their

Representations 1st Edition Ramji Lal
https://ebookmeta.com/product/algebra-4-lie-algebras-chevalley-
groups-and-their-representations-1st-edition-ramji-lal/
The American Revolution 1774 83 2nd Edition Daniel

Marston
https://ebookmeta.com/product/the-american-
revolution-1774-83-2nd-edition-daniel-marston/
BNF 83 (British National Formulary) March 2022 Joint

Formulary Committee
https://ebookmeta.com/product/bnf-83-british-national-formulary-
march-2022-joint-formulary-committee/
English Grammar Exercises with answers Part 4 Your

quest towards C2 1st Edition Daniel B. Smith
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
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.
More information about this series at http://www.springer.com/series/6253

Pierre Cartier Frédéric Patras
•
Classical Hopf Algebras

and Their Applications
123
Pierre Cartier Frédéric Patras
Institut des Hautes Études Scientifiques Laboratoire J.A.Dieudonné
Bures-sur-Yvette, France Université Côte d’Azur
Nice, France
ISSN 1572-5553 ISSN 2192-2950 (electronic)

ISBN 978-3-030-77844-6 ISBN 978-3-030-77845-3 (eBook)
https://doi.org/10.1007/978-3-030-77845-3
Mathematics Subject Classification: 16T05, 16S30, 16T15, 16T30, 16W10
© Springer Nature Switzerland AG 2021

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with regard
to jurisdictional claims in published maps and institutional affiliations.
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.
Limours, France Pierre Cartier

Nice, France Frédéric Patras
Conventions
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
Comk (resp., Com): category of commutative unital algebras.

dij : Kronecker delta function (dij ¼ 1 if i ¼ j and 0 else).
dM (resp., d): coproduct for a comodule M.
dV : identity of V viewed as an element of V V .
dx : delta function (dx ðyÞ :¼ dyx ).
DC , DH (resp., D): coproduct of the coalgebra C, the Hopf algebra H...
Dd : deconcatenation coproduct.
Du : unshuffle coproduct.
Dl : l 1-fold iteration from C to C l of a coassociative coproduct
D : C ! C C.
reduced coproduct associated to a coproduct D.
D:
: dual (V is the dual of V).
gA (resp., g): unit map k ! A of an algebra, coaugmentation gC : k ! C of a
coalgebra.
EncðVÞ ¼ V V ¼ End _ ðV Þ: vector space of linear endomorphisms of a
finite-dimensional vector space V viewed as a coalgebra.
End _ ðVÞ ¼ V V: dual coalgebra of the algebra EndðVÞ ¼ V V of linear
endomorphisms of a finite-dimensional vector space V.
End C ðXÞ (resp., EndðXÞ): set of endomorphisms of X in C.
eC , eA (resp., e): counit map C ! k of a coalgebra, augmentation eA : A ! k of
an algebra.
f jX : restriction of a map f to X, a subset of the domain of f .
f jA : when a vector space decomposes as W ¼ A B and f is a linear map from V
to W, it denotes the composition of the projection from W to A along B with f . We
call f jA the corestriction of f to A.
Fin: category whose objects are the finite (possibly empty) subsets of N and
whose morphisms are the bijections.
CðBÞ: set (resp., group) of group-like elements of a coalgebra (resp., Hopf
algebra) B.
GLðn; kÞ: n-th general linear group over k.
Grp: category of groups.
HomC ðA; BÞ: set of morphisms from A to B in the category C.
Hopk (resp., Hop): category of Hopf algebras.
Hopck (resp., Hopc ): category of complete Hopf algebras.
Id C (resp., Id): identity map of an object C in a given category.
k: ground field.
kG: group algebra of the group G.
kG : k-valued functions on G.
k½V: space of polynomials over a vector space V.
Lx f : left translate of f , Lx f ðyÞ ¼ f ðxyÞ.
Liek (resp., Lie): category of Lie algebras.
Link (resp., Lin): category of vector spaces.
Conventions xi
Linck (resp., Linc ): category of complete filtered vector spaces.

Linkf (resp., Linf ): category of filtered vector spaces.
Lingk (resp., Ling ): category of graded vector spaces.
mA , mH (resp., m): algebra product of the algebra A, the Hopf algebra H...
ml : ðl 1Þ-fold iterated product of an algebra, from Al to A.
M n : n-th tensor power M . . . M of M.
M n ðkÞ: square matrices of size n n over k.
ModA : category of left modules over an algebra A.
Mon: category of monoids.
½n :¼ f1; . . .; ng.
N: nonnegative integers.
N : positive integers.
mH (resp., m): mH :¼ gH eH , unit map of the convolution algebra End Lin ðHÞ of
linear endomorphisms of a Hopf algebra.
½0 :¼ ;.
ObðCÞ: class of objects of the category C.
Oðn; kÞ: n n orthogonal group.
Q
N
ai : in an associative algebra, and denotes the ordered product a1 . . . aN .
i¼1
PrimðCÞ: set of primitive elements of a coaugmented coalgebra C.
Q: rational numbers.
Ry f : right translate of f , Ry f ðxÞ ¼ f ðxyÞ.
RðGÞ: representative functions on a monoid or a group.
R: real numbers.
: shuffle product.
Set: category of sets.
Sn : n-th symmetric group.
Spe: category of vector species.
T: switch map (Tðx yÞ :¼ y x), also written T C;D when mapping C D to
D C.
TðVÞ: space of tensors a V n over a vector space V.
n2N
TSðVÞ: space of symmetric tensors over a vector space V.
Z: integers.
Contents
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
Part I General Theory

2 Coalgebras, Duality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1 Preliminaries on Vector Spaces and Algebras . . . . . . . . . . . . . . 9
2.2 Coalgebras: Definition and First Properties . . . . . . . . . . . . . . . . 10
2.3 Primitive and Group-Like Elements . . . . . . . . . . . . . . . . . . . . . 15
2.4 Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5 Endomorphisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.6 The Structure of Coalgebras . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.7 Representative Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.8 Comodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.9 Representations and Comodules . . . . . . . . . . . . . . . . . . . . . . . . 26
2.10 Algebra Endomorphisms and Pseudo-coproducts . . . . . . . . . . . 28
2.11 Coalgebra Endomorphisms and Quasi-coproducts . . . . . . . . . . . 31
2.12 Duals of Algebras and Convolution . . . . . . . . . . . . . . . . . . . . . 33
2.13 Graded and Conilpotent Coalgebras . . . . . . . . . . . . . . . . . . . . . 36
2.14 Bibliographical Indications . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3 Hopf Algebras and Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.1 Bialgebras, Hopf Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.2 Modules and Comodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
xiii
xiv Contents
3.3 Characteristic Endomorphisms and the Dynkin Operator . . . . . . 48

3.4 Hopf Algebras and Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.5 Algebraic Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.6 Unipotent and Pro-unipotent Groups . . . . . . . . . . . . . . . . . . . . 56
3.7 Enveloping Algebras, Groups, Tangent Spaces . . . . . . . . . . . . . 57
3.8 Filtered and Complete Hopf Algebras . . . . . . . . . . . . . . . . . . . 61
3.9 Signed Hopf Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.10 Module Algebras and Coalgebras . . . . . . . . . . . . . . . . . . . . . . . 65
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4 Structure Theorems . . . . . . . . . . . . . . . . . . . . . . .............. 71
4.1 Dilations, Unipotent Bialgebras, and Weight
Decompositions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.2 Enveloping Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.3 Cocommutative Unipotent Hopf Algebras . . . . . . . . . . . . . . . . 78
4.4 Commutative Unipotent Hopf Algebras . . . . . . . . . . . . . . . . . . 81
4.5 Cocommutative Hopf Algebras . . . . . . . . . . . . . . . . . . . . . . . . 82
4.6 Complete Cocommutative Hopf Algebras . . . . . . . . . . . . . . . . . 84
4.7 Remarks and Complements . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
5 Graded Hopf Algebras and the Descent Gebra . . . . . . . . . . . . . . . . 93
5.1 Descent Gebras of Graded Bialgebras . . . . . . . . . . . . . . . . . . . 94
5.2 Lie Idempotents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
5.3 Logarithmic Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
5.4 The Descent Gebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.5 Combinatorial Descents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
6 Pre-lie Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
6.1 The Basic Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
6.2 Symmetric Brace Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
6.3 Free Pre-Lie Algebras and Gebras of Trees . . . . . . . . . . . . . . . 124
6.4 Left-Linear Groups and Faà di Bruno . . . . . . . . . . . . . . . . . . . 129
6.5 Exponentials and Logarithms . . . . . . . . . . . . . . . . . . . . . . . . . . 131
6.6 The Agrachev–Gamkrelidze Group Law . . . . . . . . . . . . . . . . . . 134
6.7 Other Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
6.8 Brace Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
6.9 Right-Handed Tensor Hopf Algebras . . . . . . . . . . . . . . . . . . . . 144
6.10 Commutative Shuffles and Quasi-shuffles . . . . . . . . . . . . . . . . . 148
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Contents xv
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
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
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
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
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Appendix A: Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247

Appendix B: Operads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
Chapter 1
Introduction
The main purpose of this volume is to give a modern, up-to-date, presentation of

the theory of Hopf algebras and their applications, classical and recent. The Hopf
algebras we consider are the classical ones: the ones that appeared in the 40s and
50s in algebraic topology, the theory of algebraic groups and representation theory.
That is, they will be most often commutative or cocommutative and associated to a
group of characters or group-like elements. Concretely, we will consider typically
enveloping algebras of graded or complete Lie algebras and their dual Hopf algebras,
Hopf algebras of representative functions, of trees and graphs, and similar ones. The
account will include certain Hopf algebras carrying extra algebraic structures, such
as, for example, enveloping algebras of pre-Lie algebras. Applications developed
will include duality phenomena in group theory; classical Hopf algebra structures in
algebraic topology; combinatorial Hopf algebras; and Hopf algebraic renormaliza-
tion.
Let us start with some elementary and very general principles governing the
theory of Hopf algebras (and variants thereof that will be studied in this book). We
will illustrate them on the simple example of finite groups, although it will appear
later on that their range of application is much larger. We refrain from making at this
point all definitions rigorous, this introduction aiming at indicating the behavior of
some objects and structures that will be defined thoroughly later.
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
m(λ ⊗ κ)(g) = (λ · κ)(g) := λ(g)κ(g)
for λ, κ ∈ k G and g ∈ G, and is equipped with a linear map (the coproduct)
: 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
where δx stands for the delta function on G, δx (g) = 1 if g = x and 0 else.

Whereas this simple dualization process is quite satisfactory for finite groups,
properly defining the linear object dual to more general group structures (infinite
groups, continuous groups, algebraic groups, group structures in categories, renor-
malization groups...) is, in general, less straightforward and requires some care.
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
1.4 Natural Endomorphisms
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.
1.6 Structure of the Book
Part I of the book addresses the general theory.

Chapter 2 deals with coalgebras, comodules, duality, representative functions, and
structural properties of endomorphisms (notions of pseudo- and quasi-coproducts).
Chapter 3 introduces fundamental Hopf algebra notions and examples and some
direct generalizations thereof, modules and comodules, algebraic and pro-unipotent
groups, enveloping algebras, and the Dynkin operator.
Chapter 4 is about structure theorems for commutative and cocommutative Hopf
algebras, following the approach by properties of their endomorphisms.
Chapter 5 studies the properties of endomorphisms in the graded case. It introduces
the descent gebra and some of its main applications: Lie idempotents, logarithmic
derivatives.
Chapter 6 introduces pre-Lie algebras and the equivalent notion of symmetric
brace algebras, together with fundamental examples. Special attention is devoted
to the two exponential maps on the completion of their enveloping algebras and
their relations, together with the associated group-theoretical notions. The chapter
also includes a brief introduction to brace algebras, a non-symmetric version of
symmetric brace algebras. The notion is illustrated with the example of quasi-shuffle
Hopf algebras.
Part II is devoted to applications.
Chapter 7 illustrates the duality between groups and commutative Hopf algebras
with the example of compact Lie groups. The last part explains the mechanisms
underlying Tannaka duality—the construction of coalgebras and bialgebras out of
suitable linear categories.
Chapter 8 treats Hopf algebras in algebraic topology with structure theorems for
the homology and cohomology of H -spaces, including the case of nonzero charac-
teristic and the notion of divided powers. The bar construction is introduced and the
computation of the homology of Eilenberg–MacLane spaces explained. The chapter
ends with the study of the algebra of stable cohomology operations or Steenrod
algebra.
Chapter 9 considers combinatorial Hopf algebras in the context of species (func-
tors from the category of finite sets and bijections between them). Two equivalent
notions of Hopf algebras exist in this context: twisted Hopf algebras and Hopf species.
Functors to usual Hopf algebras connect the twisted theory with the classical one.
We investigate their combinatorial properties, give examples, and introduce the main
constructions.
6 1 Introduction
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
2.1 Preliminaries on Vector Spaces and Algebras
Let V, W be two vector spaces, V ∗ and W ∗ their duals. For v∗ ∈ V ∗ , v ∈ V , we

write < v |v∗ >=< v∗ |v > for v∗ (v ) ∈ k. The map from W ⊗ V ∗ to Lin(V, W ),
where Lin stands for the category of vector spaces,
(w ⊗ v∗ )(v ) := w· < v∗ |v >
is an isomorphism when V is finite dimensional. In particular, when V = W , the

map V ⊗ V ∗ → End(V ) is an isomorphism when V is finite dimensional, and to
the identity map I dV of V corresponds then an element δV ∈ V ⊗ V ∗ .
An associative unital algebra over k (or, for short, an algebra) is a triple (A, m, η)
where A is a vector space, m : A ⊗ A → A (multiplication) and η : k → A (unit)
are linear maps such that the following diagrams, expressing, respectively, the asso-
ciativity and unit axioms, commute
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
One advantage of this diagrammatic definition of algebras is that it generalizes

immediately to arbitrary categories equipped with a tensor product (tensor categories,
see Appendix A for a definition).
Recall also that a Lie algebra is a vector space L equipped with a bilinear product,
usually written as a bracket [ , ], such that for any x, y, z ∈ L,
• (Antisymmetry) [x, x] = 0 (hence [x, y] = −[y, x]),
• (Jacobi identity) [x, [y, z]] + [y, [z, x]] + [z, [x, y]] = 0.
© Springer Nature Switzerland AG 2021 9

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_2
10 2 Coalgebras, Duality
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.
2.2 Coalgebras: Definition and First Properties
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 .
We will very often use the abbreviated Sweedler notation
(c) =: c(1) ⊗ 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)
Diagrammatically, the coproduct is coassociative if the following diagram com-

mutes:
2.2 Coalgebras: Definition and First Properties 11
- C ⊗C
C
? ?Id ⊗
C ⊗C - C ⊗ C ⊗ C.
⊗ Id
We denote the map ( ⊗ I d) ◦ by 3 . When the coproduct is coassociative,

there is a unique map from C to C ⊗n induced by iterated compositions of coproducts.
It is written n and can be defined recursively by 2 := ,
n := ( ⊗ I d ⊗n−2 ) ◦ n−1 ,
the coassociativity hypothesis implying that for any i ≤ n − 2,
n = (I d ⊗i ⊗ ⊗ I d ⊗n−i−2 ) ◦ n−1 .
More generally, one gets by induction
n 1 +···+n k = (n 1 ⊗ · · · ⊗ n k ) ◦ k
for n 1 , . . . , n k > 0, with the notation 1 := I d.

In the Sweedler notation, the identity (2.1) would read, for an arbitrary element
c ∈ C,
(1) (2) (1) (2)
c(1) ⊗ c(1) ⊗ c(2) = c(1) ⊗ c(2) ⊗ c(2) .
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 ◦ = ,
where T is the switch map T (a ⊗ b) := b ⊗ a. In the Sweedler notation,
c(1) ⊗ c(2) = c(2) ⊗ c(1)
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)) .
In general, T ◦ defines a new coproduct on C, the opposite coproduct.

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)
where we use the isomorphisms k ⊗ C ∼

=C ⊗k ∼
= C. Diagrammatically,

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).
Definition 2.2.1 A coalgebra is a vector space C equipped with a coassociative

coproduct and a counit ε. It is cocommutative if is cocommutative.
We will call noncounital coalgebra a k-vector space C equipped with a coasso-

ciative coproduct but without a counit. One reason for this convention is that most
of the coalgebras we will consider will be equipped with a counit.
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
defines a coalgebra structure on C.
Given two coalgebras C and D, equipped, respectively, with coproducts C ,

D and counits εC , ε D , a morphism of coalgebras from C to D is a morphism
f ∈ Lin(C, D) that commutes to the structure maps (the coproducts and counits).
That is,
( f ⊗ f ) ◦ C = D ◦ f, ε D ◦ f = εC .
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).
Remark 2.2.1 Notice that if (C, , ε) is a coalgebra, (C, T ◦ , ε) is also a coal-

gebra: the opposite coalgebra of C, denoted C op .
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.
Lemma 2.2.1 (Dual of a coalgebra) Let C be a coalgebra, its dual C ∗ is naturally

equipped with the structure of an algebra.
Proof Indeed, there is a canonical embedding C ∗ ⊗ C ∗ → (C ⊗ C)∗ , defined by

< λ ⊗ β|c ⊗ d >=< λ|c >< β|d >, where C ∗ denotes the linear dual of C, λ, β ∈
C ∗ , c, d ∈ C. Composition with the dual map of , from (C ⊗ C)∗ to C ∗ , defines
the product on C ∗ . The associativity and unit properties follow by duality, the unit
of C ∗ being simply the counit of C viewed as a linear form.
The product on C ∗ is written as a convolution product,
λ ∗ β := m k ◦ (λ ⊗ β) ◦ C ,
where m k is multiplication in k. It is an instance of a more general construction

defining an associative product on the vector space of linear maps from a coalgebra
to an arbitrary algebra, see Sect. 2.12.
Remark 2.2.2 When C is finite dimensional, C = (C ∗ )∗ and there is a perfect dual-

ity between finite-dimensional algebras and finite-dimensional coalgebras. The cat-
egories are anti-equivalent, f : C → D corresponding to f ∗ : D ∗ → C ∗ . The struc-
ture maps of the algebras are (as always) dual to the structure maps of the dual
coalgebras, and the converse statement also holds.
The coproduct
C⊗D := (I dC ⊗ TC,D ⊗ I d D ) ◦ (C ⊗ D ),
where TC,D (c ⊗ d) := d ⊗ c, and the counit
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.
Remark 2.2.3 In the category of cocommutative coalgebras, the tensor product of

coalgebras is the categorical product:
Cog(C, D) × Cog(C, E) ∼
= Cog(C, D ⊗ E),
where the isomorphism is obtained from the composition of maps

- 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 ,
is the coproduct of A and B:
Com(A, C) × Com(B, C) ∼
= Com(A ⊗ B, C),
where the isomorphism is obtained from the composition of maps

f ⊗g
- m-
A⊗B C ⊗C C
C.
Definition 2.2.2 A subcoalgebra of a coalgebra C is a subvector space D of C such

that (D) ⊂ D ⊗ D ⊂ C ⊗ C. It is a coalgebra whose counit is the restriction to D
of the counit of C, ε D := εC|D . It is called proper in case D = C and D = 0.
Recall that an (two-sided) ideal I of an associative algebra A with product m is
a subvector space of A such that m(A ⊗ I ) ⊂ I and m(I ⊗ A) ⊂ I . This notion of
ideal is dual to the one of subcoalgebra in the following sense. Assume that A = C ∗
and let I ⊂ C ∗ be the annulator of a subcoalgebra D,
∀λ ∈ C ∗ , (λ ∈ I ) ⇐⇒ (∀d ∈ D, λ(d) = 0).
Then, for λ ∈ I , β ∈ C ∗ and d ∈ D,
λ ∗ β(d) = λ(d (1) )β(d (2) ) = 0 = β(d (1) )λ(d (2) ) = β ∗ λ(d)
since (d) = d (1) ⊗ d (2) ∈ D ⊗ D. Hence λ ∗ β and β ∗ λ are in I .

Definition 2.2.3 A coideal of C is a subvector space I such that (I ) ⊂ I ⊗ C +
C ⊗ I and I ⊂ K er (ε). For such an I , C/I is equipped with a coalgebra structure
by the induced maps
: C/I → C/I ⊗ C/I, ε : C/I → k.
It is called the quotient coalgebra of C by I .

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 .
2.3 Primitive and Group-Like Elements
Definition 2.3.1 (Group coalgebras) The vector space of functions k G on a finite

group G is a coalgebra for the coproduct (1.1) and the counit ε(λ) := λ(1), for
λ ∈ kG.
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
and dually on kG, for g and h in G
(g) = g ⊗ g, g · h = gh.
The coalgebraic part of this construction holds more generally for all sets:
Definition 2.3.2 (Set coalgebras) The maps
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.
These phenomena suggest the definition:

Definition 2.3.3 (Group-like elements) Let C be a coalgebra. A group-like element

of C is an element c such that ε(c) = 1 and
(c) = c ⊗ c.
The set of group-like elements is written (C).
Equivalently, because of counitality, a group-like element is an element of C such

that c = 0 and (c) = c ⊗ c.
As we shall see later, (C) is actually a group when C is a Hopf algebra. This
is the case of kG, for example, with (kG) = G. This last property is a general
phenomenon:
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.
Exercise 2.3.2 Let C be a coalgebra and X, Y be sets. Show that
Cog(k X, C) ∼
= Set(X, (C))
and that
Cog(k X, kY ) ∼
= Set(X, Y ).
Lemma 2.3.1 If C is a finite-dimensional coalgebra, there is a natural isomorphism

between (C) and characters (algebra morphisms to the ground field) on the dual
algebra
(C) ∼
= Alg(C ∗ , k).
Proof Indeed, in that case Alg(C ∗ , k) ⊂ (C ∗ )∗ ∼

= C and (C ∗ ⊗ C ∗ )∗ ∼ = C ⊗ C.
Therefore, for c ∈ C, c ∈ Alg(C , k) if and only if, for any a, b in C ∗ ,
∗
< a|c >< b|c >=< ab|c >=< a|c(1) >< b|c(2) >,
that is, if and only if (c) = c ⊗ c.
Definition 2.3.4 (Coaugmented coalgebra) A coaugmented coalgebra is a coalge-

bra C equipped with a morphism of coalgebras η : k → C.
Equivalently, a coaugmented coalgebra is a coalgebra with a fixed group-like

element (η(1)). We will use the notation η(1) = 1 and sometimes also, slightly
abusively, η(1) = 1.
Definition 2.3.5 (Reduced coproduct) On a coaugmented coalgebra, the reduced

coproduct ¯ is defined by
¯
(x) := (x) − x ⊗ 1 − 1 ⊗ x,
2.3 Primitive and Group-Like Elements 17
¯ : C̄ → C̄ ⊗ C̄ (recall that C̄ = K er ε). One can check

so that, as is counital,
that it makes C̄ a noncounital coalgebra. The iterated reduced coproduct
¯ n : C̄ → C̄ ⊗n

¯ 2 := ,
is defined inductively by ¯
¯ n+1 = (
¯ ⊗ I dC⊗n−1 ) ◦
¯ n.
The notion of coaugmented coalgebra is dual to the one of augmented algebra,

that is, an algebra equipped with an algebra map to the ground field. Any finite-
dimensional algebra gives rise by duality to a finite-dimensional coalgebra, and any
finite-dimensional augmented algebra to a finite-dimensional coaugmented coalge-
bra.
Definition 2.3.6 (Primitive elements) Let C be a coaugmented coalgebra. An ele-

¯
ment c of C̄ is called primitive if and only if (c) = 0 or, equivalently, (c) =
c ⊗ 1 + 1 ⊗ c.
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.
Definition 2.4.1 (Deconcatenation coalgebra) The tensor gebra T (V ) equipped

with the deconcatenation coproduct

n
d (v1 . . . vn ) := v1 . . . v p ⊗ v p+1 . . . vn ,
p=0
and the counit ε(1) = 1, ε(v1 . . . vn ) = 0 for n ≥ 1, is a coalgebra called the

deconcatenation coalgebra over V . We used the convention v1 . . . v0 := v∅ := 1 =:
vn+1 . . . vn .
Definition 2.4.2 (Unshuffle tensor coalgebra) The tensor gebra T (V ) equipped with
the unshuffle coproduct

u (v1 . . . vn ) := vI ⊗ v J

I J =[n]
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 ,
σ (v1 . . . vn ) := vσ −1 (1) . . . vσ −1 (n) .

The space T S(V ) := T S n (V ) is called the symmetric gebra or gebra of symmetric
n≥0
tensors over V .
Symmetric tensors can be studied using the polarization process, according to
which T S n (V ) is generated linearly by tensor powers γn (v) := v . . . v = v⊗n . For
example, when n = 2, for arbitrary v1 , v2 in V , we have the identity
(v1 v2 + v2 v1 ) = γ2 (v1 + v2 ) − γ2 (v1 ) − γ2 (v2 ).
The general case follows from the following identity,

valid in an arbitrary ring R.1
Let x1 , . . . , xn ∈ R and for H ⊂ [n] set x H := xi . We then have
i∈H

(−1)n xσ (1) . . . xσ (n) = (−1)Car d(H ) (x H )n .
σ ∈Sn H ⊂[n]
Lemma 2.4.1 The symmetric gebra T S(V ) is a subcoalgebra of the deconcatena-

tion coalgebra (T (V ), d ) and of the unshuffle tensor coalgebra (T (V ), u ).
Proof Since, by polarization, T S(V ) is generated by the tensor powers γn (v), the
Lemma follows from the identities:
1 See Bourbaki, Algèbre I, 1970, A I.95, Prop. 2.

2.4 Tensors 19

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
For an arbitrary finite-dimensional vector space V , the algebra structure of End(V ) =

V ⊗ V ∗ is obtained from the composition
(V ⊗ V ∗ ) ⊗ (V ⊗ V ∗ ) = V ⊗ (V ∗ ⊗ V ) ⊗ V ∗ → V ⊗ k ⊗ V ∗ ∼
= V ⊗ V ∗.
Dually, we write End ∨ (V ) := V ∗ ⊗ V the dual coalgebra of the algebra End(V ):
End ∨ (V ) → End ∨ (V ) ⊗ End ∨ (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.
Definition 2.5.1 (Endomorphism coalgebra) We denote Enc(V ) and call endomor-

phism coalgebra of V the vector space of linear endomorphisms End(V ) = V ⊗ V ∗
viewed as a coalgebra for the coproduct

n
(v ⊗ w∗ ) := (ei ⊗ w∗ ) ⊗ (v ⊗ ei∗ ),
i=1
and counit the trace map ε(v ⊗ w∗ ) :=< w∗ |v >.

It is the opposite coalgebra to End ∨ (V ∗ )—a choice dictated by the study of right
comodules later in this chapter.
2.6 The Structure of Coalgebras
A key property of coalgebras is that each element is contained in a finite-dimensional

subcoalgebra. This property explains largely why coalgebras are in many situations
easier to deal with than algebras, although using coproducts may seem more difficult
or less natural than products. This property is usually referred to as the fundamental
theorem for coalgebras.
Theorem 2.6.1 (Fundamental theorem for coalgebras) Let c ∈ C, where C is a
coalgebra. Then c is contained in a finite-dimensional subcoalgebra of C.

n
Proof Indeed, let us expand 3 (c) ∈ C ⊗3 as a finite sum ai ⊗ bi ⊗ ci with n
i=1
minimal. This implies, in particular, that the tensor products ai ⊗ bi (resp., ai ⊗ ci ,
2.6 The Structure of Coalgebras 21
bi ⊗ ci ) are linearly independent in C ⊗2 . Else, assuming for example that, up to a

reordering of the indices,

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
from which it follows that 4 (c) ∈ C ⊗ C ⊗ C ⊗ C ∩ C ⊗ C ⊗ C ⊗ C = C ⊗

C ⊗ C ⊗ C . The linear independency of the ai ⊗ ci implies then that (bi ) ∈
C ⊗ C , and similarly for (ai ) and (ci ).
k
Expand now (c) as a finite sum x j ⊗ y j with k minimal. The xi (resp., the
j=1
yi ) are then linearly independent and, from

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.
2.7 Representative Functions
Let X be a monoid and consider k X := Set(X, k). An element f in k X is called

representative if and only if there exist finitely many f i , f i ∈ k X , i ∈ I such that
for any x, y ∈ X
f (x y) = f i (x) f i (y). (2.3)
i∈I
Linear combinations of representative functions are representative.

Let us introduce further notation. For x ∈ X, f ∈ k X , we denote L x f , resp., Rx f
the left and right translates of f :
L x f (y) := f (x y), R y f (x) := f (x y).

If f is representative, its right and left translates are representative:

L x f (yz) = f (x yz) = L x f i (y) f i (z).
i∈I
Lemma 2.7.1 If f is representative, the space generated by the right translates

R y f, y ∈ X is finite dimensional.
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
and these functions are representative. To π , we associate at last its space of

coefficients C(π ), the subvector space of k X spanned linearly by the functions
πi, j , 1 ≤ i, j ≤ n. More intrinsically, C(π ) is spanned by the set of coefficient
functions cv∗ ,v , cv∗ ,v (g) :=< v∗ |π(g)(v) > for v ∈ V, v∗ ∈ V ∗ . It consists of the set
of functions c A,π (g) := T r (A · π(g)) for A ∈ End(V ).
The union of the C(π ), where π runs over the linear representations of π is
written R(X ). One can deduce that R(X ) is a vector space from the observation that
C(π ) + C(π ) = C(π ⊕ π ). Equivalently, it is the linear span of the C(π ) in k X .
Lemma 2.7.2 If the space of right translates of f is finite dimensional, f ∈ R(X )

and f is representative.
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.
The following proposition summarizes the previous results.
Proposition 2.7.1 For f ∈ k X , the following are equivalent:

1. The function f is representative.
2. The function f belongs to R(X ).
3. The space generated by the left translates L x f, x ∈ X , is finite dimensional.
4. The space generated by the right translates Rx f, x ∈ X , is finite dimensional.
2.7 Representative Functions 23
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.
In terms of representative functions the coproduct is defined in duality with the

product in X ,
( f )(x ⊗ y) = f (x y), ( f ) = f i ⊗ f i .
i∈I
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
There is another approach to group representations than representative functions. It

relies on duality phenomena and is encapsuled in the notion of comodules, which
are the object of the present section.
Definition 2.8.1 Let C be a coalgebra. A comodule (or right-comodule) over C is

a vector space M equipped with a linear morphism
δ : M → M ⊗C
such that
(I d M ⊗ ) ◦ δ = (δ ⊗ I dC ) ◦ δ
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
Proof The commutativity of the diagram
∂C - V ⊗C
V
∂C ? I dV ⊗ C ?
V ⊗C - V ⊗C ⊗C
∂C ⊗ I dC
translates into the identities (for all 1 ≤ j ≤ n)

2.8 Comodules 25

n
n
n
n
ei ⊗ (cij ) = ∂C (ek ) ⊗ ckj = ei ⊗ cki ⊗ ckj .
i=1 k=1 i=1 k=1
That is, for all 1 ≤ i, j ≤ n:

n
(cij ) = cki ⊗ ckj .
k=1
The identity (for all 1 ≤ i, j ≤ n)

n
n
γC ⊗ γC (ek ⊗ ei∗ ⊗ e j ⊗ ek∗ ) = cki ⊗ ckj
k=1 k=1
translates then into the commutativity of the diagram
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
The lemma follows.
To each comodule M is canonically associated a subcoalgebra C(M) of C. It is the

minimal subcoalgebra of C such that δ restricts to a map M → M ⊗ C(M). It can be
defined directly by choosing a basis (m i )i∈I of M. Writing δ(m j ) = i∈I m i ⊗ ci j ,
C(M) is the vector space spanned by the ci j (we let the reader check it is indeed a
coalgebra). Intrinsically, it is the image of the map from Enc(M) = M ⊗ M ∗ to C
induced from δ. When M is a subcoalgebra of C, C(M) = M.
The proof of the fundamental theorem for coalgebras, Thm 2.6.1, can be mimicked
for comodules:
Theorem 2.8.1 (Fundamental theorem for comodules) Let C be a coalgebra, M a

comodule and m ∈ M. Then, m is contained in a finite-dimensional subcomodule M
of M and there exists a finite-dimensional subcoalgebra C of C such that δ(m) ∈
M ⊗ C .
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 .
Proposition 2.8.1 The map D → D̃ is a bijection between the set of subcoalgebras

f
of C and the set of subcategories E of the abelian category CmdC such that:
f
1. Any object of CmdC which is isomorphic to an object, sub-object or quotient
object of an element of E is in E.
f
2. The category E is a full subcategory of CmdC (i.e., for A, B in E we have
f
E(A, B) = CmdC (A, B)).
3. The category E is stable by finite direct sums.
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.
2.9 Representations and Comodules
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).
We get a linear map δ from V to V ⊗ R(G). When V is a finite-dimensional

repre-
sentation, using the notations of Sect. 2.7, we have δ(ei ) = j≤n e j ⊗ π ji .
Since π(gg ) = π(g)π(g ) for arbitrary g, g ∈ G, we have
2.9 Representations and Comodules 27
δ(v)(gg ) = π(gg )(v) = π(g)π(g )(v) = δ(π(g )(v))(g).
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).
Theorem 2.9.1 The categories of locally finite G-modules and R(G)-comodules

are isomorphic.
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
g ∈ G → π(g) ∈ End(V ), π(g)(v) := (I dV ⊗ g) ◦ δ(v),
where we view g ∈ G as a linear form on R(G), defines a locally finite G-module

structure on V . The local finiteness follows from the fundamental theorem for comod-
ules.
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).
The tensor product of R(G)-comodules is therefore induced by the product in

R(G): this phenomenon will be understood in a broader generality later, using the
existence of a Hopf algebra structure on R(G), see in particular Prop. 3.2.2.
2.10 Algebra Endomorphisms and Pseudo-coproducts
In many situations, it is convenient to extend the notion of representative functions

to more general frameworks such as spaces of linear endomorphisms of algebras,
coalgebras, and Hopf algebras. This idea will be essential later in this book, especially
when studying the fine structure of graded Hopf algebras.
Let (A, m, η) be an associative algebra over a field k of characteristic 0, possibly
but not necessarily augmented. When this is the case, we denote by ε : A → k the
augmentation map and A+ the augmentation ideal of A (the kernel of ε). Recall also
that we write then ν := η ◦ ε.
Definition 2.10.1 An element f of End(A), the algebra of linear endomorphisms of

A is said to be a representative endomorphism of A if there exists a F = f F(1) ⊗ f F(2) ∈
End(A) ⊗ End(A) such that for any a, b ∈ A, f F(1) (a) f F(2) (b) = m ◦ F(a ⊗ b) =
f (ab). We then say that F is a pseudo-coproduct for f .
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.
Example 2.10.1 The map f admits the pseudo-coproduct f ⊗ I d + I d ⊗ f if and

only if it is a derivation: f (ab) = f (a)b + a f (b).
Example 2.10.2 Let A := C[x1 , ..., xn ]. The linear endomorphisms f of A with

pseudo-coproduct f ⊗ I d + I d ⊗ f (the derivations) correspond bijectively to poly-
n
nomial vector fields (linear endomorphisms of the form f i ∂xi with f i ∈ A).
i=1
Definition 2.10.2 If the algebra is augmented and f admits the pseudo-coproduct

f ⊗ ν + ν ⊗ f , we say that f is an infinitesimal endomorphism of A.
2.10 Algebra Endomorphisms and Pseudo-coproducts 29
In particular, such a f is the null application on the scalars ( f ◦ ν = 0) since

f (1) = f (1)ν(1) + ν(1) f (1) = 2 f (1). A map f is an infinitesimal endomorphism
if and only it is null on the scalars and its restriction to (A+ )2 , the square of the
augmentation ideal, is the null application (i.e., f is zero on products of elements of
A+ ).
Definition 2.10.3 If f admits the pseudo-coproduct f ⊗ f and f (1) = 1, we say
that f is a group-like element of End(A). An element of End(A) is group-like if
and only if it is an algebra endomorphism of A.
These notions generalize in a straightforward way to linear maps from A to another

algebra B: assume that an algebra map ρ from A to B is fixed, then f ∈ Lin(A, B)
is a derivation, resp., group-like, resp., (when A is augmented) infinitesimal if and
only if for any a, b ∈ A, f (ab) = f (a)ρ(b) + ρ(a) f (b), resp., f (ab) = f (a) f (b)
and f (1) = 1, resp., f (ab) = f (a)η B ◦ ε A (b) + η B ◦ ε A (a) f (b), where η B is the
unit map of B, and ε A the augmentation of A.
Following standard terminology, when B is commutative, we will also call group-
like morphisms (B-valued) characters and infinitesimal morphisms (B-valued)
infinitesimal characters.
The concept of pseudo-coproduct is very flexible, as shown by the following
result.
Theorem 2.10.1 If f, g admit the pseudo-coproducts F, G and α ∈ k, then f +

g, α · f, f ◦ g admit, respectively, the pseudo-coproducts F + G, α · F, F ◦ G,
where the composition product ◦ is naturally extended from End(A) to End(A) ⊗
End(A).
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.
Corollary 2.10.2 The sets of derivations and infinitesimal endomorphisms written,

respectively, Der (A) and Endinf (A) are Lie subalgebras of End(A).
The corollary is an obvious consequence of the definitions, but we include a proof

to illustrate the functioning of pseudo-coproduct calculus.
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].
Let f, g be infinitesimal elements in End(A), a pseudo-coproduct for [ f, g] :=

f ◦ g − g ◦ f is given by (recall that f ◦ ν = 0 and similarly for 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.
The exponential and the logarithm define bijections between a neighborhood of

the identity in a Lie group and a neighborhood of 0 in the tangent space to the identity
(which is a Lie algebra). The same properties hold true for EndAlg (A) and Der (A);
however, some restrictions on A and on the maps under consideration are required
to make sure that the formal power series defining the exponential and the logarithm
maps are well defined. This kind of ideas will be used later in this book to revisit
some parts of the classical theory of free Lie algebras.
Let us fix a framework allowing
to deal with formal power series. In the remaining
part of this section, A = An is a graded connected algebra (the product maps
n≥0
Ai ⊗ A j to Ai+ j and A0 ∼ = k; we write A+ := An ). Notice that we could also
n>0
have assumed that A is a complete augmented algebra, a natural framework for the
following developments, but for didactical reasons we postpone the formal definition
of such an algebra to Section 3.8.
An identity-tangent algebra endomorphism of A is a graded algebra endomor-
phism f such that f (1) = 1 and ( f − I d)(A+ ) ⊂ (A+ )2 . Since ( f − I d)(A+ ) p ⊂
(A+ ) p+1 , ( f − I d)n is the null application on the first n + 1 components, Ai , i =
0, . . . , n, of the graded algebra A and formal power series in ( f − I d) are well
defined. Identity-tangent algebra endomorphisms are therefore invertible
2.10 Algebra Endomorphisms and Pseudo-coproducts 31

f −1 = (I d + ( f − I d))−1 = (−1)n ( f − I d)n .
n≥0
A tangent-to-zero derivation is a graded derivation D such that D(A) ⊂ (A+ )2 .

We write EndAlg 1
(A), resp., Der 0 (A) for the group of identity-tangent elements in
EndAlg (A), resp., the sub-Lie algebra of tangent-to-zero elements in Der (A).
Notice first that, as for Lie groups, it is not true that the exponential and the
logarithm maps define a bijection between EndAlg (A) and Der (A): if it was the
case, any element of EndAlg (A) would be invertible (since e−x is always the inverse
of e x , whenever the corresponding series make sense) but ν is not invertible if A = k.
Theorem 2.10.2 The logarithm and the exponential series define inverse bijections
1
between EndAlg (A) and Der 0 (A).
Proof Let us show that the logarithm is a map from EndAlg 1

(A) to Der 0 (A)
(the reverse property is proved similarly). Any f ∈ EndAlg (A) has the pseudo-
1
coproduct f ⊗ f and ( f − I d) has the pseudo-coproduct f ⊗ f − I d ⊗ I d. There-

(−1)n−1
fore, log( f ) = n
( f − I d)n has the pseudo-coproduct:
n≥1
(−1)n−1
( f ⊗ f − I d ⊗ I d)n = log( f ⊗ f ) = log(( f ⊗ I d) ◦ (I d ⊗ f ))
n≥1
n
which, since f ⊗ I d and I d ⊗ f commute in End(A) ⊗ End(A), equals
log( f ⊗ I d) + log(I d ⊗ f ) = log( f ) ⊗ I d + I d ⊗ log( f ),
which concludes the proof since log( f )(A) ⊂ (A+ )2 .
2.11 Coalgebra Endomorphisms and Quasi-coproducts
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 .
Definition 2.11.1 An element f of End(C), the algebra of linear endomorphisms of

C, is said to be a characteristic endomorphism of C if there exists a F ∈ End(C) ⊗
End(C) such that F ◦ = ◦ f . We then say that F is a quasi-coproduct for f .
Graphically, this amounts to the commutation of the following diagram:

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
Remark 2.11.1 The map f admits the quasi-coproduct f ⊗ I d + I d ⊗ f if and

only if it is a coderivation.
Definition 2.11.2 If the coalgebra is coaugmented, f (1) = 0, and f admits the

quasi-coproduct f ⊗ ν + ν ⊗ f , we say that f is a primitive endomorphism of C.
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:
( f (c)) = ( f ⊗ ν + ν ⊗ f ) ◦ (c) = f (c) ⊗ 1 + 1 ⊗ f (c).
Definition 2.11.3 If f admits the quasi-coproduct f ⊗ f and ε ◦ f = ε, we say

that f is a group-like element of End(C). An element of End(C) is group-like if
and only if it is a coalgebra endomorphism of C.
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.
These notions generalize in a straightforward way to linear maps from C to another

coalgebra D.
As the concept of pseudo-coproduct, the one of quasi-coproduct is very flexible:
Theorem 2.11.1 If f, g admit the quasi-coproducts F, G and α ∈ k, then f + g, α ·

f, f ◦ g admit, respectively, the quasi-coproducts F + G, α · F, F ◦ G, where the
product ◦ is naturally extended to End(C) ⊗ End(C).
Proof We have ◦ ( f ◦ g) = F ◦ (g) = F ◦ G ◦ , so that F ◦ G is a quasi-

coproduct for f ◦ g. The other assertions are straightforward.
Corollary 2.11.1 The sets of coderivations and primitive endomorphisms written,

respectively, CoDer (C) and Endprim (C) are Lie subalgebras of End(C).
Another random document with
no related content on Scribd:
[Inhoud]
HOOFDSTUK XIII.
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 …”
„Betalen, ja wel … Meneer heeft nu en dan wat bij mij gehaald, en …”
„Toch nooit buiten de maandelijksche rente?” vroeg zij.
„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]
„Je liegt, schurk!”

Wiechen was met stomheid geslagen. Even bewoog hij de lippen,
waaruit de weinige kleur nu geheel was verdwenen. Toen begon een
dun straaltje bloed uit zijn neus te loopen, afdruppelend op den
grond.
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.
Op straat gekomen, bedaarde Betsy langzamerhand. Zij begon

zichzelf kinderachtig te vinden, om op het zien van een beetje bloed
op den loop te gaan. Misschien was die neusbloeding maar een
soort vertooning geweest. Maar dit denkbeeld vond ze weer wat al te
kras. Hoewel, voor iemand, die met een effen gezicht zulke vuile
insinuaties over Jan durfde loslaten, was eigenlijk niets te kras.
Inmiddels had zij door haar drift haar doel gemist. Hoe dom van
haar, en wat zou Jan ervan zeggen? Zij begon de gevolgen van haar
daad te overwegen, en opeens keek zij in de straat om, angstig, of
ook iemand gezien kon hebben, dat zij uit het huis van dien man
kwam.… „O!”
De kreet was haar ontsnapt. Marie stond vóór haar.
„Was je bij mij?” vroeg deze, niet zonder verwondering.
„Neen,” antwoordde Betsy hard, „dergelijken omgang heb ik niet.”
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.
„Mevrouw van Vleuten?” vroeg de advocaat. „Waarmee kan ik u van

dienst zijn?”
Betsy legde het geval bloot, en overhandigde den advocaat de

stukken.
Deze doorliep ze even, en trok toen de schouders op.
„Hetzelfde, precies hetzelfde,” mompelde hij. En toen luider:

„Mevrouw, ik kan u geen opwekkende mededeeling doen. Er is juist
een proces beëindigd, op nagenoeg gelijke stukken als dit. Ik houd
het ervoor, dat het een piège is, opzettelijk zoo gemaakt om er de
menschen te laten inloopen. Kent u den inhoud van deze acte?”
„Ik heb dat groote stuk gelezen,” zeide zij.
„De acte van deposito. Kijkt u nu zelf nog eens na, of u daarin iets
vinden kunt over de verschuldigde rente.”
„Neen,” erkende zij. „Maar wat doet er dat toe?”
„De quitanties, die ik hier zie, vermelden evenmin de strekking. Ze

kunnen zijn, zooals uw man waant, voor overeengekomen rente.
Zeker, maar het staat er niet op. Van den kant van Wiechen zal men
beweren, dat het disposities zijn, in rekening op het deposito.”
„Zoodat dit steeds kleiner wordt?” vroeg zij angstig. [140]
„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.”
„Maar als wij hem eens een eed lieten doen.…?”
„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.”
„Zweren de menschen dan allemaal valsch?” vroeg Betsy naief.
„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.”
„Zeer waarschijnlijk, en zelfs moet u er op rekenen, dat hij nog een

tijdlang daarna rekt.”
„Rekt? Hoe bedoelt u dat?” [141]
„Langzaam procedeeren. Veel uitstel vragen. Trouwens, door de

fraaie inrichting ten onzent, is men altijd zeker van een jaar
vertraging, als men eenvoudig pleidooi vraagt. Er zijn zóó veel
pleidooien ingeschreven, dat we binnen het jaar geen pleitdag
krijgen.”
„Maar zijn de rechters de zaak dan niet weer vergeten?”
„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.”
„O ja, gelukkig,” zeide zij. „Anders hadden we de zaak geen drie

maanden kunnen laten loopen. Maar we zullen zuinig moeten zijn.”
„We zouden iets kunnen probeeren. Heeft uw man een paar

vrienden, die hem voor een kleine comedie van dienst zouden willen
zijn?”
„We hebben weinig omgang,” zeide Betsy. „Maar … hij kent hier een
zekeren Boom, dien hij met geld heeft bijgestaan.”
„Voor zijn uitvinding zeker?” lachte de advocaat.
„Ja …” deed Betsy verbluft. „Maar dat mocht ik niet vertellen.”

„Is het veel?”
„Tweeduizend gulden.”
„Altijd jammer. Uw man heeft het dan wèl getroffen.”
„Is dat geld dan weg?”
„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]
„Maar die Boom leek me zoo’n fatsoenlijk man!”
„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?”
Betsy zweeg, terwijl de tranen haar in de oogen schoten.
„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?”
„Vanavond, of morgen om half één. Schikt dat?”
„Hij komt vanavond,” zeide Betsy, met de haar eigen beslistheid.

Thuiskomende zag Betsy de deur van van Vleuten’s kantoortje open
staan. Meenende, dat zij vergeten had die te sluiten, wilde zij haar
verzuim herstellen, toen zij bemerkte, dat haar man thuis was.
„Ben je al thuis?” riep zij uit. „Wat lees je daar?”
„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.”
„Eerst naar het kantoor van Wiechen, en toen naar je advocaat,”

zeide zij. [143]
„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.”
Betsy had van haar zenuwen gevergd wat ze verdragen konden. Nu

was bet uit met haar geestkracht, en het duurde lang, eer Van
Vleuten een eenigzins geregeld verslag van haar wedervaren kon
krijgen. Toen betrok zijn gezicht.
En nu bekende hij zelf al sinds geruimen tijd ongerust te zijn

geweest. Ook dat hij feitelijk niet had durven handelen, uit vrees dat
er een débacle uit voortkomen zou. Toch was hij blij, dat Betsy het
nu gedaan had. Immers nu wisten zij, dat Wiechen een geraffineerde
schelm was, maar ook op welke manier zij de dupe van hem waren.
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.
„Maar iedere dag is er toch één, in zulke gevallen,” merkte Van

Vleuten op. „De man leeft, en voort zelfs een zekeren train de vivre,
die hem geld moet kosten. In Indië kennen we dat van de
Chineezen. Men kan dan zeker zijn, dat zij bezig zijn geld en
goederen te verduisteren. Dan is bij ons het wachtwoord: haast
maken; en als de Javasche Bank het niet doet, slaan gewoonlijk een
paar lui de handen ineen, ter verdeeling van de gerechtskosten, en
vragen zoo spoedig mogelijk faillissement aan. Dan ligt tenminste
alles vast; want wat eenmaal weg is, komt nooit meer terecht.”
„Verduistering op die manier is strafbaar …”
„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.”
„Wel,” meende de advocaat, „als ik u dan goed begrijp, ziet u tegen

snel handelen niet op. Ik geloof vast, dat hij [145]geld geborgen heeft,
en misschien is de tijd nog niet rijp, dat het faillissement, waar hij op
aan schijnt te sturen, hem nu reeds welkom zou zijn. Dan zou er met
hem te accordeeren zijn. Ik wil dat voor u probeeren, maar heb me
altijd het best bevonden bij de methode om eerst eenige
gerechtelijke stappen te doen, en dan te gaan praten. In dit geval
zou ik u in overweging geven uw deposito reeds morgen te doen
opzeggen. Daar kan hij alvast niet van af. Mogelijk geeft hij wel
geluid, en als hij eenig voorstel doet, zal ik u er dadelijk kennis van
geven …”
„Kunt u geen beslag laten leggen?” vroeg Van Vleuten.

„Op uw vordering niet, daar die niet liquide is. Maar ik zal eens
rondkijken, of er niet een ander te vinden is, waarmee we hem dan
zoolang ophouden, tot ook de uwe opeischbaar is. Veel zou het
waard zijn, als we erachter konden komen, waar hij zijn geld gestopt
heeft.”
„Ik weet, dat hij naar Parijs is geweest.”
„Dat wordt wanhopig zoeken, als hij het daar heeft.
„Enfin, laat mij maar eerst eens probeeren, en mocht u iets

wetenswaardigs ontdekken, dan houd ik mij voor onmiddellijk bericht
aanbevolen.”
Vrij mistroostig kwam van Vleuten thuis. De zaak hinderde hem te

meer, daar hij gevoelde dat het zijn eigen schuld was.
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.
Zonder Betsy’s aandringen op sparen, zou hij nu geheel zonder

middelen zijn. Hij had evengoed kunnen speculeeren in den blinde.
Misschien zou hem de kans dan gunstiger [146]geweest zijn, daar bij
dan alleen met het toeval had te rekenen, terwijl hier boos opzet
geen gering woord had meegesproken.
Plotseling bedacht hij geheel vergeten te hebben over de zaak met

Boom te sproken. Bah! Eén bende, had zijn advocaat tot Betsy
gezegd. Toch interesseerde het hem, hoe Boom zich zou
verantwoorden. Iets moest hij toch zeggen, en vooral hij, die zoo
soeverein hoog over zijn eigen fatsoen sprak. Het was wat laat …
hoewel, Boom was iemand, die ’s avonds in een koffiehuis wel te
vinden was. In C e n t r a l bijvoorbeeld. Daar scheen alles van dat
slag bijeen te komen. Van Vleuten herinnerde zich de uitdrukking
„boevenkroeg”, bij de studenten in gebruik, voor de verzamelplaats
van niet-corpsleden. Geen kwade variant!
Al in zichzelf mopperend, was hij van richting veranderd, en kwam

ten slotte bij het genoemde café. In zoover trof hij het, daar de man
dien hij zocht, er was, toekijkend bij de biljarten.
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.
„Meneer Boom, een oogenblikje alsublieft.”
De aangesprokene toonde zich onmiddellijk bereid, en volgde van

Vleuten naar een hoektafeltje. Op van Vleutens directe vraag naar
den stand van zaken, noodigde hij dezen uit eens te zijnent te
komen. Hij zou hem dan inzage geven van de transacties, tot nog
toe door hem voor gezamenlijke rekening gedaan.
„Jawel,” zeide van Vleuten, „maar de hoofdzaak, uw machine?” [147]
„Ja,” meende Boom, „een mensch kan maar één ding tegelijk doen.
Ik heb het met de zaken zóó druk gehad …”
„Dat zou tijdelijk zijn, en moest toch bijzaak blijven.”
„Goed—maar u heeft niettemin uw deel opgestreken. Ik geloof dat u

zelfs nagenoeg al uw geld terug hebt.”
„Terug? O, als u bedoelt … Maar neen. U heeft mij in het geheel, als
ik me wel herinner zevenhonderd tachtig gulden als winst
afgedragen.”
„Nu, dat is dan op tweehonderd twintig gulden na uw deel van het

kapitaal,” zeide Boom. „Die kunt u krijgen, als ik ze binnen heb.”
Van Vleuten keek den ander aan. Was hij dronken? Dat scheen
overigens toch niet.
„U is in de war,” zeide hij. „Ik stelde u twee mille terhand, om uw

model te maken, en.…”
„Nu ja, maar zoolang ik daar niet aan beginnen kon, zouden we het
geld gezamenlijk uitzetten.”
„Dat klopt. Ik heb u ook niet gedrongen.…”
„Welnu, ik heb dus ieder voor de helft, dat is duizend gulden

geboekt, en daarvan heeft u zevenhonderd tachtig terug genomen.
Ik heb uw ontvangbewijsjes bij elkaar.”
„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.”
„Nu begrijp ik u niet. We zouden de zaak toch samen doen? Dat is

volgens mij, elk de helft. U kunt mijn boek zien: zooveel als u nam,
heb ik ook genomen. Wilt u dus uitscheiden, dan geeft ik u
tweehonderd twintig gulden, zoodra ik ze los heb, en ’t is
afgeloopen.” [148]
„Inderdaad, dat is een aardige rekening,” zeide van Vleuten, met
geweld zijn kalmte bewarend.
„Wat wilt u? Ik houd me aan de afspraak. Samen, dus ieder de helft.

U schijnt daar van af te willen …”
„Er heerscht een misverstand. Wanneer is het model van uw

machine klaar, denkt u?”
„Dat zeide ik u reeds. Ik heb het met de geldzaken te druk gehad,

om daar wat aan te kunnen doen.”
„Dat was toch de bedoeling niet. Enfin, laat dan die geldzaken
rusten, en werk uw machine af.”
„Waarvan?”
„Wel, van de twee mille, die u daarvoor gekregen hebt.”
„Daar hebben we elk zevenhonderd tachtig gulden van gehad.”
„Parbleu! Om die er af te nemen, heeft u toch geen drukte bezorgd?”
„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.”
Van Vleuten bleef staan, planté là, zooals de franschen zeggen.
Eindelijk draaide hij zich om, en ging heen, voornemens morgen

Boom schriftelijk te beduiden, dat hij hem het geld voor een bepaald
doel had gegeven, en niet wenschte, dat er langer op deze wijze
mee gewerkt werd.
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.
Dr. Arnolds zat in zijn studeerkamer, toen Boom werd aangediend.
„Wat mankeert eraan?” vroeg hij eenigszins verbaasd, den op het

oog volmaakt gezonden jongen man aanziend.
„Ik wou gaan trouwen,” zeide Boom.
„O, en u wenscht een geneeskundig onderzoek? Ja, dat doet men

tegenwoordig veel, en ik kan het niet kwaad vinden.”
„Pardon, dokter. Maar trouwen kost geld.”
„Dat zult u toch niet bij mij zoeken?” vroeg de oude man treurig.
„Toch wel.” verklaarde Boom. „U is, naar mij destijds uw zoon

vertelde, penningmeester van een dames-vereeniging, die gelden
voorschiet, als men … nu, ’n zoogenaamde fout repareeren wil. Is
dat zoo?”
„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.”
„Zoo! Nu, ik denk niet dat er moeielijkheid zal zijn. De vereeniging

heeft in langen tijd niets te doen gehad. U begrijpt … men meldt zich
niet gaarne aan.” [150]
„Neen, dat begrijp ik niet,” zeide Boom. „Ik zou eer denken, dat als
men het maar meer algemeen wist, de aanvragen niet zouden
ontbreken.”
„Integendeel. Het kost soms vrij wat moeite de mannen te bewegen,

om de meisjes te trouwen, die zij ongelukkig gemaakt hebben.”
„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.
„Foei, meneer Boom! Dat zou een immoraliteit zijn.”
„Wat? Het te doen, of er een premie op stellen?”
„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?”
„Zoo spoedig mogelijk, dokter. Ik wacht erop om aan te teekenen.”
„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?”
„Het is natuurlijk een zaak van vertrouwen,” zeide de oude dokter,

toen hij teekende. „Trouwt u niet, dan krijg ik het accept of het geld
terug. Overigens heeft u vijf jaar den tijd.…
„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?”
„O, best! Hij wordt vrijgesproken. Waarschijnlijk zal de instructie

reeds de volgende week zijn beëindigd, en er is veel hoop, dat hij
dan al losgelaten wordt.”
„Dat is een goed bericht,” vond Boom. „Dus komen de jongelui nu

eerstdaags vrij?”
„De jongelui? Neen, Viehof natuurlijk niet.”
„U zegt.…? Hoe zit dat dan in elkaar? Medunkt, ze hebben het

allebei gedaan, of.…”
„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

Classical Hopf Algebras and Their Applications Algebra and Applications 29 Pierre Cartier Frederic Patras Online Ebook Texxtbook Full Chapter PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Classical Hopf Algebras and Their Applications Algebra and Applications 29 Pierre Cartier Frederic Patras Online Ebook Texxtbook Full Chapter PDF

Uploaded by

Copyright:

Available Formats

Classical Hopf Algebras and Their

Applications Algebra and Applications

Algebra and Applications 2 Combinatorial Algebra and

Algebra and Applications 1: Non-associative Algebras

Chess Explained The c3 Sicilian 1st Edition Sam Collins

Starting Out The c3 Sicilian 1st Edition John Emms

Algebra 4: Lie Algebras, Chevalley Groups, and their

The American Revolution 1774 83 2nd Edition Daniel

BNF 83 (British National Formulary) March 2022 Joint

English Grammar Exercises with answers Part 4 Your

More information about this series at http://www.springer.com/series/6253

Classical Hopf Algebras

ISSN 1572-5553 ISSN 2192-2950 (electronic)

Mathematics Subject Classiﬁcation: 16T05, 16S30, 16T15, 16T30, 16W10

© Springer Nature Switzerland AG 2021

Limours, France Pierre Cartier

Comk (resp., Com): category of commutative unital algebras.

Linck (resp., Linc ): category of complete ﬁltered vector spaces.

Part I General Theory

3.3 Characteristic Endomorphisms and the Dynkin Operator . . . . . . 48

Appendix A: Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247

The main purpose of this volume is to give a modern, up-to-date, presentation of

m(λ ⊗ κ)(g) = (λ · κ)(g) := λ(g)κ(g)

for λ, κ ∈ k G and g ∈ G, and is equipped with a linear map (the coproduct)

where δx stands for the delta function on G, δx (g) = 1 if g = x and 0 else.

1.4 Natural Endomorphisms

1.6 Structure of the Book

Part I of the book addresses the general theory.

2.1 Preliminaries on Vector Spaces and Algebras

Let V, W be two vector spaces, V ∗ and W ∗ their duals. For v∗ ∈ V ∗ , v ∈ V , we

(w ⊗ v∗ )(v ) := w· < v∗ |v >

is an isomorphism when V is finite dimensional. In particular, when V = W , the

One advantage of this diagrammatic definition of algebras is that it generalizes

© Springer Nature Switzerland AG 2021 9

2.2 Coalgebras: Definition and First Properties

We will very often use the abbreviated Sweedler notation

(c) =: c(1) ⊗ c(2) .

Diagrammatically, the coproduct is coassociative if the following diagram com-

We denote the map ( ⊗ I d) ◦ by 3 . When the coproduct is coassociative,

the coassociativity hypothesis implying that for any i ≤ n − 2,

More generally, one gets by induction

for n 1 , . . . , n k > 0, with the notation 1 := I d.

where T is the switch map T (a ⊗ b) := b ⊗ a. In the Sweedler notation,

c(1) ⊗ c(2) = c(2) ⊗ c(1)

In general, T ◦ defines a new coproduct on C, the opposite coproduct.

where we use the isomorphisms k ⊗ C ∼

Definition 2.2.1 A coalgebra is a vector space C equipped with a coassociative

We will call noncounital coalgebra a k-vector space C equipped with a coasso-

defines a coalgebra structure on C.

Given two coalgebras C and D, equipped, respectively, with coproducts C ,

Remark 2.2.1 Notice that if (C, , ε) is a coalgebra, (C, T ◦ , ε) is also a coal-

Lemma 2.2.1 (Dual of a coalgebra) Let C be a coalgebra, its dual C ∗ is naturally

Proof Indeed, there is a canonical embedding C ∗ ⊗ C ∗ → (C ⊗ C)∗ , defined by

The product on C ∗ is written as a convolution product,

where m k is multiplication in k. It is an instance of a more general construction

Remark 2.2.2 When C is finite dimensional, C = (C ∗ )∗ and there is a perfect dual-

C⊗D := (I dC ⊗ TC,D ⊗ I d D ) ◦ (C ⊗ D ),

where TC,D (c ⊗ d) := d ⊗ c, and the counit

Remark 2.2.3 In the category of cocommutative coalgebras, the tensor product of

where the isomorphism is obtained from the composition of maps

is the coproduct of A and B:

where the isomorphism is obtained from the composition of maps

Definition 2.2.2 A subcoalgebra of a coalgebra C is a subvector space D of C such

Proposition 2.8.1 The map D → D̃ is a bijection between the set of subcoalgebras

g ∈ G → π(g) ∈ End(V ), π(g)(v) := (I dV ⊗ g) ◦ δ(v),