STUDENT MATHEMATICAL LIBRARY

Volume 59

Introduction to
Representation Theory
Pavel Etingof
Oleg Golberg
Sebastian Hensel
Tiankai Liu
Alex Schwendner
Dmitry Vaintrob
Elena Yudovina

with historical interludes by

Slava Gerovitch

AMS
American Mathematical Society
Providence, Rhode Island
 Editorial Board
Gerald B. Folland Brad G. Osgood (Chair)
Robin Forman John Stillwell
2010 Mathematics Subject Classification. Primary 16Gxx, 2OGxx.

For additional information and updates on this book, visit

www.ams.org/bookpages/stml-59
Library of Congress Cataloging-in-Publication Data
Introduction to representation theory / Pavel Etingof . . . [et al.] ; with historical
interludes by Slava Gerovitch.
p. cm. — (Student mathematical library ; v. 59)
Includes bibliographical references and index.
ISBN 978-0-8218-5351-1 (alk. paper)
1. Representations of algebras. 2. Representations of groups. I. Etingof, P. I.
(Pavel I.), 1969­
c;A1551586 2011
512A46——dc22
2011004787

Copying and reprinting. Individual readers of this publication, and nonprofit

libraries acting for them, are permitted to make fair use of the material, such as to
copy a chapter for use in teaching or research. Permission is granted to quote brief
passages from this publication in reviews, provided the customary acknowledgment of
the source is given.
Republication, systematic copying, or multiple reproduction of any material in this
publication is permitted only under license from the American Mathematical Society.
Requests for such permission should be addressed to the Acquisitions Department,
American Mathematical Society, 201 Charles Street, Providence, Rhode Island 02904­
2294 USA. Requests can also be made by e-mail to reprint-permissionthams.org.
© 2011 by the American Mathematical Society. All rights reserved.
The American Mathematical Society retains all rights
except those granted to the United States Government.
Printed in the United States of America.
The paper used in this book is acid-free and falls within the guidelines
established to ensure permanence and durability.
Visit the AMS home page at http: //www.ams.org/
10987654321 161514131211
Contents

Chapter 1. Introduction
Chapter 2. Basic notions of representation theory
§2.1. What is representation theory?
§2.2. Algebras
§2.3. Representations CDOOCHCJ1

§2.4. Ideals 15

§2.5. Quotients 15

§2.6. Algebras defined by generators and relations 16

§2.7. Examples of algebras 17
§2.8. Quivers 19

§2.9. Lie algebras 22

§2.10. Historical interlude: Sophus Lie’s trials and
transformations 26
§2.11. Tensor products 30
§2.12. The tensor algebra 35
§2.13. Hilbert’s third problem 36
§2.14. Tensor products and duals of representations of Lie
algebras 36
§2.15. Representations of s[(2) 37

Ill
iv Contents
§2.16. Problems on Lie algebras 39
Chapter 3. General results of representation theory 41
§3.1. Subrepresentations in semisimple representations 41
§3.2. The density theorem 43
§3.4. Filtrations 45
§3.3. Representations of direct sums of matrix algebras 44

§3.5. Finite dimensional algebras 46

§3.6. Characters of representations 48
§3.7. The Jordan-Holder theorem 50
§3.9. Problems 53
§3.8. The Krull-Schmidt theorem 51
§3.10. Representations of tensor products 56
Chapter 4. Representations of finite groups: Basic results 59
§4.1. Maschke’s theorem 59
§4.2.
§4.3. Characters
Examples 6261
§4.4. Duals and tensor products of representations 65
§4.5. Orthogonality of characters 65
§4.6. Unitary representations. Another proof of Maschke’s
theorem for complex representations 68

tables 74
§4.7. Orthogonality of matrix elements 70
§4.8. Character tables, examples 71
§4.9. Computing tensor product multiplicities using character

§4.10. Frobenius determinant 75

Horse Trade” 77
§4.11. Historical interlude: Georg Frobenius’s “Principle of

§4.12. Problems 81
poetry 86
§4.13. Historical interlude: William Rowan Hamilton’s
quaternion of geometry, algebra, metaphysics, and
Contents v
Chapter 5. Representations of finite groups: Further results 91
§5.1. Frobenius-Schur indicator 91
§5.2. Algebraic numbers and algebraic integers 93
§5.3. Frobenius divisibility 96
§5.4. Burnside’s theorem 98
§5.5. Historical interlude: William Burnside and intellectual
harmony in mathematics 100
§5.6. Representations of products 104
§5.7. Virtual representations 105
§5.8. Induced representations 105
representation 106
§5.9. The Frobenius formula for the character of an induced

§5.10. Frobenius reciprocity 107

§5.11. Examples 110
of Sn 112
§5.12. Representations of Sn 110
§5.13. Proof of the classification theorem for representations

§5.14.. Induced representations for Sn 114

§5.16. Problems 118
§5.15. The Frobenius character formula 115
§5.17. The hook length formula 118
§5.18. Schur-Weyl duality for g[(V) 119
§5.19. Schur-Weyl duality for GL(V) 122
§5.20. Historical interlude: Hermann Weyl at the intersection
of limitation and freedom 122
§5.21. Schur polynomials 128
§5.22. The characters of L A 129
§5.24. Problems 131
§5.23. Algebraic representations of GL(V) 130

§5.25. Representations of GL2 (lFq) 132

§5.26. Artin’s theorem 141
vi Contents

§5.27. Representations of semidirect products 142

Chapter 6. Quiver representations 145

§6.1. Problems 145
§6.2. Indecomposable representations of the quivers
A1, A2, A3 150
§6.3. Indecomposable representations of the quiver D4 154
§6.4. Roots 160
§6.5. Gabriel’s theorem 163
§6.6. Reflection functors 164
§6.7. Coxeter elements 169
§6.8. Proof of Gabriel’s theorem 170
§6.9. Problems 173

Chapter 7. Introduction to categories 177

§7.1. The definition of a category 177
§7.2. Functors 179
§7.3. Morphisms of functors 181
§7.4. Equivalence of categories 182
§7.5. Representable functors 183
§7.6. Adjoint functors 184
§7.7. Abelian categories 186
§7.8. Complexes and cohomology 187
§7.9. Exact functors 190
§7.10. Historical interlude: Eilenberg, Mac Lane, and
“general abstract nonsense” 192

Chapter 8. Homological algebra 201

§8.1. Projective and injective modules 201
§8.2. Tor and Ext functors 203

Chapter 9. Structure of finite dimensional algebras 209

§9.1. Lifting of idempotents 209
§9.2. Pro jective covers 210
Contents vii
§9.3. The Cartan matrix of a finite dimensional algebra 211

§9.5. Blocks 213

§9.4. Homological dimension 212
§9.6. Finite abelian categories 214
§9.7. Morita equivalence 215
References for historical interludes 217
Mathematical references 223
Index 225
Chapter 1

Introduction

Very roughly speaking, representation theory studies symmetry in

linear spaces. It is a beautiful mathematical subject which has many
applications, ranging from number theory and combinatorics to ge­
ometry, probability theory, quantum mechanics, and quantum field
theory.
Representation theory was born in 1896 in the work of the Ger­
man mathematician F. G. Frobenius. This work was triggered by a
letter to Frobenius by R. Dedekind. In this letter Dedekind made the
following observation: take the multiplication table of a finite group
G’ and turn it into a matrix XG by replacing every entry g of this
table by a variable :59. Then the determinant of XG factors into a
product of irreducible polynomials in {mg}, each of which occurs with
multiplicity equal to its degree. Dedekind checked this surprising fact
in a few special cases but could not prove it in general. So he gave
this problem to Frobenius. In order to find a solution of this problem
(which we will explain below), Frobenius created the representation
theory of finite groups.
The goal of this book is to give a “holistic” introduction to rep­
resentation theory, presenting it as a unified subject which studies
representations of associative algebras and treating the representa­
tion theories of groups, Lie algebras, and quivers as special cases. It
is designed as a textbook for advanced undergraduate and beginning

1
2 1 . Introduction
graduate students and should be accessible to students with a strong
background in linear algebra and a basic knowledge of abstract al­
gebra. Theoretical material in this book is supplemented by many
problems and exercises which touch upon a lot of additional topics;
the more difficult exercises are provided with hints.
The book covers a number of standard topics in representation
theory of groups, associative algebras, Lie algebras, and quivers. For
a more detailed treatment of these topics, we refer the reader to the
textbooks [S], and [CR]. We mostly follow [FH], with the
exception of the sections discussing quivers, which follow [BGP], and
the sections on homological algebra and finite dimensional algebras,
for which we recommend and [CR], respectively.
The organization of the book is as follows.
Chapter 2 is devoted to the basics of representation theory. Here
we review the basics of abstract algebra (groups, rings, modules,
ideals, tensor products, symmetric and exterior powers, etc.), as well
as give the main definitions of representation theory and discuss the
objects whose representations we will study (associative algebras,
groups, quivers, and Lie algebras).
Chapter 3 introduces the main general results about representa­
tions of associative algebras (the density theorem, the Jordan—H6lder
theorem, the Krull—Schmidt theorem, and the structure theorem for
finite dimensional algebras).
In Chapter 4 we discuss the basic results about representations of
finite groups. Here we prove Maschke’s theorem and the orthogonality
of characters and matrix elements and compute character tables and
tensor product multiplicities for the simplest finite groups. We also
discuss the Frobenius determinant, which was a starting point for
development of the representation theory of finite groups.
We continue to study representations of finite groups in Chapter
5, treating more advanced and special topics, such as the Frobenius—
Schur indicator, the Frobenius divisibility theorem, the Burnside the­
orem, the Frobenius formula for the character of an induced repre­
sentation, representations of the symmetric group and the general
1. Introduction 3
linear group over (C, representations of GL2(IFq), representations of
semidirect products, etc.
In Chapter 6, we give an introduction to the representation theory
of quivers (starting with the problem of the classification of configura­
tions of n subspaces in a vector space) and present a proof of Gabriel’s
theorem, which classifies quivers of finite type.
In Chapter 7, We give an introduction to category theory, in par­
ticular, abelian categories, and explain how such categories arise in
representation theory.
In Chapter 8, We give a brief introduction to homological algebra
and explain how it can be applied to categories of representations.
Finally, in Chapter 9 we give a short introduction to the repre­
sentation theory of finite dimensional algebras.
Besides, the book contains six historical interludes Written by Dr.
Slava Gerovitch.1 These interludes, Written in an accessible and ab­
sorbing style, tell about the life and mathematical work of some of
the mathematicians who played a major role in the development of
modern algebra and representation theory: F. G. Frobenius, S. Lie,
W. Burnside, W. R. Hamilton, H. Weyl, S. Mac Lane, and S. Eilen—
berg. For more on the history of representation theory, we recommend
that the reader consult the references to the historical interludes, in
particular the excellent book [Cu].
Acknowledgments. This book arose from the lecture notes of
a representation theory course given by the first author to the re­
maining six authors in March 2004 within the framework of the Clay
Mathematics Institute Research Academy for high school students
and its extended version given by the first author to MIT undergrad­
uate mathematics students in the fall of 2008.
The authors are grateful to the Clay Mathematics Institute for
hosting the first version of this course. The first author is very in­
debted to Victor Ostrik for helping him prepare this course and thanks

11 wish to thank Prof. Pavel Etingof and his co—authors for adding technical notes
to my historical monograph. While they have made a commendable effort at a concise
exposition, their notes, unfortunately, have grown in size and in the end occupied
a better part of this volume. I hope the reader will forgive this preponderance of
technicalities in what, in essence, is a history book. — S. Gerovitch.
4 1. Introduction
Josh Nichols—Barrer and Thomas Lam for helping run the course in
2004 and for useful comments. He is also very grateful to Darij Grin­
berg for his very careful reading of the text, for many useful comments
and corrections, and for suggesting Problems 2.11.6, 3.3.3, 3.8.3, 3.8.4,
4.52, 5.10.2 and Exercises 5.27.2, 5.27.3, 7.98. Finally, the authors
gratefully acknowledge the use of the Dynkin diagram pictures pre­
pared by W. Casselman.
Chapter 2

Basic notions of
representation theory

2.1. What is representation theory?

In technical terms, representation theory studies representations of
associative algebras. Its general content can be very briefly summa­
rized as follows.
An associative algebra over a field k is a vector space A over
It equipped with an associative bilinear multiplication a,b +—> ab,
a,b E A. We will always consider associative algebras with unit,
i.e., with an element 1 such that 1 - a = a - 1 = a for all a E A. A
basic example of an associative algebra is the algebra EndV of linear
operators from a vector space V to itself. Other important examples
include algebras defined by generators and relations, such as group
algebras and universal enveloping algebras of Lie algebras.
A representation of an associative algebra A (also called a left
A-module) is a vector space V equipped with a homomorphism p :
A —> EndV, i.e., a linear map preserving the multiplication and unit.
A subrepresentation of a representation V is a subspace U C V
which is invariant under all operators p(a), a E A. Also, if V1, V2 are
two representations of A, then the direct sum V1 69 V2 has an obvious
structure of a representation of A.

5
6 2. Basic notions of representation theory
A nonzero representation V of A is said to be irreducible if its
only subrepresentations are 0 and V itself, and it is said to be inde­
composable if it cannot be written as a direct sum of two nonzero
subrepresentations. Obviously, irreducible implies indecomposable,
but not vice versa.
Typical problems of representation theory are as follows:
(1) Classify irreducible representations of a given algebra A.
(2) Classify indecomposable representations of A.
(3) Do (1) and (2) restricting to finite dimensional representa­
tions.
As mentioned above, the algebra A is often given to us by gener­
ators and relations. For example, the universal enveloping algebra U
of the Lie algebra 5l(2) is generated by h, e, f with defining relations
(2.1.1) he—eh=2e, hf—fh=—2f, ef—fe=h.
This means that the problem of finding, say, N -dimensional represen­
tations of A reduces to solving a bunch of nonlinear algebraic equa­
tions with respect to a bunch of unknown N X N matrices, for example
system (2.1.1) with respect to unknown matrices h, e, f.
It is really striking that such, at first glance hopelessly compli­
cated, systems of equations can in fact be solved completely by meth­
ods of representation theory! For example, we will prove the following
theorem.

Theorem 2.1.1. Let k = (C be the field of complex numbers. Then:

( i ) The algebra U has exactly one irreducible representation Vd of
each dimension, up to equivalence; this representation is realized in

0 0 0 ('3
the space of homogeneous polynomials of two variables :13, y of degree
d — 1 and is defined by the formulas

901) - wa - .715; 9(6) — wa—y, p(f) — .7455­

( ii ) Any indecomposable finite dimensional representation of U is
irreducible. That is, any finite dimensional representation of U is a
direct sum of irreducible representations.

As another example consider the representation theory of quivers.

2.1. What is representation theory? 7
A quiver is an oriented graph Q (which we will assume to be
finite). A representation of Q over a field k is an assignment of
a kt-vector space Vi to every vertex i of Q and of a linear operator
Ah : Vi ——> to every directed edge h going from i to j (loops and
multiple edges are allowed). We will show that a representation of a
quiver Q is the same thing as a representation of a certain algebra
PQ called the path algebra of Q. Thus one may ask: what are the
indecomposable finite dimensional representations of Q?
More specifically, let us say that Q is of finite type if it has
finitely many indecomposable representations.
We will prove the following striking theorem, proved by P. Gabriel
in early 1970s:
Theorem 2.1.2. The finite type property of Q does not depend on
the orientation of edges. The connected graphs that yield quivers of
finite type are given by the following list:
0 An :
o——o——o-----0:010
0 Dn:
o——o——o-----0:0

0 E5
o——o—-o—-0:0
O

0 E7
O:-O:—O:"O:—OjO
O

0 E3
o—-o:o:o:o:o:o
8 2. Basic notions of representation theory
The graphs listed in the theorem are called (simply laced) Dyn­
kin diagrams. These graphs arise in a multitude of classification
problems in mathematics, such as the classification of simple Lie al­
gebras, singularities, platonic solids, reflection groups, etc. In fact, if
we needed to make contact with an alien civilization and show them
how sophisticated our civilization is, perhaps showing them Dynkin
diagrams would be the best choice!
As a final example consider the representation theory of finite
groups, which is one of the most fascinating chapters of represen­
tation theory. In this theory, one considers representations of the
group algebra A = <C[G] of a finite group G — the algebra with basis
ag, g E G, and multiplication law agah = agh. We will show that any
finite dimensional representation of A is a direct sum of irreducible
representations, i.e., the notions of an irreducible and indecompos­
able representation are the same for A (Maschke’s theorem). Another
striking result discussed below is the Frobenius divisibility theorem:
the dimension of any irreducible representation of A divides the or­
der of G. Finally, we will show how to use the representation theory
of finite groups to prove Burnside’s theorem: any finite group of or­
der paqb, where p, q are primes, is solvable. Note that this theorem
does not mention representations, which are used only in its proof; a
purely group-theoretical proof of this theorem (not using representa­
tions) exists but is much more difficult!

2.2. Algebras
Let us now begin a systematic discussion of representation theory.
Let kt be a field. Unless stated otherwise, we will always assume
that kt is algebraically closed, i.e., any nonconstant polynomial with
coefficients in kt has a root in kt. The main example is the field of
complex numbers (C, but we will also consider fields of characteristic
p, such as the algebraic closure E, of the finite field IFP of p elements.

Definition 2.2.1. An associative algebra over kt is a vector space

A over kt together with a bilinear map A X A —> A, (a, b) +—> ab, such
that (ab)c = a(bc).
2.3. Representations 9
Definition 2.2.2. A unit in an associative algebra A is an element
1 E A such that 1a = a1 = a.

Proposition 2.2.3. If a unit exists, it is unique.

Proof. Let 1, 1’ be two units. Then 1 = 11’ = 1’. C]
From now on, by an algebra A We will mean an associative algebra
with a unit. We Will also assume that A yé 0.
Example 2.2.4. Here are some examples of algebras over k:
1. A = k.
2. A = k[:z:1, . . . ,:z:.,,] — the algebra of polynomials in variables
51:1, . . . , sun.

3. A =: EndV — the algebra of endomorphisms of a vector space

V over kt (i.e., linear maps, or operators, from V to itself). The
multiplication is given by composition of operators.
4. The free algebra A = /c(:1:1, . . . ,:n.n). A basis of this algebra
consists of Words in letters :31, . . . ,:z:.,,, and multiplication in this basis
is simply the concatenation of Words.
5. The group algebra A = k[G] of a group G. Its basis is
{ag, g E G}, With multiplication law agah = agh.
Definition 2.2.5. An algebra A is commutative if ab = ba for all
a, b E A.

For instance, in the above examples, A is commutative in cases 1

and 2 but not commutative in cases 3 (if dim V > 1) and 4 (if n > 1).
In case 5, A is commutative if and only if G is commutative.
Definition 2.2.6. A homomorphism of algebras f : A —> B is a
linear map such that f(:z:y) = f(:z:)f(y) for all 51:, y E A and f(1) = 1.

2.3. Representations
Definition 2.3.1. A representation of an algebra A (also called a
left A-module) is a vector space V together with a homomorphism
of algebras p : A —> EndV.
10 2. Basic notions of representation theory
Similarly, a right A-module is a space V equipped with an
antihomomorphism p : A —> EndV; i.e., p satisfies p(ab) = p(b)p(a)
and p(1) =2 1.
The usual abbreviated notation for p(a)?) is av for a left mod­
ule and va for a right module. Then the property that p is an
(anti)homomorphism can be written as a kind of associativity law:
(ab)'v 2- a(bv) for left modules, and (va)b = 'v(ab) for right modules.
Remark 2.3.2. Let M be a left module over a commutative ring
A. Then one can regard M as a right A-module, with ma 2: am.
Similarly, any right A-module can be regarded as a left A-module. For
this reason, for commutative rings one does not distinguish between
left and right A-modules and just calls them A-modules.

Here are some examples of representations.

Example 2.3.3. 1. V = O.
2. V = A, and p : A —> EndA is defined as follows: p(a) is the
operator of left multiplication by a, so that p(a)b :— ab (the usual
product). This representation is called the regular representation of
A. Similarly, one can equip A with a structure of a right A-module
by setting p(a)b := ba.
3. A = k. Then a representation of A is simply a vector space
over k.
4. A = k(:z:1, . . . ,:I:.,,). Then a representation of A is just a vec­
tor space V over k with a collection of arbitrary linear operators
p(:z:1), . . . ,p(:z:.n) : V —> V (explain why!).
Definition 2.3.4. A subrepresentation of a representation V of
an algebra A is a subspace W C V which is invariant under all the
operators p(a) : V —> V, a E A.

For instance, 0 and V are always subrepresentations.

Definition 2.3.5. A representation V 7E O of A is irreducible (or
simple) if the only subrepresentations of V are 0 and V.

Definition 2.3.6. Let V1, W be two representations of an algebra A.

A homomorphism (or intertwining operator) qb : V1 ——> V2 is a
2.3. Representations 1 1
linear operator which commutes with the action of A, i.e., ¢(a'u) =
aqb('v) for any '0 6 V1. A homomorphism qb is said to be an isomor­
phism of representations if it is an isomorphism of vector spaces.
The set (space) of all homomorphisms of representations V1 ——> V2 is
denoted by HomA(V1, V2).
Note that if a linear operator qb : V1 —> V2 is an isomorphism of
representations, then so is the linear operator ¢‘1 : V2 —> V1 (check
it!).
Two representations between which there exists an isomorphism
are said to be isomorphic. For practical purposes, two isomorphic
representations may be regarded as “the same”, although there could
be subtleties related to the fact that an isomorphism between two
representations, when it exists, is not unique.
Definition 2.3.7. Let V1,V2 be representations of an algebra A.
Then the space V1 613 V2 has an obvious structure of a representation
of A, given by a(v1 EB '02) = av1 EB av2. This representation is called
the direct sum of V1 and lé.
Definition 2.3.8. A nonzero representation V of an algebra A is said
to be indecomposable if it is not isomorphic to a direct sum of two
nonzero representations.
It is obvious that an irreducible representation is indecomposable.
On the other hand, we will see below that the converse statement is
false in general.
One of the main problems of representation theory is to classify
irreducible and indecomposable representations of a given algebra up
to isomorphism. This problem is usually hard and often can be solved
only partially (say, for finite dimensional representations). Below we
will see a number of examples in which this problem is partially or
fully solved for specific algebras.
We will now prove our first result — Schur’s lemma. Although
it is very easy to prove, it is fundamental in the whole subject of
representation theory.
Proposition 2.3.9 (Schur’s lemma). Let V1, V2 be representations of
an algebra A over any field F (which need not be algebraically closed).
12 2. Basic notions of representation theory
Let qb : V1 —> V2 be a nonzero homomorphism of representations.
Then:
If V1 is irreducible, qb is injective.
If V2 is irreducible, qb is surjective.
Thus, if both V1 and V2 are irreducible, qb is an isomorphism.

Proof. The kernel K of qb is a subrepresentation of V1. Since

I = V2. Cl
qb 75 0, this subrepresentation cannot be V1. So by irreducibility of V1
we have K = 0.
(ii) The image I of qb is a subrepresentation of V2. Since qb 75 0,
this subrepresentation cannot be 0.’ So by irreducibility of V2 We have

Corollary 2.3.10 (Schur’s lemma for algebraically closed fields). Let

V be a finite dimensional irreducible representation of an algebra A
over an algebraically closed field k, and let qb : V ——> V be an inter­
twining operator. Then 45 = A - Id for some A E k ( a scalar operator).
Remark 2.3.11. Note that this corollary is false over the field of
real numbers: it suffices to take A = (C (regarded as an ]R—algebra)
and V = A.

Proof. Let A be an eigenvalue of qb (a root of the characteristic poly­

nomial of ¢). It exists since k is an algebraically closed field. Then
the operator qb — A Id is an intertwining operator V ——> V, Which is not
an isomorphism (since its determinant is zero). Thus by Proposition
2.3.9 this operator is zero, hence the result. D
Corollary 2.3.12. Let A be a commutative algebra. Then every
irreducible finite dimensional representation V of A is 1—dimensional.
Remark 2.3.13. Note that a 1-dimensional representation of any
algebra is automatically irreducible.
Proof. Let V be irreducible. For any element a E A, the operator
p(a) 2 V ——> V is an intertwining operator. Indeed,
p(a)p(b)'v = p(ab)'v = p(ba)'v = p(b)p(a)'v
(the second equality is true since the algebra is commutative). Thus,
by Schur’s lemma, p(a) is a scalar operator for any a E A. Hence
2.3. Representations 13
(since V 74 0). Cl
every subspace of V is a subrepresentation. But V is irreducible, so
0 and V are the only subspaces of V. This means that dimV = 1

Example 2.3.14. 1. A = k. Since representations of A are simply

vector spaces, V = A is the only irreducible and the only indecom­
posable representation.
2. A = Since this algebra is commutative, the irreducible
representations of A are its 1-dimensional representations. As we
discussed above, they are defined by a single operator In the 1­
dimensional case, this is just a number from Is. So all the irreducible
representations of A are V,\ = k, A E k, in which the action of A
is defined by p(a:) = A. Clearly, these representations are pairwise
nonisomorphic.
The classification of indecomposable representations of k[a:] is
more interesting. To obtain it, recall that any linear operator on
a finite dimensional vector space V can be brought to Jordan nor­
mal form. More specifically, recall that the Jordan block J ,\,n is the
operator on k” which in the standard basis is given by the formulas
J,\,,,e,- = A6, + e,--1 for 2' > 1 and J,\,ne1 = A61. Then for any linear
operator B : V —> V there exists a basis of V such that the matrix
of B in this basis is a direct sum of Jordan blocks. This implies that
all the indecomposable representations of A are V,\,,, = k", A E k,
with p(a:) = J ,\,.,,. The fact that these representations are indecom­
posable and pairwise nonisomorphic follows from the Jordan normal
form theorem (which in particular says that the Jordan normal form
of an operator is unique up to permutation of blocks).
This example shows that an indecomposable representation of an
algebra need not be irreducible.
3. The group algebra A = k[G], where G is a group. A represen­
tation of A is the same thing as a representation of G, i.e., a vector
space V together with a group homomorphism p : G —> Aut(V),
where Aut(V) = GL(V) denotes the group of invertible linear maps
from the space V to itself (the general linear group of V).
14 2. Basic notions of representation theory
Problem 2.3.15. Let V be a nonzero finite dimensional representa­
tion of an algebra A. Show that it has an irreducible subrepresen­
tation. Then show by example that this does not always hold for
infinite dimensional representations.

Problem 2.3.16. Let A be an algebra over a field k. The center

Z (A) of A is the set of all elements 2: E A which commute with all
elements of A. For example, if A is commutative, then Z (A) = A.
(a) Show that if V is an irreducible finite dimensional representa­
tion of A, then any element 2: E Z (A) acts in V by multiplication by
some scalar X1/(z). Show that XV : Z (A) —> k is a homomorphism.
It is called the central character of V.
(b) Show that if V is an indecomposable finite dimensional rep­
resentation of A, then for any 2 E Z (A), the operator p(z) by which 2
acts in V has only one eigenvalue X1/(2), equal to the scalar by which 2:
acts on some irreducible subrepresentation of V. Thus XV : Z (A) —> k
is a homomorphism, which is again called the central character of V.
(c) Does p(z) in (b) have to be a scalar operator?

Problem 2.3.17. Let A be an associative algebra, and let V be

a representation of A. By End A(V) one denotes the algebra of all
homomorphisms of representations V —> V. Show that End A(A) =
A°P, the algebra A with opposite multiplication.

Problem 2.3.18. Prove the following “infinite dimensional Schur

lemma” (due to Dixmier): Let A be an algebra over C and let V
be an irreducible representation of A with at most countable basis.
Then any homomorphism of representations qb : V —> V is a scalar
operator.
Hint: By the usual Schur’s lemma, the algebra D 2: End A(V) is
an algebra with division. Show that D is at most countably dimen­
sional. Suppose (15 is not a scalar, and consider the subfield (C(q5) C D.
Show that (C(q5) is a transcendental extension of (C. Derive from this
that (C(q5) is uncountably dimensional and obtain a contradiction.
2.5. Quotients 15
2.4. Ideals
A left ideal of an algebra A is a subspace I Q A such that aI Q I
for all a E A. Similarly, a right ideal of an algebra A is a subspace
I Q A such that Ia Q I for all a E A. A two-sided ideal is a
subspace that is both a left and a right ideal.
Left ideals are the same as subrepresentations of the regular rep­
resentation A. Right ideals are the same as subrepresentations of the
regular representation of the opposite algebra A°P.
Below are some examples of ideals:
o If A is any algebra, O and A are two-sided ideals. An algebra
A is called simple if 0 and A are its only two-sided ideals.
o If (15 : A —+ B is a homomorphism of algebras, then kerqb is
a two-sided ideal of A.
o If S is any subset of an algebra A, then the two-sided ideal
generated by S is denoted by (S) and is the span of ele­
ments of the form asb, Where a, b E A and s E S. Similarly,
We can define (S)g = spa.n{as} and (S), = spa.n{sb}, the
left, respectively right, ideal generated by S.
Problem 2.4.1. A maximal ideal in a ring A is an ideal I gé A
such that any strictly larger ideal coincides with A. (This definition
is made for left, right, or two-sided ideals.) Show that any unital
ring has a maximal left, right, and two-sided ideal. (Hint: Use Zorn’s
lemma.)

2.5. Quotients
Let A be an algebra and let I be a two-sided ideal in A. Then A/ I
is the set of (additive) cosets of I. Let 7r : A —> A/ I be the quotient
map. We can define multiplication in A/ I by 7r(a) - 7r(b) := 7r(ab).
This is well defined because if 7r(a) = 7r(a’), then
7r(a'b) = 7r(ab + (a' — a)b) = 7r(ab) + 7r((a' — a)b) = 7r(ab)
because (a’ — a)b E I b Q I = ker 7r, as I is a right ideal; similarly, if
7r(b) = 7r(b’), then
7r(ab') = 7r(ab + a(b' — b)) = 7r(ab) + 7r(a(b' — b)) = 7r(ab)
16 2. Basic notions of representation theory
because a(b’ — b) E aI Q I = ker 7r, as I is also a left ideal. Thus, A/I
is an algebra.
Similarly, if V is a representation of A and W C V is a subrep­
resentation, then V/ W is also a representation. Indeed, let 7r : V —>
V/W be the quotient map, and set pv/W(a)7r(:z:) := 7r(pv(a):z:).
Above We noted that left ideals of A are subrepresentations of the
regular representation of A, and vice versa. Thus, if I is a left ideal
in A, then A/ I is a representation of A.

Problem 2.5.1. Let A = k[:z:1,...,:z:,,] and let I 75 A be any ideal

in A containing all homogeneous polynomials of degree 2 N. Show
that A / I is an indecomposable representation of A.

Problem 2.5.2. Let V 75 0 be a representation of A. We say that a

vector v E V is cyclic if it generates V, i.e., Av = V. A representation
admitting a cyclic vector is said to be cyclic. Show the following:
(a) V is irreducible if and only if all nonzero vectors of V are
cyclic.
(b) V is cyclic if and only if it is isomorphic to A / I , Where I is a
left ideal in A.
(c) Give an example of an indecomposable representation Which
is not cyclic.
Hint: Let A = (C[:z:,y]/I2, Where I2 is the ideal spanned by ho­
mogeneous polynomials of degree 2 2 (so A has a basis 1,:z:,y). Let
V = A* be the space of linear functionals on A, with the action of A
given by (p(a)f = f (ba). Show that V provides such an example.

2.6. Algebras defined by generators and

relations
If f1,...,fm are elements of the free algebra k(:z:1, . . . ,a:,,), We say
that the algebra A := k(:z:1, . . . ,:z:,,)/({f1, . . .,fm}) is generated by
51:1, . . . ,:z:,, with defining relations fl = 0, ..., fm = O.
2.7. Examples of algebras 17
2.7 . Examples of algebras
The following two examples are among the simplest interesting ex­
amples of noncommutative associative algebras:
(1) the Weyl algebra, k(:z:,y)/(ya: — my — 1);
(2) the q-Weyl algebra, generated by :13, :z:"1, y, y”1 with defin­
ing relations ya: = qczry and :z::z:’1 = :z:“1:z: = yy’1 = y’1y =
1.

Proposition 2.7.1. ( A basis for the Weyl algebra A is {criyj , i, j 2

0}.
A basis for the q- Weyl algebra Aq is {:cly7,i,j E Z}.
Proof. First let us show that the elements cztlyj are a spanning set
for A. To do this, note that any word in :13, y can be ordered to have
all the :z:’s on the left of the y’s, at the cost of interchanging some :1:
and y. Since ya: -133; = 1, this will lead to error terms, but these terms
will be sums of monomials that have a smaller number of letters :13, 3;
than the original word. Therefore, continuing this process, we can
order everything and represent any word as a linear combination of
cztlyj .

The proof that cztiyj are linearly independent is based on represen­

tation theory. Namely, let a be a variable, and let E = t“k[a][t, t“1]
(here t“ is just a formal symbol, so really E = k[a][t, t‘1]). Then E
is a representation of A with action given by :1: f = t f and y f = it
(where 0“ H H
d(t“+")
:= (a + 'n.)t“+"“1). Suppose now that we have a non­
trivial linear relation 2 c,-J-:z:'y9 = 0. Then the operator
idj
L = EC,'jt
acts by zero in E. Let us write L as
1' d j
L = Z c2.~<t> 5 .
j=0
where Q, 79 0. Then we have

Lt“ = Z362,-(t)a(a — 1) . . . (a — j + 1)t“"j.

i=0
18 2. Basic notions of representation theory
This must be zero, so we have 23:20 Q,-(t)a(a—1)...(a—j+1)t‘7' = 0
in k[a][t,t"1]. Taking the leading term in a, we get Q,.(t) = 0, a
contradiction.
(ii) Any word in :13, y, :13”, y‘1 can be ordered at the cost of mul­
tiplying it by a power of q. This easily implies both the spanning
property and the linear independence. Cl
Remark 2.7.2. The proof of shows that the Weyl algebra A can
be viewed as the algebra of polynomial differential operators in one
variable t.

The proof of also brings up the notion of a faithful represen­

tation.
Definition 2.7.3. A representation p : A —> End V of an algebra A
is faithful if p is injective.
For example, k[t] is a faithful representation of the Weyl algebra if
is has characteristic zero (check it!), but not in characteristic p, where
(d/dt)pQ = 0 for any polynomial Q. However, the representation
E = t“k[a][t, t”1], as we’ve seen, is faithful in any characteristic.
Problem 2.7.4. Let A be the Weyl algebra.
(a) If char k = 0, what are the finite dimensional representations
of A? What are the two-sided ideals in A?
Hint: For the first question, use the fact that for two square
matrices B,C', Tr(BC') = Tr(C'B). For the second question, show
that any nonzero two-sided ideal in A contains a nonzero polynomial
in :13, and use this to characterize this ideal.
Suppose for the rest of the problem that char k = p.
(b) What is the center of A?
Hint: Show that cup and yp are central elements.
(c) Find all irreducible finite dimensional representations of A.
Hint: Let V be an irreducible finite dimensional representation
of A, and let 1) be an eigenvector of y in V. Show that the collection
of vectors {v,:z:v, 1:22), . . . ,:z:p‘1v} is a basis of V.
Problem 2.7.5. Let q be a nonzero complex number, and let A be
the q-Weyl algebra over (C.
2.8. Quivers 19
(a) What is the center of A for different q? If q is not a root of
unity, what are the two—sided ideals in A?
(b) For which q does this algebra have finite dimensional repre­
sentations?
Hint: Use determinants.
(c) Find all finite dimensional irreducible representations of A for
such q.
Hint: This is similar to part (c) of the previous problem.

2.8. Quivers
Definition 2.8.1. A quiver Q is a directed graph, possibly with
self-loops and / or multiple edges between two vertices.
Example 2.8.2.

We denote the set of vertices of the quiver Q as I and the set

of edges as E. For an edge h E E, let h’, h” denote the source and
target of h, respectively:
C ——-% C
h/ h hll
Definition 2.8.3. A representation of a quiver Q is an assign­
ment to each vertex 2' E I of a vector space V, and to each edge h E E
of a linear map 13;, : Vhr —> Vhu.

It turns out that the theory of representations of quivers is a part

of the theory of representations of algebras in the sense that for each
quiver Q, there exists a certain algebra PQ, called the path algebra
of Q, such that a representation of the quiver Q is “the same” as
a representation of the algebra PQ. We shall first define the path
algebra of a quiver and then justify our claim that representations of
these two objects are “the same”.
 20 2. Basic notions of representation theory
Definition 2.8.4. The path algebra PQ of a quiver Q is the algebra
whose basis is formed by oriented paths in Q, including the trivial
paths pi, i E I, corresponding to the vertices of Q, and multiplication
is the concatenation of paths: ab is the path obtained by first tracing
b and then a. If two paths cannot be concatenated, the product is
defined to be zero.
Remark 2.8.5. It is easy to see that for a finite quiver 2 pi = 1, so
z'eI
PQ is an algebra with unit.
Problem 2.8.6. Show that the algebra PQ is generated by pi for
2' E I and ah for h E E with the following defining relations:
(1) P? = pi, pi-pi = 0 for i 75 J’,
(2) ahpru = ah, ahpj = 0 f01‘ J 7'9 h’,
(3) piinaii = ah, piaii = 0 for 2' 75 h”.
We now justify our statement that a representation of a quiver is
the same thing as a representation of the path algebra of a quiver.
Let V be a representation of the path algebra PQ. From this
representation, we can construct a representation of Q as follows: let
Vi = piV, and for any edge h, let sch = ahlpiirv I p;izV ——> p;i~V be
the operator corresponding to the one-edge path h.
Similarly, let (Vi, xii) be a representation of a quiver Q. From this
representation, we can construct a representation of the path algebra
PQ: let V = 6}, Vi, let pi : V —> Vi —> V be the projection onto Vi,
' and for any path p = h1...hm let ai, = :z:;i1...:z:;im : Viirm —> Viirlr be
the composition of the operators corresponding to the edges occurring
in p (and the action of this operator on the other Vi is zero).
It is clear that the above assignments V I—> (piV) and I—>
$1. Vi are inverses of each other. Thus, we have a bijection between
isomorphism classes of representations of the algebra PQ and of the
quiver Q.
Remark 2.8.7. In practice, it is generally easier to consider a rep­
resentation of a quiver as in Definition 2.8.3.
We lastly define several previous concepts in the context of quiver
representations.
2.8. Quivers 21
Definition 2.8.8. A subrepresentation of a representation (V,-, sch)
of a quiver Q is a representation (W,-, where W, Q V, for all 2' E I
and where :z:;,(Whr) Q Whn and 513;, = :z:;,|wh, : Whr —> Whu for all
h E E.
Definition 2.8.9. The direct sum of two representations (V,-,:z:;,)
and (W,-, yh) is the representation (Vi GB W,-, 513;, GB yh).

As with representations of algebras, a nonzero representation

of a quiver Q is said to be irreducible if its only subrepresentations
are (0) and itself, and it is said to be indecomposable if it is not
isomorphic to a direct sum of two nonzero representations.
Definition 2.8.10. Let (V,-,:z:;,) and (W,-,y;,) be representations of
the quiver Q. A homomorphism (,0 : —> of quiver
representations is a collection of maps «,0, : V, —> W, such that
yh o (phz = (phu o sch for all h E E.

Problem 2.8.11. Let A be a Z+-graded algebra, i.e., A = @700 A[n],

and A[n] - A[m] C A[n + m]. If A[n] is finite dimensional, it is— useful
to consider the Hilbert series h A(t) = ZdimA[n]t” (the generating
function of dimensions of Often this series converges to a ratio­
nal function, and the answer is written in the form of such a function.
For example, if A = k[:z:] and deg(:z:”) = n, then

hA(t)=1+t+t2+---+t"+---=E.
Find the Hilbert series of the following graded algebras:
(a) A = k[:z:1, . . . ,:z:m] (where the grading is by degree of polyno­
mials).
(b) A = k( 1:1, . . . ,:z:m) (the grading is by length of words).
(c) A is the exterior (= Grassmann) algebra /\k[:z:1, . . . ,:z:m] gen­
erated over some field k by 51:1, . . .,:z:m with the defining relations
:13,-:13, + :13,-51:, = O and == 0 for all z',j (the grading is by degree).
(d) A is the path algebra PQ of a quiver Q (the grading is defined
by deg(p.-) = 0, deg(ah) = 1)­
Hint: The closed answer is written in terms of the adjacency
matrix MQ of Q.
22 2. Basic notions of representation theory
2.9. Lie algebras
Let g be a vector space over a field k, and let [ , ] : g x g ——> g
be a skew-symmetric bilinear map. (That is, [a, a] = O, and hence
[a,b] = -[baa]-)
Definition 2.9.1. (g, [ , is a Lie algebra if [ , ] satisfies the Jacobi
identity
(2.9.1) [[a, b] ,c] + [[b, c] ,a] + [[c, a] ,b] = 0.
Example 2.9.2. Some examples of Lie algebras are:
(1) Any space g with [ , ] = O (abelian Lie algebra).
(2) Any associative algebra A with [a, b] = ab—ba , in particular,
the endomorphism algebra A = End(V), Where V is a vector
space. When such an A is regarded as a Lie algebra, it is
often denoted by g[(V) (general linear Lie algebra).
(3) Any subspace U of an associative algebra A such that [a, b] E
U for all a,b E U.
(4) The space Der(A) of derivations of an algebra A, i.e. linear
maps D : A —> A which satisfy the Leibniz rule:
D(ab) = D(a)b + aD(b).
(5) Any subspace a of a Lie algebra g which is closed under the
commutator map [ , ], i.e., such that [a, b] E a if a,b E a.
Such a subspace is called a Lie subalgebra of g.
Remark 2.9.3. Ado’s theorem says that any finite dimensional Lie
algebra is a Lie subalgebra of g[(V) for a suitable finite dimensional
vector space V.
Remark 2.9.4. Derivations are important because they are the “in­
finitesimal version” of automorphisms (i.e., isomorphisms onto itself
For example, assume that g(t) is a differentiable family of automor­
phisms of a finite dimensional algebra A over R or (C parametrized
by t E (—e,e) such that g(O) = Id. Then D 2: g’(0) : A —> A is a
derivation (check it!). Conversely, if D : A —> A is a derivation, then
e‘D is a 1-parameter family of automorphisms (give a proof !).
2.9. Lie algebras 23
This provides a motivation for the notion of a Lie algebra. Namely,
We see that Lie algebras arise as spaces of infinitesimal automorphisms
(= derivations) of associative algebras. In fact, they similarly arise as
spaces of derivations of any kind of linear algebraic structures, such
as Lie algebras, Hopf algebras, etc., and for this reason play a very
important role in algebra.
Here are a few more concrete examples of Lie algebras:
(1) R3 with [u, v] = u x v, the cross-product of u and v.

0 1 0 0 1 0
(2) 5[(n), the set of n x n matrices With trace 0.
For example, s[(2) has the basis

e=(o 0)’ f=(1o)w h=(o -1)

With relations
[hael:2ea [h’a.fl=‘2.fa
(3) The Heisenberg Lie algebra H of matrices ( ).
0 0 0 0 1 0 0 0
ac=OO1,y=O0O,c==OOO
1
It has the basis
COO CO!­ O-x- -1­

000000000
with relations [y,ac] =cand [y,c] = [ac,c] =0.
(4) The algebra aff(1) of matrices (5 5 .
Its basis consists ofX = ((1, 8) and Y = (8 (1,), With [X, Y] =
Y.
(5) 5o(n), the space of skew-symmetric n x n matrices, With
[a, b] = ab — ba.
Exercise 2.9.5. Show that example (1) is a special case of example
(5) (for n == 3).
Definition 2.9.6. Let g1,g2 be Lie algebras. A homomorphism
of Lie algebras go : g1 —> g2 is a linear map such that <,0([a, b]) =
[s0(a). s0(b)]­

Definition 2.9.7. A representation of a Lie algebra g is a vector

space V With a homomorphism of Lie algebras p : g —> End V.
24 2. Basic notions of representation theory
Example 2.9.8. Some examples of representations of Lie algebras
are:

(1) V = O.
(2) Any vector space V with p = O (the trivial representation).
(3) The adjoint representation V = g with p(a)(b) := [a, b].
That this is a representation follows from equation (2.9.1).
Thus, the meaning of the Jacobi identity is that it is equiv­
alent to the existence of the adjoint representation.

It turns out that a representation of a Lie algebra g is the same

thing as a representation of a certain associative algebra U Thus,
as With quivers, We can view the theory of representations of Lie alge­
bras as a part of the theory of representations of associative algebras.
Definition 2.9.9. Let g be a Lie algebra with basis ac, and [ , ]
defined by [ac,-, ac]-] = 2k cf]-wk. The universal enveloping algebra
H (g) is the associative algebra generated by the ac,-’s with the defining
relations ac,-ac_,- — :r_,-(E, = 2k cf]-wk.

Remark 2.9.10. This is not a very good definition since it depends

on the choice of a basis. Later We will give an equivalent definition
which will be basis-independent.
Exercise 2.9.11. Explain Why a representation of a Lie algebra is
the same thing as a representation of its universal enveloping algebra.
Example 2.9.12. The associative algebra I/{(5[(2)) is the algebra
generated by e, f, h, with relations
he-eh=2e, hf—fh=—2f, ef—fe=h.
Example 2.9.13. The algebra I/{(”H), Where H is the Heisenberg Lie
algebra, is the algebra generated by ac, y, c With the relations
yac—acy=c, yc—cy=0, acc—cac=0.
Note that the Weyl algebra is the quotient of I/{(”H) by the relation
c = 1.
Remark 2.9.14. Lie algebras were introduced by Sophus Lie (see
Section 2.10) as an infinitesimal version of Lie groups (in early texts
2.9. Lie algebras 25
they were called “infinitesimal groups” and were called Lie algebras
by Hermann Weyl in honor of Lie). A Lie group is a group G which is
also a manifold (i.e., a topological space which locally looks like IR")
such that the multiplication operation is differentiable. In this case,
one can define the algebra of smooth functions G°°(G) which carries
an action of G by right translations ((g o 2: f(acg)), and the
Lie algebra Lie(G) of G consists of derivations of this algebra which
are invariant under this action (with the Lie bracket being the usual
commutator of derivations). Clearly, such a derivation is determined
by its action at the unit element e E G, so Lie(G) can be identified
as a vector space with the tangent space TBG to G at e.
Sophus Lie showed that the attachment G r—> Lie(G) is a bijec­
tion between isomorphism classes of simply connected Lie groups (i.e.,
connected Lie groups on which every loop contracts to a point) and
finite dimensional Lie algebras over IR. This allows one to study (dif­
ferentiable) representations of Lie groups by studying representations
of their Lie algebras, which is easier since Lie algebras are “linear” ob­
jects while Lie groups are “nonlinear”. Namely, a finite dimensional
representation of G can be differentiated at e to yield a representa­
tion of Lie(G), and conversely, a finite dimensional representation of
Lie(G) can be exponentiated to give a representation of G. Moreover,
this correspondence extends to certain classes of infinite dimensional
representations.
The most important examples of Lie groups are linear algebraic
groups, which are subgroups of GLn(IR) defined by algebraic equa­
tions (such as, for example, the group of orthogonal matrices On(IR)).
Also, given a Lie subalgebra g C g[,,(IR) (which, by Ado’s theo­
rem, can be any finite dimensional real Lie algebra), we can define G
to be the subgroup of GLn(IR) generated by the elements ex , X E g.
One can show that this group has a natural structure of a connected
Lie group, whose Lie algebra is g (even though it is not always a
closed subgroup). While this group is not always simply connected,
its universal covering G is, and it is the Lie group corresponding to g
under Lie’s correspondence.
For more on Lie groups and their relation to Lie algebras, the
reader is referred to textbooks on this subject, e.g.
26 2. Basic notions of representation theory
2.10. Historical interlude: Sophus Lie’s trials
and transformations
To call Sophus Lie (1842-1899) an overachiever would be an under­
statement. Scoring first at the 1859 entrance examinations to the
University of Christiania (now Oslo) in Norway, he was determined
to finish first as well. When problems with his biology class derailed
this project, Lie received only the second-highest graduation score.
He became depressed, suffered from insomnia, and even contemplated
suicide. At that time, he had no desire to become a mathematician.
He began working as a mathematics tutor to support himself, read
more and more on the subject, and eventually began publishing re­
search papers. He was 26 when he finally decided to devote himself
to mathematics.
The Norwegian government realized that the best way to edu­
cate their promising scientists was for them to leave Norway, and
Lie received a fellowship to travel to Europe. Lie went straight to
Berlin, a leading European center of mathematical research, but the
mathematics practiced by local stars — Weierstrass and Kronecker —
did not impress him. There Lie met young Felix Klein, who eagerly
shared this sentiment. The two had a common interest in line geome­
try and became friends. Klein’s and Lie’s personalities complemented
each other very well. As the mathematician Hans Freudenthal put it,
“Lie and Klein had quite different characters as humans and math­
ematicians: the algebraist Klein was fascinated by the peculiarities
of charming problems; the analyst Lie, parting from special cases,
sought to understand a problem in its appropriate generalization”
[16, p. 323].
Lie liked to bounce ideas off his friend’s head, and Klein’s returns
were often quite powerful. In particular, Klein pointed out an anal­
ogy between Lie’s research on the tetrahedral complex and the Galois
theory of commutative permutation groups. Blissfully unaware of
the difficulties on his path, Lie enthusiastically embraced this sugges­
tion. Developing a continuous analog of the Galois theory of algebraic
equations became Lie’s idée fiace for the next several years.
2.10. Historical interlude: Sophus Lie 27
Lie and Klein traveled to Paris together, and there Lie produced
the famous contact transformation, which mapped straight lines into
spheres. An application of this expertise to the Earth sphere, however,
did not serve him well. After the outbreak of the Franco-Prussian war,
Lie could not find a better way to return to Norway than by first
hiking to Italy. With his peculiar hiking habits, such as taking off his
clothes in the rain and putting them into his backpack, he was not
able to flee very far. The French quickly apprehended him and found
papers filled with mysterious symbols. Lie’s efforts to explain the
meaning of his mathematical notation did not dispel the authorities’
suspicion that he was a German spy. A short stay in prison afforded
him some quiet time to complete his studies, and upon return to
Norway, Lie successfully defended his doctoral dissertation. Unable
to find a job in Norway, Lie resolved to go to Sweden, but Norwegian
patriots intervened, and the Norwegian National Assembly voted by
a large majority to establish a personal extraordinary professorship
for Lie at the University of Christiania. Although the salary offered
was less than extraordinary, he stayed.
Lie’s research on sphere mapping and his lively exchanges with
Klein led both of them to think of more general connections between
group theory and geometry. In 1872 Klein presented his famous Er­
langen Program, in which he suggested unifying specific geometries
under a general framework of projective geometry and using group
theory to organize all geometric knowledge. Lie and Klein clearly ar­
ticulated the notion of a transformation group, the continuous analog
of a permutation group, with promising applications to geometry and
differential equations, but they lacked a general theory of the sub­
ject. The Erlangen Program implied one aspect of this project — the
group classification problem — but Lie had no intention of attacking
this bastion at the time. As he later wrote to Klein, “[I]n your essay
the problem of determining all groups is not posited, probably on the
grounds that at the time Such a problem seemed to you absurd or
impossible, as it did to me” [22, pp. 41-42].
By the end of 1873, Lie’s pessimism gave way to a much brighter
outlook. After dipping into the theory of first order differential equa­
tions, developed by Jacobi and his followers, and making considerable
28 2. Basic notions of representation theory
advances with his idée fizce, Lie finally acquired the mathematical
weaponry needed to answer the challenge of the Erlangen Program
and to tackle the theory of continuous transformation groups.
Living on the outskirts of Europe, Lie felt quite marginalized in
the European mathematics community. No students and very few for­
eign colleagues were interested in his research. He wrote his papers
in German but published them almost exclusively in Norwegian jour­
nals, preferring publication speed over wide accessibility. A few years
later he learned, however, that one French mathematician had won
the Grand Prix from the Académie des Sciences for independently
obtained results that yielded some special cases of Lie’s work on dif­
ferential equations. Lie realized that his Norwegian publications were
not the greatest publicity vehicle, and that he needed to make his
work better known in Europe. “If only I could collect together and
edit all my results,” he wistfully wrote to Klein [22, p. 77]. Klein’s
practical mind quickly found a solution. Klein, who then taught at
Leipzig, arranged for the young mathematician Friedrich Engel, a re­
cent doctoral student of his colleague, to go to Christiania and to
render Lie a helping mathematical hand.
Lie and Engel met twice daily for a polite conversation about
transformation groups. As Engel recalled, Lie carried his theory al­
most entirely in his head and dictated to Engel an outline of each
chapter, “a sort of skeleton, to be clothed by me with flesh and blood”
[22, p. 77]. Lie read and revised Engel’s notes, eventually producing
the first draft of a book-length manuscript.
When Klein left Leipzig to take up a professorship at Gottingen,
he arranged for the vacated chair of geometry to be offered to Lie.
Lie somewhat reluctantly left his homeland and arrived at Leipzig
with the intention of building “a healthy mathematical school” there
[22, p. 226]. He continued his collaboration with Engel, which cul­
minated in the publication of their joint three-volume work, Theorie
der Transformationsgmppen.
Lie’s ideas began to spread around Europe, finding a particularly
fertile ground in Paris. Inspired by Lie, Henri Poincaré remarked that
all mathematics was a tale about groups, and Emile Picard wrote to
Lie, “Paris is becoming a center for groups; it is all fermenting in
2.10. Historical interlude: Sophus Lie 29
young minds, and one will have an excellent wine after the liquors
have settled a bit”. German mathematicians were less impressed.
Weierstrass believed that Lie’s theory lacked rigor and had to be re­
built from the foundations, and Frobenius labeled it a “theory of
methods” for solving differential equations in a roundabout way, in­
stead of the natural methods of Euler and Lagrange [22, pp. 186,
188-189}
Students flocked to Lie’s lectures on his own research, but this
only exacerbated his heavy teaching load at Leipzig — 8-10 lectures
per week — compared to the leisurely pace of his work in Christia­
nia. An outdoor man, who was used to weeks-long hikes in Norway,
Lie felt homesick, longing for the forests and mountains of his native
country. All this began taking its toll on Lie. Most importantly, he
felt underappreciated and became obsessed with the idea that others
plundered his work and betrayed his trust. His relations with col­
leagues gradually deteriorated, particularly with those closest to him.
He broke with Engel and eventually with Klein. Lie felt that his role
in the development of the Erlangen Program was undervalued, and he
publicly attacked Klein, claiming, “I am no pupil of Klein, nor is the
opposite the case, although this might be closer to the truth. I value
Klein’s talent highly and will never forget the sympathetic interest
with which he has always followed my scientific endeavors. But I do
not feel that he has a satisfactory understanding of the difference be­
tween induction and proof, or between a concept and its application”
[56, p. 371]. Whoever was right in this dispute, Lie’s public accu­
sations against widely respected and influential Klein reflected badly
on Lie’s reputation.
Eventually Lie suffered a nervous breakdown and was diagnosed
with “neurasthenia”, a popular mental disease dubbed the “American
Nervousness”, or “Americanitis”. Its cause was ascribed to the stress
of modern urban life and the exhaustion of an individual’s “nervous
energy”. Lie spent some months in the supposedly less stressful envi­
ronment of a psychiatric clinic and upon some reflection decided that
he was better off in his mathematics department. His mathemati­
cal abilities returned, but his psyche never fully recovered. Rumors
30 2. Basic notions of representation theory
spread of his mental illness, possibly fueled by his opponents, who
tried to invalidate his accusations.
In the meantime, trying to assert its cultural (and eventually
political) independence from Sweden, Norway took steps to bring back
its leading intellectuals. The Norwegian National Assembly voted to
establish a personal chair in transformation group theory for Lie,
matching his high Leipzig salary. Lie was anxious to return to his
homeland, but his wife and three children did not share his nostalgia.
He eventually returned to Norway in 1898 with only a few months to
live.
Lie “thought and wrote in grandiose terms, in a style that has
now gone out of fashion, and that would be censored by our scientific
journals”, wrote one commentator [26, p. iii]. Lie was always more
concerned with originality than with rigor. “Let us reason with con­
cepts!” he often exclaimed during his lectures and drew geometrical
pictures instead of providing analytical proofs [22, p. 244]. “With—
out Phantasy one would never become a Mathematician”, he wrote.
“[W]hat gave me a Place among the Mathematicians of our Day,
despite my Lack of Knowledge and Form, was the Audacity of my
Thinking” [56, p. 409]. Hardly lacking relevant knowledge, Lie indeed
had trouble putting his ideas into publishable form. Due to Engel’s
diligence, Lie’s research on transformation groups was summed up
in three grand volumes, but Lie never liked this ghost-written work
and preferred citing his own earlier papers [47 , p. 310]. He had even
less luck with the choice of assistant to write up results on contact
transformations and partial differential equations. Felix Hausdorff ’s
interests led him elsewhere, and Lie’s thoughts on these subjects were
never completely spelled out [16, p. 324]. Thus we may never discover
the “true Lie”.

2.11. Tensor products

In this subsection we recall the notion of tensor product of vector
spaces, which will be extensively used below.
Definition 2.11.1. The tensor product V 8) W of vector spaces V
and W over a field k is the quotient of the space V * W whose basis
2.11. Tensor products 31
is given by formal symbols '0 ® 111, v E V, w 6 W, by the subspace
spanned by the elements
(‘U1 +’U2)®’LU—’U1 ®’LU—’U2®’LU,
’U®(’LU1 +’LU2)—’U®’LU1 —’U®'LU2,
av®w—a('U®w),
v®aw—a(v®w),
Where '0 E V,wE W,aE k.

Exercise 2.11.2. Show that V (8) W can be equivalently defined as

the quotient of the free abelian group V o W generated by '0 (8) w,
v E V, w E W by the subgroup generated by
(’U1+’U2)®1U-’U1®1U—’U2®1Ua
v®(w1+w2)—v®w1-v®w2,
av®w—v®aw,
WherevEV,w€l/V,aEk.
The elements '0 ® 111 E V (X) W, for ‘U E V, w E W are called pure
tensors. Note that in general, there are elements of V 8) W Which are
not pure tensors.

This allows one to define the tensor product of any number of

vector spaces, V1 ®- - - ® Vn. Note that this tensor product is associa­
tive, in the sense that (V1 (8) l6) 8) V3 can be naturally identified with
14 69 (K6 <2» V3).
In particular, people often consider tensor products of the form
V®" = V (3) - - - 8) V (n times) for a given vector space V, and, more
generally, E := V®" ® (V*)®"”. This space is called the space of
tensors of type (m, n) on V. For instance, tensors of type (0,1)
are vectors, tensors of type (1,0) — linear functionals (covectors),
tensors of type (1,1) — linear operators, of type (2,0) — bilinear
forms, tensors of type (2, 1) — algebra structures, etc.
If V is finite dimensional with basis e,-, 72 = 1, . . . ,N, and ei is the
dual basis of V*, then a basis of E is the set of vectors
ez.1®...®ez.n®e.71®...®e.7m,
32 2. Basic notions of representation theory
and a typical element of E is
N
E Tjllmjgleil ®...®ez.n ®e.71 ®...®e.7m,
i1s'''sin9j1s°''9jTn=1 ­
where T is a multidimensional table of numbers.
Physicists define a tensor as a collection of such multidimensional
tables TB attached to every basis B in V, which change according to a
certain rule when the basis B is changed (derive this rule!). Here it is
important to distinguish upper and lower indices, since lower indices
of T correspond to V and upper ones to V*. The physicists don’t
write the sum sign, but remember that one should sum over indices
that repeat twice — once as an upper index and once as lower. This
convention is called the Einstein summation, and it also stipulates
that if an index appears once, then there is no summation over it,
while no index is supposed to appear more than once as an upper
index or more than once as a lower index.
One can also define the tensor product of linear maps. Namely,
if A : V —> V’ and B : W —> W’ are linear maps, then one can define
the linear map A (X) B : V (8) W —> V’ (8) W’ given by the formula
(A (X) B)(v (8) w) = AU (8) Bw (check that this is well definedl). The
most important properties of tensor products are summarized in the
following problem.

Problem 2.11.3. (a) Let U be any Is:-vector space. Construct a

natural bijection between bilinear maps V x W —> U and linear maps
V (8) W —> U (“natural” means that the bijection is defined without
choosing bases).
(b) Show that if is a basis of V and {w_,-} is a basis of W,
then {vi (8) w_,-} is a basis of V (8) W.
(c) Construct a natural isomorphism V* (8) W —> Hom(V, W) in
the case when V is finite dimensional.
(d) Let V be a vector space over a field 19. Let S"V be the quotient
of V® " (n-fold tensor product of V) by the subspace spanned by the
tensors T — s(T) where T E V®" and s is a transposition. Also let
/\”V be the quotient of V® " by the subspace spanned by the tensors
T such that s(T) = T for some transposition 3. These spaces are
2.11. Tensor products 33
called the nth symmetric power, respectively exterior power of
V. If is a basis of V, can you construct a basis of S"V, /\"V?
If dimV = m, what are their dimensions?
(e) If is has characteristic zero, find a natural identification of S "V
with the space of T E V® " such that T = sT for all transpositions s,
and find a natural identification of /\”V with the space of T E V®"
such that T = —sT for all transpositions s.
(f) Let A : V —> W be a linear operator. Then we have an op­
erator A®" : V®" —> W®” and its symmetric and exterior powers
S”A : S"V —> S"W, /\"A : /\”V —> /\"W which are defined in an
obvious way. Suppose that V = W and that dimV = N, and that
the eigenvalues of A are A1, . . . , AN. Find Tr(S”A) and Tr(/\"A).
(g) Show that /\N A = det(A)Id, and use this equality to give a
one-line proof of the fact that det(AB) = det(A) det(B).
Remark 2.11.4. Note that a similar definition to the above can be
used to define the tensor product V (8),; W, where A is any ring, V
is a right A-module, and W is a left A-module. Namely, V (8),; W
is the abelian group which is the quotient of the group V 0 W freely
generated by formal symbols '0 (8) w, v E V, w E W, modulo the
relations
(v1+v2)®w—v1®w—v2®w,
v®(w1+w2)—v®w1—v®w2,
va®w—v®aw, aEA.
Exercise 2.11.5. Let K be a field, and let L be an extension of K.
If A is an algebra over K, show that A ®K L is naturally an algebra
over L. Show that if V is an A-module, then V ®K L has a natural
structure of a module over the algebra A ®K L.
Problem 2.11.6. Throughout this problem, we let k be an arbi­
trary field (not necessarily of characteristic zero and not necessarily
algebraically closed).
If A and B are two k-algebras, then an (A,B)-bimodule will
mean a Is:-vector space V with both a left A-module structure and a
right B-module structure which satisfy (av) b = a (vb) for any '0 E V,
a E A, and b E B. Note that both the notions of “left A-module”
34 2. Basic notions of representation theory
and “right A-module” are particular cases of the notion of bimodules;
namely, a left A-module is the same as an (A, kt)-bimodule, and a right
A-module is the same as a (k, A)-bimodule.
Let B be a Is:-algebra, W a left B-module, and V a right B­
module. We denote by V (83 W the Is:-vector space
(V®k W)/(vb®w—v®bw | v E V, w E W, b E B). Wedenotethe
projection of a pure tensor '0 (8) 10 (With U E V and w E W) onto the
space V ®B W by v <83 10. (Note that this tensor product V 8)}; W
is the one defined in Remark 2.11.4.)
If, additionally, A is another Is:-algebra and if the right B-module
structure on V is part of an (A, B)-bimodule structure, then V®B W
becomes a left A-module by a (v <83 10) = av <83 10 for any a E A,
v E V, and w E W.
Similarly, if C is another Is:-algebra, and if the left B-module
structure on W is part of a (B, C)-bimodule structure, then V 8)}; W
becomes a right C-module by ('0 <83 10) c = v ®B wc for any c 6 C,
v E V, and w E W.
If V is an (A,B)-bimodule and W is a (B, C)-bimodule, then
these two structures on V (83 W can be combined into one (A, C)­
bimodule structure on V (83 W.
(a) Let A, B, C, D be four algebras. Let V be an (A, B)­
bimodule, W a (B,C)-bimodule, and X a (C,D)-bimodule. Prove
that (V ®B W) ®C X E’ V ®B (W ®C X) as (A, D)-bimodules. The
isomorphism (from left to right) is given by the formula

(v®3w)®Ca:»—>v®B(w®Ca:)

forallv€V,wEW,anda:EX.
(b) If A, B, C are three algebras and if V is an (A, B)-bimodule
and W an (A, C)-bimodule, then the vector space HomA (V, W) (the
space of all left A-linear homomorphisms from V to W) canonically
becomes a (B, C)-bimodule by setting (bf) (12) = f (vb) for all b E B,
f E HomA (V, W), and v E V and setting (fc) (v) = f(v)c for all
CE C, f E HomA(V,W) and ‘U E V.
2.12. The tensor algebra 35
Let A, B, C, D be four algebras. Let V be a (B, A)-bimodule,
W a (C, B)-bimodule, and X a (C, D)-bimodule. Prove that
HomB (V, Home (W,X)) E‘ Homc (W (83 V,X)
as (A, D)-bimodules. The isomorphism (from left to right) is given
by
f»—> (w®Bv»—> f(v)w)
for all '0 E V, w E W and f E HomB (V, Home (W,X)).
Exercise 2.11.7. Show that if M and N are modules over a commu­
tative ring A, then M (8),; N has a natural structure of an A-module.

2.12. The tensor algebra

The notion of tensor product allows us to give more conceptual (i.e.,
coordinate-free) definitions of the free algebra, polynomial algebra,
exterior algebra, and universal enveloping algebra of a Lie algebra.
Namely, given a vector space V, define its tensor algebra TV
over a field Is: to be TV = @n>0 V® ", With multiplication defined by
a - b := a8) b, a. E V®", b E V®'". Observe that a choice of a basis
a:1, . . . ,a:N in V defines an isomorphism of TV with the free algebra
k(a:1,...,a:n).
Also, one can make the following definition.
Definition 2.12.1. The symmetric algebra SV of V is the
quotient of TV by the ideal generated by '0 (8) w — w (8) v, v, w E V.
(ii) The exterior algebra /\V of V is the quotient of TV by the
ideal generated by '0 ® '0, ‘U E V.
(iii) If V is a Lie algebra, the universal enveloping alge­
bra I/l(V) of V is the quotient of TV by the ideal generated by
v®w—w®v—[v,w], v,wEV.
It is easy to see that a choice of a basis a:1, . . . ,a:N in V identifies
S V With the polynomial algebra k:[a:1, . . . ,a:N], /\V with the exterior
algebra /\k(a:1, . . . ,a:N), and the universal enveloping algebra U (V)
with one defined previously.
36 2. Basic notions of representation theory
Moreover, it is easy to see that we have decompositions

n20 n20
SV = Qgsnv, /\V = ®/\"V.

2.13. Hilbert’s third problem

Problem 2.13.1. It is known that if A and B are two polygons of the
same area, then A can be cut by finitely many straight cuts into pieces
from which one can make B (check it —- it is fun!). David Hilbert
asked in 1900 whether it is true for polyhedra in three dimensions. In
particular, is it true for a cube and a regular tetrahedron of the same
volume?
The answer is “no”, as was found by Dehn in 1901. The proof is
very beautiful. Namely, to any polyhedron A, let us attach its “Dehn
invariant” D(A) in V = IR (8) (IR/Q) (the tensor product of Q-vector
spaces). Namely,
DA=l— 3(0)
( ) (a) <2» 7, ,
where a runs over edges of A and l(a), B(a) are the length of a and
the angle at a.
(a) Show that if you cut A into B and C by a straight cut, then
D(A) = D(B) + D(C).
(b) Show that a = arccos(1 / 3) /7r is not a rational number.
Hint: Assume that a = 2m/n, for integers m, n. Deduce that
roots of the equation a: + a:’1 = 2/ 3 are roots of unity of degree n.
Then show that wk +a:”’° has denominator 3" and get a contradiction.
(c) Using (a) and (b), show that the answer to Hilbert’s question
is negative. (Compute the Dehn invariant of the regular tetrahedron
and the cube.)

2.14. Tensor products and duals of

representations of Lie algebras
Definition 2.14.1. The tensor product of two representations
V, W of a Lie algebra g is the space V (X) W with
Pv®W(5’3) = PVC”) ‘3’ Id + Id ®PW(93)­
2.15. Representations of 5[(2) 37
Definition 2.14.2. The dual representation V* to a representa­
tion V of a Lie algebra g is the dual space V* to V with pv-(a3) =
—Pv(93)*­

It is easy to check that these are indeed representations.

Problem 2.14.3. Let V, W, U be finite dimensional representations
of a Lie algebra g. Show that the space Homg (V® W, U) is isomorphic
to Homg(V, U (8) W*). (Here Homg 2: Homu(g).)

2.15. Representations of 5[(2)

This subsection is devoted to the representation theory of 5[(2), which
is of central importance in many areas of mathematics. It is useful to
study this topic by solving the following sequence of exercises, Which
every mathematician should do, in one form or another.

Problem 2.15.1. According to the above, a representation of 5[(2)

is just a vector space V with a triple of operators E, F, H such that
HE—EH = 2E, HF—FH = —2F, EF—FE = H (the correspond­
ing map p is given by 9(6) = E.p(f) = F. p(h) = H)­
Let V be a finite dimensional representation of s[(2) (the ground
field in this problem is (C).
(a) Take eigenvalues of H and pick one with the biggest real part.
Call it x\. Let V()\) be the generalized eigenspace corresponding to x\.

(b) Let W be any representation of 5[(2) and let u) E W be a

nonzero vector such that Ew = 0. For any is > 0 find a polynomial
Pk(a:) of degree k such that E’°F’°w = Pk(H)w. (First compute
EF’°w; then use induction in k.)
(c) Let 1) E V()\) be a generalized eigenvector of H with eigenvalue
A. Show that there exists N > 0 such that F N v = 0.
(d) Show that H is diagonalizable on V(/\). (Take N to be such
that FN = 0 on V()\), and compute ENFN12, v E V(/\), by Use
the fact that Pk(a:) does not have multiple roots.)
(e) Let N” be the smallest N satisfying Show that /\ = N,,—1.
38 2. Basic notions of representation theory
(f) Show that for each N > 0, there exists a unique up to isomor­
phism irreducible representation of s[(2) of dimension N. Compute
the matrices E, F, H in this representation using a convenient basis.
(For V finite dimensional irreducible take /\ as in (a) and v E V(/\)
an eigenvector of H. Show that 1), F12, . . . , F"v is a basis of V, and
compute the matrices of the operators E, F, H in this basis.)
Denote the (/\+1)-dimensional irreducible representation from (f)
by V)‘. Below you will show that any finite dimensional representation
is a direct sum of V)‘.
(g) Show that the operator C = EF + FE + H 2 / 2 (the so-called
Casimir operator) commutes with E, F, H and equals 5%” Id on
V)‘.

Now it is easy to prove the direct sum decomposition. Namely,

assume the contrary, and let V be a reducible representation of the
smallest dimension, which is not a direct sum of smaller representa­
tions.
(h) Show that C has only one eigenvalue on V, namely 3‘(—)“,,'l—m
for some nonnegative integer /\ (use the fact that the generalized
eigenspace decomposition of C must be a decomposition of represen­
tations).
(i) Show that V has a subrepresentation W = V)‘ such that
V/ W = nV;\ for some n (use (h) and the fact that V is the smallest
reducible representation which cannot be decomposed).
(j) Deduce from that the eigenspace V(/\) of H is (n + 1)­
dimensional. If 121, . . . , vn_,.1 is its basis, show that Fjv,-, 1 3 73 3 n+1,
0 g j 3 x\, are linearly independent and therefore form a basis of V
(establish that if Fa: = 0 and Ha: = pa: for a: 75 0, then Ca: = 5%2la:
and hence /1 = —x\).
(k) Define W,- = span(v,-, F12,-, . . . , FAQ),-). Show that W, are sub­
representations of V and derive a contradiction to the fact that V
cannot be decomposed.
(l) (J acobson-Morozov lemma) Let V be a finite dimensional com­
plex vector space and A : V —> V a nilpotent operator. Show that
there exists a unique, up to an isomorphism, representation of 5[(2)
2.16. Problems on Lie algebras 39
on V such that E = A. (Use the classification of the representations
and the Jordan normal form theorem.)
(m) (Clebsch-Gordan decomposition) Find the decomposition of
the representation V), (8) V,‘ of s[(2) into irreducibles components.
Hint: For a finite dimensional representation V of 5[(2) it is use­
ful to introduce the character xx/(a:) = Tr(e“’H), a: E (C. Show that
Xveaw(93) = Xx/(93) + XW(93) and Xv®W(93) = Xv(5'3)XW(93)- Then
compute the character of V), and of V), (8) V,‘ and derive the decompo­
sition. This decomposition is of fundamental importance in quantum
mechanics.
(11) Let V = (CM (8) CN and A = Jo,M ® IdN +IdM ®J0,N, Where
Jam is the Jordan block of size n with eigenvalue zero (i.e., J0,ne,- =
e,--1, i = 2,. . . ,n, and Jomel = 0). Find the Jordan normal form of
A using (1) and

2.16. Problems on Lie algebras

Problem 2.16.1 (Lie’s theorem). The commutant K (g) of a Lie
algebra g is the linear span of elements [a:, y], a:, y E g. This is an ideal
in g (i.e., it is a subrepresentation of the adjoint representation). A
finite dimensional Lie algebra g over a field k is said to be solvable if
there exists n such that K "(g) = 0. Prove the Lie theorem: if k = (C
and V is a finite dimensional irreducible representation of a solvable
Lie algebra g, then V is 1-dimensional.
Hint: Prove the result by induction in dimension. By the in­
duction assumption, K (g) has a common eigenvector v in V; that is,
there is a linear function X : K (g) —> (C such that av = x(a,)v for any
a E K Show that g preserves common eigenspaces of K (For
this you will need to show that x([a:, a]) = 0 for as E g and a E K(g).
To prove this, consider the smallest subspace U containing 12 and
invariant under a:. This subspace is invariant under K (g) and any
a E K (g) acts with trace dim(U)x(a,) in this subspace. In particular
0 = TY([$» 0]) = dim(U)x([9=» 0])-)
40 2. Basic notions of representation theory
Problem 2.16.2. Classify irreducible finite dimensional representa­
tions of the two-dimensional Lie algebra With basis X, Y and commu­
tation relation [X , Y] = Y. Consider the cases of zero and positive
characteristic. Is the Lie theorem true in positive characteristic?
Problem 2.16.3. (Hard!) For any element as of a Lie algebra g
let ad(a:) denote the operator g —> g,y +—> [a:,y]. Consider the Lie
algebra gn generated by two elements as, y with the defining relations
3»d($)2(y) = 3»d(y)"+1(9=) = 0­
(a) Show that the Lie algebras g1, g2, g3 are finite dimensional
and find their dimensions.
(b) (Harder!) Show that the Lie algebra g4 has infinite dimension.
Construct explicitly a basis of this algebra.
Problem 2.16.4. Classify irreducible representations of the Lie al­
gebra 5[(2) over an algebraically closed field k of characteristic 1? > 2.
Problem 2.16.5. Let k be an algebraically closed field of character­
istic zero, and let q E k’‘ , q yé :|:1. The quantum enveloping algebra
I/lq(s[(2)) is the algebra generated by e, f, K, K ‘1 with relations
K — K ‘1
KeK‘1 = (126, KfK'1 = q'2f, [6, f] =
q - q‘1
(if you formally set K = qh, you’ll see that this algebra, in an appro­
priate sense, “degenerates” to I/((5[(2)) as q —> 1). Classify irreducible
representations of I/lq(s[(2)). Consider separately the cases of q being
a root of unity and q not being a root of unity.
Chapter 3

General results of
representation theory

3.1. Subrepresentations in semisimple

representations
Let A be an algebra.

Definition 3.1.1. A semisimple (or completely reducible) rep­

resentation of A is a direct sum of irreducible representations.

Example 3.1.2. Let V be an irreducible representation of A of di­

mension n. Then Y = End(V), with action of A by left multipli­
cation, is a semisimple representation of A, isomorphic to nV (the
direct sum of n copies of V). Indeed, any basis 121, . . . , 12,, of V gives
rise to an isomorphism of representations End(V) —> nV, given by
as —> (53121, . . .,a:'v,,).

Remark 3.1.3. Note that by Schur’s lemma, any semisimple repre­

sentation V of A is canonically identified with EBX HomA(X, V) ®X ,
where X runs over all irreducible representations of A. Indeed, We
have a natural map f : EBX Hom(X, V) (X) X —> V, given by g (X) as —>
g(a:), a: E X, g E Hom(X, V), and it is easy to verify that this map is
an isomorphism.
42 3. General results of representation theory
We’ll see now how Schur’s lemma allows us to classify subrepre­
sentations in finite dimensional semisimple representations.
Proposition 3.1.4. Let Vi, 1 3 i 3 m, be irreducible finite dimen­
sional pairwise nonisomorphic representations of A, and let W be
a subrepresentation of V = niVi. Then W is isomorphic to
€921 riVi, ri 3 ni, and the inclusion gt 2 W —> V is a direct sum
of inclusions gbi : riVi —> niVi given by multiplication of a row vector
of elements of Vi ( of length ri) by a certain ri >< ni matrix Xi with
linearly independent rows: ¢>('v1,. .. ,'vi.,) = (121, . . . , 'ui1.)Xi.

Proof. The proof is by induction in n := ni. The base of

induction (n = 1) is clear. To perform the induction step, let us
assume that W is nonzero, and fix an irreducible subrepresentation
P C W. Such P exists (Problem 2.3.15).1 Now, by Schur’s lemma,
P is isomorphic to Vi for some i, and the inclusion gb| p : P —> V
factors through niVi and upon identification of P with Vi is given by
the formula 12 v—> (vql, . . . ,'uq,i1.), where qi E kt are not all zero.
Now note that the group Gi = GL,i,(kt) of invertible ni x ni
matrices over k acts on niVi by (121, . . . ,'v,i,) —> (121, . . . ,'u,i1.)gi (and
by the identity on n_iV_i, j 7E and therefore acts on the set of
subrepresentations of V, preserving the property we need to estab­
lish: namely, under the action of gi, the matrix Xi goes to Xigi,
while the matrices X]-, j 75 i, don’t change. Take gi E Gi such
that (q1,. ..,q,i1.)gi = (1, O, . . . ,0). Then Wgi contains the first sum­
mand Vi of niVi (namely, it is Pgi); hence Wgi = Vi 69 W’, where
W’ C n1V1 69- - -GB — 1)Vi 69- - - €Bn,,iVi,i is the kernel of the projec­
tion of Wgi to the first summand Vi along the other summands. Thus
the required statement follows from the induction assumption. El
Remark 3.1.5. In Proposition 3.1.4, it is not important that kt is
algebraically closed, nor does it matter that V is finite dimensional.
If these assumptions are dropped, the only change needed is that the
entries of the matrix Xi are no longer in kt but in Di = End A(Vi),
1Another proof of the existence of P, which does not use the finite dimensionality
of V, is by induction in n. Namely, if W itself is not irreducible, let K be the kernel
of the projection of W to the first summand V1. Then K is a subrepresentation of
(n1 — 1)V1 G3 - - - 63 nm Vm, which is nonzero since W is not irreducible, so K contains
an irreducible subrepresentation by the induction assumption.
3.2. The density theorem 43
which is, as We know, a division algebra. The proof of this generalized
version of Proposition 3.1.4 is the same as before (check it!).

3.2. The density theorem

Let A be an algebra over an algebraically closed field kt.
Corollary 3.2.1. Let V be an irreducible finite dimensional repre­
sentation of A, and let 121, . . . gun E V be any linearly independent
vectors. Then for any u21,. . . ,u2n E V there exists an element a E A
such that av, = w,-.

Proof. Assume the contrary. Then the image of the map A —> nV
given by a —> (a'v1,...,a'vn) is a proper subrepresentation, so by
Proposition 3.1.4 it corresponds to an r X n matrix X, r < n. Thus,
taking a = 1, We see that there exist vectors u1, . . .,u,. E V such
that (u1,. . .,u,.)X = (121, . . . gun). Let (q1,...,qn) be a nonzero vec­
tor such that X(q1, . . . ,qn)T = 0 (it exists because r < Then
Sq,-12,-= (u1,...,ur)X(q1,...,qn)T = O, i.e. Z:q,-12,-= O — a contra­
diction to the linear independence of 12,-. D
Theorem 3.2.2 (The density theorem). Let V be an irreducible
finite dimensional representation of A. Then the map p : A —> EndV
is surjectiue.
Let V = V1 EB 69 V,., where V,- are irreducible pairwise
nonisomorphic finite dimensional representations of A. Then the map
@:=1 p,- : A —> EB:=1 End(V,-) is surjectiue.

Proof. Let B be the image of A in End(V). We Want to show

that B = End(V). Let c E End(V), let 121, . . . gun be a basis of V,
and let w,- = cu,-. By Corollary 3.2.1, there exists a E A such that
av, = 111,-. Then a maps to c, so c E B, and We are done.
(ii) Let B, be the image of A in End(V,-), and let B be the im­
age of A in €B:=1 End(V,-). Recall that as a representation of A,
@:=1 End(V,-) is semisimple: it is isomorphic to 69;, d,-V,-, Where
(1,; = dim V,-. Then by Proposition 3.1.4, B = EB, B,-. On the other
hand, implies that B,- = End(V,-). Thus (ii) follows. Cl
44 3. General results of representation theory
3.3. Representations of direct sums of matrix
algebras
In this section We consider representations of algebras A = ED 1. Matdz. (kt)
for any field kt.

Theorem 3.3.1. Let A = 69;, lVl3.tdz.(kt). Then the irreducible rep­

resentations of A are V1 = k7d1,. ..,V,. = ktd"‘, and any finite dimen­
sional representation of A is a direct sum of copies of V1, . . . , V...

In order to prove Theorem 3.3.1, We shall need the notion of a

dual representation.
Definition 3.3.2 (Dual representation). Let V be a representation
of any algebra A. Then the dual representation V* is the represen­
tation of the opposite algebra A°p (or, equivalently, right A-module)
with the action
(f - a)(v) == f(av)­

Proof of Theorem 3.3.1. First, the given representations are clearly

irreducible, since for any 21 75 0, w E V,-, there exists a E A such that
av = 112. Next, let X be an n-dimensional representation of A. Then,
X * is an n-dimensional representation of A°p. But (Matd,(kt))°p E“
Matd,(kt) with isomorphism go(X) = XT, as (BC')T = CTBT. Thus,

mi
A E’ A°p and X * may be viewed as an n-dimensional representation
of A. Define
¢>:A€B---€BA——->X*
n copies
by
¢(a1a-‘-van) = all/1 +“’+anyn
Where is a basis of X *. The map (I) is clearly surjective, as
kt C A. Thus, the dual map ¢>* : X ——-> A"* is injective. But
A"* E A" as representations of A (check it!). Hence, Im ¢>* E’ X is a
subrepresentation of A". Next, Matd,(kt) = dz-V,-, so A = 69;, dz-V,-,
A" = EBA, nd,-V,-, as a representation of A. Hence by Proposition
3.1.4, X = €B:=1 m,-V,-, as desired. CI
3.4. Filtrations 45
Problem 3.3.3. The goal of this problem is to give an alternative
proof of Theorem 3.3.1, not using any of the previous results of Chap­
ter 3.
Let A1, A2, ..., An be 71, algebras with units 11, 12, ..., 1,,,
respectively. Let A = A1 69 A269 - - - 69 An. Clearly, 1,1_,- = 6,-_,-1,-, and
theunit ofAis 1=11+12+---+1.1.
For every representation V of A, it is easy to see that 1,-V is a
representation of A,- for every 13 6 {1,2, . . . ,n}. Conversely, if V1,
V2, ..., V” are representations of A1, A2, ..., An, respectively, then
V1 GB V2 69 - - - 69 V” canonically becomes a representation of A (with
(a1,a2,...,an) E A acting on V1€BV2€B---€BVn as (v1,v2,...,vn) v—>
(a1v1,a2v2, . . .,anvn)).
(a) Show that a representation V of A is irreducible if and only
if 1,-V is an irreducible representation of A,- for exactly one 72 6
{1,2, . . . ,n}, While 1,-V = O for all the other 71. Thus, classify the
irreducible representations of A in terms of those of A1, A2, . . . , An.
(b) Let d E N. Show that the only irreducible representation of
Matd(k) is kd, and every finite dimensional representation of Matd(k)
is a direct sum of copies of kd.
Hint: For every ('i,j) E {1,2,...,d}2, let E,-_,- E Matd(k) be the
matrix with 1 in the with row of the jth column and 0’s everywhere
else. Let V be a finite dimensional representation of Matd(k). Show
that V = E11V G3 EQQV G9 ‘ ‘ ‘ G3 EddV, and that I E11V —) E11V,
v »—> E,-1v is an isomorphism for every 73 6 {1,2, . . .,d}. For every
11 E E11V, denote S'(v) = (E11v,E21v,...,Ed1v). Prove that S(v)
is a subrepresentation of V isomorphic to kd (as a representation of
Matd(k)), and that v E S(v). Conclude that V = S(v1) EB S(v2) EB
EB S'(vk), Where {v1,v2, . . . ,v;,} is a basis of E11V.
(c) Deduce Theorem 3.3.1.

3.4. Filtrations
Let A be an algebra. Let V be a representation of A.
Definition 3.4.1. A (finite) filtration of V is a sequence of subrep­
resentations 0 = V0 C V1 C C Vn = V.
46 3. General results of representation theory
Lemma 3.4.2. Any finite dimensional representation V of an algebra
A admits a finite filtration O = V0 C V1 C C Vn = V such that the
successive quotients V1-/V1-_1 are irreducible.

Proof. The proof is by induction in dim(V). The base is clear, and

only the induction step needs to be justified. Pick an irreducible
subrepresentation V1 C V, and consider the representation U = V/ V1.
Then by the induction assumption U has a filtration 0 = U0 C U1 C
- C Un_1 = U such that U,-/U1-_1 are irreducible. Define V1 for
i 2 2 to be the preimages of U1-_1 under the tautological projection
Then0=V0CV1CV2C---CVn=Visa.
filtration of V with the desired property. E]
3.5. Finite dimensional algebras
Definition 3.5.1. The radical of a finite dimensional algebra A is
the set of all elements of A which act by O in all irreducible represen­
tations of A. It is denoted Rad(A).

Proof. Easy. Cl
Proposition 3.5.2. Rad(A) is a two—sided ideal.

Proposition 3.5.3. Let A be a finite dimensional algebra.

Let I be a nilpotent two—sided ideal in A; i.e., I " = O for some
n. Then I C Rad(A).
Rad(A) is a nilpotent ideal. Thus, Rad(A) is the largest
nilpotent two—sided ideal in A.

Proof. Let V be an irreducible representation of A. Let v E V.

Then Iv C V is a subrepresentation. If Iv 75 0, then Iv = V so there
is at E I such that ccv = v. Then at" 75 O, a. contradiction. Thus
Iv = 0, so I acts by 0 in V and hence I C Rad(A).
(ii) Let 0 = A0 C A1 C C An = A be a filtration of the
regular representation of A by subrepresentations such that A,-+1/A1
are irreducible. It exists by Lemma 3.4.2. Let a: E Rad(A). Then at
Rad(A)" = 0, as desired. [3
acts on A,-+1/A1 by zero, so at maps A,-+1 to A,-. This implies that
3.5. Finite dimensional algebras 47
Theorem 3.5.4. A finite dimensional algebra A has only finitely
many irreducible representations V,- up to an isomorphism. These
representations are finite dimensional, and
A/Rad(A) 2 @EndV,-.

Proof. First, for any irreducible representation V of A and for any

nonzero v E V, Av Q V is a finite dimensional subrepresentation
of V. (It is finite dimensional as A is finite dimensional.) As V is
irreducible and Av 75 O, V = Av and V is finite dimensional.
Next, suppose We have nonisomorphic irreducible representations
V1, V2, . . . , V... By Theorem 3.2.2, the homomorphism
@p,-:A—>@EndV,­
is surjective. So r 3 21. dim EndV,- 3 dim A. Thus, A has only
finitely many nonisomorphic irreducible representations (not more
than dim A).
Now, let V1, V2, . . . , V,. be all nonisomorphic irreducible finite di­
mensional representations of A. By Theorem 3.2.2, the homomor­
phism
613;), ; A —> {[3 End V,­

is surjective. The kernel of this map, by definition, is exactly Rad(A).

D

Corollary 3.5.5. 21. (dim V,-)2 3 dim A, where the V, ’s are the irre­
dvcible representations of A.

Proof. As dim End V,- = (dim V,-)2, Theorem 3.5.4 implies that dim A

:3, (dim V,-)2 g dim A. D

— dim Rad(A) = 2,. dim End V,- = 2, (dim V,-)2. As dim Rad(A) 2 0,

Example 3.5.6. 1. Let A = / This algebra has a unique

irreducible representation, which is a 1-dimensional space is, in Which
SE acts by zero. So the radical Rad(A) is the ideal
2. Let A be the algebra of upper triangular n x n matrices.
It is easy to check that the irreducible representations of A are V,-,
i = 1,. ..,n, which are 1-dimensional, and any matrix at acts by
48 3. General results of representation theory
So the radical Rad(A) is the ideal of strictly upper triangular matrices
(as it is a nilpotent ideal and contains the radical). A similar result
holds for block-triangular matrices.
Definition 3.5.7. A finite dimensional algebra A is said to be semi­
simple if Rad(A) = 0.
Proposition 3.5.8. For a finite dimensional algebra A, the following
are equivalent:
(1) A is semisimple.
(2) (dimV,-)2 = dim A, where the V,-’s are the irreducible
representations of A.
(3) A § $1. Matd1.(k) for some d,-.
(4) Any finite dimensional representation of A is completely re­
ducible (that is, isomorphic to a direct sum of irreducible
representations
(5) A is a completely reducible representation of A.

Proof. As dimA — dim Rad(A) = (dim V,-)2, clearly dimA =

Z1. (dim V,-)2 if and only if Rad(A) = 0. Thus, (1) <=>
By Theorem 3.5.4, if Rad(A) = 0, then clearly A 2 $1. Matd1.(k:)
for d,- = dim V,-. Thus, (1) => Conversely, if A E’ Matd1.(k),
then by Theorem 3.3.1, Rad(A) = 0, so A is semisimple. Thus (3) =>
(1)­
Next, (3) => (4) by Theorem 3.3.1. Clearly (4) => To see
that (5) => (3), let A = $1. n,-V,-. Consider End A(A) (endomorphisms
of A as a representation of A). As the V,-’s are pairwise nonisomor­
phic, by Schur’s lemma, no copy of V, in A can be mapped to a

as desired. [:1
distinct Also, again by Schur’s lemma, EndA = k. Thus,
EndA(A) 2 $1.Matn1.(k). But EndA(A) E A°p by Problem 2.3.17,
so A°p 2’ $1. Matn1.(k). Thus, A § Matn1.(k))°p = $1. Matn1.(k:),

3.6. Characters of representations

Let A be an algebra and V a finite dimensional representation of
A with action p. Then the character of V is the linear function
3.6. Characters of representations 49
X1/:A ——> is given by

xv(a) = TY|v(p(a))­

If [A, A] is the span of commutators [x, y] := my — ya: over all x, y E A,

then [A, A] Q ker XV. Thus, We may view the character as a mapping
X1/:A/[A,A]—> k.

Exercise 3.6.1. Show that if W C V are finite dimensional repre­

sentations of A, then XV = Xw + XV/W.

Theorem 3.6.2. ( Characters of (distinct) irreducible finite dimen­

sional representations of A are linearly independent.
If A is a finite dimensional semisimple algebra, then these
characters form a basis of (A/[A, A])*.

Proof. If V1, . . . , V,. are nonisomorphic irreducible finite dimension­

al representations of A, then the map

pV1EB---EBpV1_ZA—)EIld V,­
is surjective by the density theorem, so X1/1, . . . , XV, are linearly inde­
pendent. (Indeed, ifz A,-XV,.(a) = O for all a E A, then 2 /\,-Tr(M,-) =
0 for all M, E EndkV,-. But each can range independently over
k, so it must be that A1 = = A7. = 0.)
(ii) First we prove that [Matd(k), Matd(k)] = 5Id(k:), the set of all
matrices with trace 0. It is clear that [Matd(k), Matd(k)] Q 5Id(k:). If
We denote by E,-j the matrix with 1 in the ith row of the jth column
and 0’s everywhere else, we have [Ez-J-,E3-m] = E,-m for i aé m and
]Ei,i+la Ei+1,i] = En — Ei+1,i+1- NOW U {Eu — Ei+1,i+1} f0I‘mS
a basis in 5Id(k), so indeed [Matd(k:), Matd(k)] = 5[d(k), as claimed.
By semisimplicity, We can write A = Matd, EB - - - EB Matd,_(k).
Then [A, A] = sldl - -EBsld,_(k:), and A/[A, A] E’ k’. By Theorem
3.3.1, there are exactly r irreducible representations of A (isomor­
phic to k:d1,...,k:d", respectively) and therefore r linearly indepen­
the characters form a basis. Cl
dent characters on the r-dimensional vector space A/[A, A]. Thus,
50 3. General results of representation theory
3.7. The Jordan-Holder theorem
We will now state and prove two important theorems about represen­
tations of finite dimensional algebras — the Jordan-Holder theorem
and the Krull-Schmidt theorem.

Theorem 3.7.1 (Jordan—Holder theorem). Let V be a finite dimen­

sional representation of A, and let 0 = V0 C V1 C C V” = V,
0 = V0’ C C V,f, = V be filtrations of V, such that the represen­
tations W,- := V,-/V,-_1 and W,-’ := V,-’/V,-’_1 are irreducible for all i.
Then n = m, and there exists a permutation or of 1, . . . , n such that
Wa(,-) is isomorphic to W,-’ .

Proof. First proof (for k: of characteristic zero). The character of

V obviously equals the sum of characters of W, and also the sum of
characters of W,-’ . But by Theorem 3.6.2, the characters of irreducible
representations are linearly independent, so the multiplicity of every
irreducible representation W of A among W,- and among W,-’ is the
same. This implies the theorem?
Second proof (general). The proof is by induction on dim V.
The base of induction is clear, so let us prove the induction step.
If W1 = W1’ (as subspaces), we are done, since by the induction
assumption the theorem holds for V/W1. So assume W1 aé In
this case W1 fl W’ = 0 (as W1, W1’ are irreducible), so we have an
embedding f : W1 69 W1’ —-> V. Let U = V/(W1 69 W1’), and let
0 = U0 C U1 C C Up = U be a filtration of U with simple
quotients Z, = U,-/U,-_1 (it exists by Lemma 3.4.2). Then we see the
following:
1) V/ W1 has a filtration with successive quotients W1’, Z1, . . . , Zp
and another filtration with successive quotients W2, . . . ., W”.
2) V/W1’ has a filtration with successive quotients W1, Z1, . . . , Zp
and another filtration with successive quotients W5, . . . .,

2This proof does not work in characteristic p because it only implies that the
multiplicities of W, and W,-' are the same modulo p, which is not sufficient. In fact,
the character of the representation pV, where V is any representation, is zero.
3.8. The Krull-Schmidt theorem 51
By the induction assumption, this means that the collection of
irreducible representations with multiplicities W1, W1’, Z1, . . . , Zp co­

W{, . . . , We are done. CI

incides on one hand with W1, . . . , W” and on the other hand with

The Jordan—H6lder theorem shows that the number n of terms in

a filtration of V with irreducible successive quotients does not depend
on the choice of a filtration and depends only on V. This number is
called the length of V. It is easy to see that n is also the maximal
length of a filtration of V in which all the inclusions are strict.
The sequence of the irreducible representations W1, . . . , Wn enu­
merated in the order they appear from some filtration of V as succes­
sive quotients is called a J ordan-Holder series of V.

3.8. The Krull-Schmidt theorem

Theorem 3.8.1 (Krull-Schmidt theorem). Any finite dimensional
representation of A can be uniquely (up to an isomorphism and the
order of summands) decomposed into a direct sum of indecomposable
representations.

Proof. It is clear that a decomposition of V into a direct sum of in­

decomposable representations exists, so we just need to prove unique­
ness. We will prove it by induction on dim V. Let V = V1€B- - -€BVm =
V1'€B---€BV,;. Let is : V3 ——> V, : V3’ —-> V, ps : V —-> V3,
p; : V —> V3’ be the natural maps associated with these decompo­
sitions. Let 03 = p1i;p;i1 : V1 —> V1. We have 22:1 03 = 1. Now we
need the following lemma.

Lemma 3.8.2. Let W be a finite dimensional indecomposable repre­

sentation of A. Then:
( Any homomorphism 0 : W —> W is either an isomorphism or
nilpotent.
If 03 : W —> W, s = 1,. . . ,n, are nilpotent homomorphisms,
then so is0 := 01 +---+0”.
52 3. General results of representation theory
Proof. Generalized eigenspaces of 0 are subrepresentations of W,
and W is their direct sum. Thus, 0 can have only one eigenvalue A.
If A is zero, 0 is nilpotent; otherwise it is an isomorphism.
(ii) The proof is by induction in n. The base is clear. To make
the induction step (n — 1 to n), assume that 0 is not nilpotent. Then
by (i), 0 is an isomorphism, so Z?=10"101 = 1. The morphisms
0‘10,- are not isomorphisms, so they are nilpotent. Thus 1 - 0‘10,, =

the induction assumption. C!

0‘101 + - - - + 0‘10,,_1 is an isomorphism, which is a contradiction to

By the lemma, we find that for some 3, 03 must be an isomor­

phism; we may assume that s = 1. In this case, V1’ = Im(p’1'i1)€B
Ker(p1'i’1), so since V1’ is indecomposable, we get that f := p’17I1 :
V1 —> V1’ and g := p1'i’1 : V1’ —> V1 are isomorphisms.
Let B = ®j>1 V1-, B’ = ®j>1 Vj’; then we have V = V1 EB B =
V1’ EB B’. Consider the map h : B —> B’ defined as a composition of
the natural maps B —> V —-> B’ attached to these decompositions.
We claim that h is an isomorphism. To show this, it suffices to show
that Ker h = 0 (as h is a map between spaces of the same dimension).
Assume that v E Kerh C B. Then 12 6 V1’. On the other hand, the
projection of v to V1 is zero, so gv = 0. Since g is an isomorphism,
we get 12 = 0, as desired.
Now by the induction assumption, m = n, and E‘ V411.) for
some permutation 0 of 2,. . . ,n. The theorem is proved.

Problem 3.8.3. The above proof of Lemma 3.8.2 uses the condition
that k is an algebraically closed field. Prove Lemma 3.8.2 (and hence
the Krull-Schmidt theorem) without this condition.

Problem 3.8.4. Let V, W be finite dimensional representations

of an algebra A over a (not necessarily algebraically closed) field K .
Let L be a field extension of K . Suppose that V ®K L is isomorphic
to W ®K L as a module over the L-algebra A ®K L. Show that V
and W are isomorphic as A-modules.
Hint: Reduce to the case of finitely generated, then finite ex­
tension, of some degree n. Then regard V ®K L and W ®K L as
3.9. Problems 53
A-modules, and show that they are isomorphic to V" and W", re­
spectively. Deduce that V” is isomorphic to W", and use the Krull­
Schmidt theorem (valid over any field by Problem 3.8.3) to deduce
that V is isomorphic to W.
(ii) (The Noether-Deuring theorem) In the setting of (i), suppose
that V ®K L is a direct summand in W ®K L (i.e., W ®K L E’
V ®K L 69 Y, where Y is a module over A ®K L). Show that V is a
direct summand in W.
Problem 3.8.5. Let A be the algebra of real—valued continuous func­
tions on IR which are periodic with period 1. Let M be the A-module
of continuous functions f on R which are antiperiodic with period 1,
i.e., f(a: + 1): —f(a:).
(i) Show that A and M are indecomposable A-modules.
(ii) Show that A is not isomorphic to M but A 69 A is isomorphic
to M ED M.
Remark 3.8.6. Thus, we see that, in general, the Krull-Schmidt
theorem fails for infinite dimensional modules. However, it still holds
for modules of finite length, i.e., modules M such that any filtration
of M has length bounded above by a certain constant l = l(M
3.9. Problems
Problem 3.9.1. Extensions of representations. Let A be an
algebra, and let V, W be a pair of representations of A. We would like
to classify representations U of A such that V is a subrepresentation of
U and U / V = W. Of course, there is an obvious example U = VEBW,
but are there any others?
Suppose we have a representation U as above. As a vector space,
it can be (nonuniquely) identified with V EB W, so that for any a E A
the corresponding operator pg (0.) has block triangular form
_P()f()
pU(a) _ ( V0a PwC(l0«)),
where f : A —> Homk(W, V) is a linear map.
(a) What is the necessary and sufficient condition on f (a) under
which pg (0.) is a representation? Maps f satisfying this condition are
54 3. General results of representation theory
called 1-cocycles (of A with coefficients in Homk( W, They form
a vector space denoted by Z1(W, V).
(b) Let X : W —> V be a linear map. The coboundary of X, dX,
is defined to be the function A —> Homk(W, V) given by dX (a) =
pv(a)X — X pw(a). Show that dX is a cocycle which vanishes if
and only if X is a homomorphism of representations. Thus cobound­
aries form a subspace B1(W, V) C Z1(W, V), which is isomorphic
to Homk(W, V)/HomA(W, V). The quotient Z1(W, V)/B1(W, V) is
denoted by Ext1(W, V).
(c) Show that if f, f’ E Z1(W, V) and f — f’ E B1(W, V), then
the corresponding extensions U, U’ are isomorphic representations of
A. Conversely, if (15 : U —> U’ is an isomorphism such that

¢<a> = (1; 1;) .

then f — f’ E B1(V, Thus, the space Ext1(W, V) “classifies”
extensions of W by V.
(d) Assume that W, V are finite dimensional irreducible represen­
tations of A. For any f E Ext1(W, V), let U f be the corresponding
extension. Show that U f is isomorphic to U f/ as representations if
and only if f and f’ are proportional. Thus isomorphism classes
(as representations) of nontrivial extensions of W by V (i.e., those
not isomorphic to W 69 V) are parametrized by the projective space
IP’Ext1(W, V). In particular, every extension is trivial if and only if
Ext1(W, V) = 0.
Problem 3.9.2. (a) Let A = C[a:1, . . . ,a:,,], and let Va, Vb be 1­
dimensional representations in which the elements :13, act by a.,- and b,-,
respectively (a,-, b,- E C). Find Ext1(Va, Vb) and classify 2-dimensional
representations of A.
(b) Let B be the algebra over C generated by 3:1, . . . ,a:,, with the
defining relations :13,-a:j = 0 for all 2', j. Show that for n > 1 the algebra
B has infinitely many nonisomorphic indecomposable representations.
Problem 3.9.3. Let Q be a quiver without oriented cycles, and let
PQ be the path algebra of Q. Find irreducible representations of PQ
and compute Extl between them. Classify 2-dimensional representa­
tions of PQ.
3.9. Problems 55
Problem 3.9.4. Let A be an algebra, and let V be a representation
of A. Let p : A ——> EndV. A formal deformation of V is a formal
series
,5=p0+tp1+---+tnpn+...,
where p,- : A —> End(V) are linear maps, p0 = p, and p(ab) = ,5(a),5(b).

If b(t) = 1+b1t+b2t2 +. . . , where b,- E End(V), and ,5 is a formal

deformation of p, then bpb-1 is also a deformation of p, which is said
to be isomorphic to ,5.
(a) Show that if Ext1(V, V) = 0, then any deformation of p is
trivial, i.e., isomorphic to p.
(b) Is the converse to (a) true? (Consider the algebra of dual
numbers A =
Problem 3.9.5. The Clifford algebra. Let V be a finite dimen­
sional complex vector space equipped with a symmetric bilinear form
( , The Clifford algebra Cl(V) is the quotient of the tensor alge­
bra TV by the ideal generated by the elements v (8)?) — (v, v)1, v E V.
More explicitly, if :13,-, 1 3 2' 3 N, is a basis of V and (.’II»;,.’I.‘j) = aij

_2_
then Cl(V) is generated by :13, with defining relations
.’L‘i.’I7_7' + .’I3j.’I.‘»g — 2a,-J-, IL‘, — a,-,.

Thus, if ( , ) = 0, C1(V) = /xv.

(i) Show that if ( , ) is nondegenerate, then C1(V) is semisimple
and has one irreducible representation of dimension 2" if dim V = 271
(so in this case Cl(V) is a matrix algebra) and two such representa­
tions if dim(V) = 2n + 1 (i.e., in this case Cl(V) is a direct sum of
two matrix algebras).
Hint: In the even case, pick a basis a1,...,an,b1,...,bn of V
in which (a,-,aj) = (b,-,bj) = 0, (a,-,bj) = 6,-J-/2, and construct a
representation of Cl(V) on S := /\(a1, . . . ,an) in which bi acts as
“differentiation” with respect to a,-. Show that S is irreducible. In the
odd case the situation is similar, except there should be an additional
basis vector c such that (c, a,-) = (c, b,-) = 0, (c, c) = 1 and the action
of c on S may be defined either by (—1)deg'ee or by (—1)deg'ee+1,
giving two representations S+, S_ (why are they nonisomorphic?).
56 3. General results of representation theory
Show that there is no other irreducible representations by finding a
spanning set of Cl(V) with Zdimv elements.
(ii) Show that Cl(V) is semisimple if and only if ( , ) is nonde­
generate. If ( , ) is degenerate, what is Cl(V)/ Rad(Cl(V))?

3.10. Representations of tensor products

Let A, B be algebras. Then A <8) B is also an algebra, with multipli­
cation (a1 <8) b1)(a2 <8) 02) = alag <8) blbg.

Exercise 3.10.1. Show that Matm(k) <8) Matn(k) E’ Matmn(k).

The following theorem describes irreducible finite dimensional

representations of A <8) B in terms of irreducible finite dimensional
representations of A and those of B.
Theorem 3.10.2. ( Let V be an irreducible finite dimensional rep­
resentation of A and let W be an irreducible finite dimensional rep­
resentation of B. Then V <8) W is an irreducible representation of
A <8) B.
( ii ) Any irreducible finite dimensional representation M of A®B
has the form for unique V and W.
Remark 3.10.3. Part (ii) of the theorem typically fails for infinite
dimensional representations; e.g. it fails when A is the Weyl algebra
in characteristic zero. Part may also fail. E.g. let A = B = V =
W = Then fails, as A <8) B is not a field.
Proof. By the density theorem, the maps A —> End V and B —>
End W are surjective. Therefore, the map A®B ——> End V®End W =
End(V <8) W) is surjective. Thus, V <8) W is irreducible.
(ii) First we show the existence of V and W. Let A’, B’ be the
images of A, B in End M. Then A’, B’ are finite dimensional algebras,
and M is a representation of A’ <8) B’, so we may assume without loss
of generality that A and B are finite dimensional.
In this case, We claim that
Rad(A <8) B) = Rad(A) <8) B + A <8) Rad(B).
3.10. Representations of tensor products 57
Indeed, denote the latter by J. Then J is a nilpotent ideal in A®B, as
Rad(A) and Rad(B) are nilpotent. On the other hand, (A <8) B)/J =
(A / Rad(A)) <8) (B / Rad(B)), which is a product of two semisimple al­
gebras, hence semisimple. This implies J D Rad(A ®B). Altogether,
by Proposition 3.5.3, we see that J = Rad(A <8) B), proving the claim.
Thus, we see that
(A <8) B)/ Rad(A <8) B) = A/ Rad(A) <8) B/ Rad(B).
Now, M is an irreducible representation of (A <8) B) / Rad(A <8) B),
so it is clearly of the form M = V <8) W, where V is an irreducible
representation of A / Rad(A) and W is an irreducible representation
of B / Rad(B). Also, V, W are uniquely determined by M (as all of
the algebras involved are direct sums of matrix algebras). C]
Chapter 4

Representations of finite
groups: Basic results

Recall that a representation of a group G’ over a field kt is a kt—vector

space V together with a group homomorphism p : G’ —> GL(V). As
we have explained above, a representation of a group G’ over kt is the
same thing as a representation of its group algebra kt[G].
In this section, we begin a systematic development of representa­
tion theory of finite groups.

4.1. Maschke’s theorem

Theorem 4.1.1 (Maschke). Let G be a finite group and let kt be a
field whose characteristic does not divide |G Then:
The algebra kt[G] is semisimple.
There is an isomorphism of algebras zb : kt[G] —> EndV,­
defined by g I—> G9,, g|V,., where V, are the irreducible representations
of G. In particular, this is an isomorphism of representations of G
(where G acts on both sides by left multiplication). Hence, the regu­
lar representation kt[G] decomposes into irreducibles as dim(V,-)V,­
and one has the “sum of squares formula”

IGI = Zdim<vt->2.

O-‘Ino
60 4. Representations of finite groups: Basic results
Proof. By Proposition 3.5.8, implies (ii), and to prove (i), it is
sufficient to show that if V is a finite dimensional representation of
G’ and W C V is any subrepresentation, then there exists a subrep­
resentation W’ C V such that V = W EB W’ as representations.
Choose any complement W of W in V. (Thus V = W 69 W
as vector spaces, but not necessarily as representations.) Let P be
the projection along W onto W, i.e., the operator on V defined by
P|w = Id and PIW = 0. Let
—1
P == E Zg€G'p(9)Pp(9“’),
where p(g) is the action of g on V, and let
W’ = ker P.
Now Plw = Id and P(V) Q W, so P2 = P, and so P is a projection
along W’. Thus, V = W GB W’ as vector spaces.
Moreover, for any h E G’ and any y E W’,

Pp(h)y = Ilag€G'
Z p(9)Pp(9"h)y

= lg, ZKEG’
p<he>Pp<e-1>y = p<h>?y = o.

so p(h)y E kerP = W’. Thus, W’ is invariant under the action of G’

and is therefore a subrepresentation of V. Thus, V = W 69 W’ is the
desired decomposition into subrepresentations. C]
The converse to Theorem 4.1.1(i) also holds.
Proposition 4.1.2. If k:[G'] is semisimple, then the characteristic of
k does not divide
Proof. Write k:[G'] = 6927:, End V,;, where the V,- are irreducible rep­
resentations and V1 = k: is the trivial 1-dimensional representation.
Then

k[G] = keaésnd V, = keBG”9d,-I/.,;,

i=2 i=2
4.2. Characters 61
where d,; = dim V,-. By Schur’s lemma,
HOInk[(;](k?, k[G]) = kA,
Homk[(;](kt[G], kt) = kte,
for nonzero homomorphisms of representations 6 : kt[G] —> kt and
A : kt ——> k[G] unique up to scaling. We can take 6 such that e(g) = 1
for all g 6 G, and we can take A such that A(1) = ZQEG g. Then

eoA(1) =.(Zg)
g€G' =g€G'
21 = .01.
This is a contradiction. CI
If |G| = 0, then A has no left inverse, as (as) oA(1) = 0 for any a E kt.

Example 4.1.3. If G’ = Z/pZ and kt has characteristic p, then every

irreducible representation of G’ over kt is trivial (so kt[Z/pZ] indeed is
not semisimple). Indeed, an irreducible representation of this group
is a 1-dimensional space on which the generator acts by a pth root of
unity. But every pth root of unity in kt equals 1, as :13" — 1 = (:1: — 1)”
over kt.

Problem 4.1.4. Let G’ be a group of order p". Show that every

irreducible representation of G over a field kt of characteristic p is
trivial.

4.2. Characters
If V is a finite dimensional representation of a finite group G, then its
character XV : G —> kt is defined by the formula Xx/(g) = Tr|V(p(g)).
Obviously, Xx/(g) is simply the restriction of the character X‘/(a) of
V as a representation of the algebra A = kt[G] to the basis G C A, so
it carries exactly the same information. The character is a central
function, or class function: X‘/(g) depends only on the conjugacy
class of g; i.e., XV(hgh‘1) = X‘/(g).
Denote by F (G, kt) the space of kt-valued functions on G and by
Fc(G, kt) C F (G, kt) the subspace of class functions.
Theorem 4.2.1. If the characteristic of kt does not divide |G I, char­
acters of irreducible representations of G form a basis in the space
Fc(G, k).
62 4. Representations of finite groups: Basic results
Proof. By the Maschke theorem, k[G] is semisimple, so by Theo­
rem 3.6.2, the characters are linearly independent and are a basis of
(A/ [A, A])*, where A = kt[G]. It suffices to note that, as vector spaces
over kt,
(A/[A.A])* E’ {r E Hom;.(k[G]. k) I ah - hg 6 kerr V9.71 6 G}
3 {f E F(G.k) I f(9h) = f(h9) V9.11 6 G}.
which is precisely Fc(G, kt). Cl
Corollary 4.2.2. The number of isomorphism classes of irreducible
representations of G’ equals the number of conjugacy classes of G’
|G’| 79 O in kt).
Exercise 4.2.3. Show that if |G| = O in kt, then the number of iso­
morphism classes of irreducible representations of G’ over kt is strictly
less than the number of conjugacy classes in G.
Hint: Let P = Zgeag E kt[G]. Then P2 = 0. So P has zero
trace in every finite dimensional representation of G over kt.
Corollary 4.2.4. Any representation of G is determined by its char­
acter if kt has characteristic 0; namely, XV = XW implies V E’ W.

4.3. Examples
The following are examples of representations of finite groups over (C.
(1) Finite abelian groups G = Zn, x - - - x Znk. Let GV be the set
of irreducible representations of G. Every element of G forms a con­
jugacy class, so |GV| —-— |G Recall that all irreducible representations
over (C (and algebraically closed fields in general) of commutative al­
gebras and groups are 1-dimensional. Thus, GV is an abelian group:
if p1, pg : G ——> (C’‘ are irreducible representations, then so are the
representations p1(g)p2(g) and p1(g)‘1. The group GV is called the
dual group or character group of G.
For given n 2 1, define p : Zn —> (C’‘ by p(m) = e2’”"”/". Then
Z): = {pk : kt = 0,...,n— 1}, so Z): ‘é Zn. In general,
(G1 xG2 x xG,,)V =G\1’ xG§’ x x(}x,
so GV E’ G for any finite abelian group G. This isomorphism is,
however, noncanonical: the particular decomposition of G as
4.3. Examples 63
Zn, x - - - X Znk is not unique as far as which elements of G’ correspond
to Zn,, etc., is concerned. On the other hand, G’ E (GV)V is a canon­
ical isomorphism, given by cp : G’ —> (GV)V, where cp(g)(X) = X(g).
(2) The symmetric group S3. In the symmetric group Sn, con­
jugacy classes are determined by cycle decomposition sizes: two per­
mutations are conjugate if and only if they have the same number
of cycles of each length. For S3, there are three conjugacy classes,
so there are three difierent irreducible representations over (C. If
their dimensions are d1,d2,d3, then df + dg + d3 = 6, so S3 must
have two 1-dimensional and one 2-dimensional representations. The
1-dimensional representations are the trivial representation (C+ given
by p(0) = 1 and the sign representation (C_ given by p(0) = (—1)‘’.
The 2-dimensional representation can be visualized as represent­
ing the symmetries of the equilateral triangle with vertices 1, 2, 3 at
the points (cos 120°, sin 120°), (cos 240°, sin 240°), (1,0) of the coor­
dinate plane, respectively. Thus, for example,
1 0 cos 120° — sin 120°
p((12)) Z (0 -1) ’ ’0((123)) : (sin120° cos 120° ) °
To show that this representation is irreducible, consider any subrep­
resentation V. The space V must be the span of a subset of the
eigenvectors of p(( 12)), which are the nonzero multiples of (1,0) and
(0,1). Also, V must be the span of a subset of the eigenvectors of
p((123)), which are different vectors. Thus, V must be either (C2 or
0.

(3) The quaternion group Q3 = {:l:1, :l:z', :l:j, :l:k:}, with defining
relations
z'=jk=-kj, j=kz'=—z'k, k=z'j=—j2I, —1=z'2=j2-—-k2.
The five conjugacy classes are {1}, {-1}, {ii}, {:l:j}, {:l:k}, so there
are five difierent irreducible representations, the sum of the squares
of whose dimensions is 8, so their dimensions must be 1, 1, 1, 1, and
2.

The center Z(Q3) is {:l:1}, and Q3/Z(Q3) ’—‘_-’ Z2 x Z2. The four
1-dimensional irreducible representations of Z2 X Z2 can be “pulled
back” to Q3. That is, if q 2 Q3 —> Q3/Z(Q3) is the quotient map and
64 4. Representations of finite groups: Basic results
/3 is any representation of Q3 / Z (Q3), then poq gives a representation
Of Q3.
The 2-dimensional representation is V = (C2, given by p(—1) =
—Id and

p(i) = (_01 3) .
(4.31) p<j>=
O — -1
P0“) = (_\/fl 0
These are the Pauli matrices, which arise in quantum mechanics.
Exercise 4.3.1. Show that the 2-dimensional irreducible representa­
tion of Q3 can be realized in the space of functions f : Q3 —> (C such
that f(gi) = \/ -1 f (g) (the action of G’ is by right multiplication,
9 0 f($) = f(rI:9))­

(4) The symmetric group S4. The order of S4 is 24, and there are
five conjugacy classes: e, (12), (123), (1234), (12)(34). Thus the sum
of the squares of the dimensions of five irreducible representations
is 24. As with S3, there are two of dimension 1: the trivial and
sign representations, (C+ and (C_. The other three must then have
dimensions 2, 3, and 3. Because S3 ‘é S4 /Z2 X Z2, where Z2 x Z2
is {e,(12)(34),(13)(24),(14)(23)}, the 2-dimensional representation
of S3 can be pulled back to the 2-dimensional representation of S4,
which we will call (C2.
We can consider S4 as the group of rotations of a cube acting by
permuting the interior diagonals (or, equivalently, on a regular octahe­
dron permuting pairs of opposite faces); this gives the 3-dimensional
representation (C1.
The last 3-dimensional representation is (Ci , the product of (C1
with the sign representation. (C1 and (C3 are difi'erent, for if g is
a transposition, detg|C§r = 1 while detg|Ca; = (—1)3 = -1. Note
that another realization of (CE is by action of S4 by symmetries (not
necessarily rotations) of the regular tetrahedron. Yet another real­
ization of this representation is the space of functions on the set of
4.5. Orthogonality of characters 65
four elements (on which S4 acts by permutations) with zero sum of
values.

4.4. Duals and tensor products of

representations
If V is a representation of a group G’, then V* is also a representation,
v1a
Pv*(9)=(Pv(9)*)_1 = (pv(9)‘1)* = pv(9‘1)*­
The character is XV. (g) = X‘/(g‘1).
We have Xx/(g) = Z2)‘,-, where the A, are the eigenvalues of g
in V. These eigenvalues must be roots of unity because p(g)'Gl =
p(g'G') = p(e) = Id. Thus for complex representations

Xv*(9) = Xv(9—1)= 2:)»-1: EX; = EN = Xv(9)­

In particular, V § V* as representations (not just as vector spaces)
if and only if Xx/(g) E R for all g E G’.
If V, W are representations of G, then V (8) W is also a represen­
tation, via
Pv®w(9) = Pv(9) 8’ Pw(9)­
Therefore, Xv®w(9) = Xv(9)Xw(9)­
An interesting problem discussed below is decomposing V (8) W
(for irreducible V, W) into the direct sum of irreducible representa­
tions.

4.5. Orthogonality of characters

We define a positive definite Hermitian inner product on F c(G', (C)
(the space of central functions) by

(flu f2) = i Z f1<g)?2’<.?z3.

IGI gee
The following theorem says that characters of irreducible represen­
tations of G’ form an orthonormal basis of FC(G, (C) under this inner
product.
66 4. Representations of finite groups: Basic results
Theorem 4.5.1. For any representations V, W
(XV) = HOInG'(‘/V‘) V):
and
(XV’XW
) _ 1, if“ V '3 W,
0, ifV;W
if V, W are irreducible.

Proof. By the definition

gEG g€G
(xv, xw) = fl Z Xv(9)Xw(9) = % Z xv(g)xw— (9)
1

=@ Zyea
Xv®w*(9) = TF|v®w*(P),

where P = lfi 2966 g E Z((C[G]). (Here Z((C[G]) denotes the center

of (C[G].) If X is an irreducible representation of G, then
IdifX=<C,
PIX: 0 X7é<C.
Therefore, for any representation X the operator P| X is the G-invari­
ant projector onto the subspace X G of G-invariants in X. Thus,

Tr |V®w-= (P) = dim HomG((C, V (8) W*)

= dim(V ® W*)G = dim HomG(l/V, V).
Cl

Theorem 4.5.1 gives a powerful method of checking if a given

complex representation V of a finite group G is irreducible. Indeed,
it implies that V is irreducible if and only if (XV, XV) = 1.
Problem 4.5.2. Let G be a finite group. Let V,- be the irreducible
complex representations of G.
For every i, let
__ dim V,­
¢z' — |G| ZXV.-(.9)°.9—1
gEG
€ClGl­
4.5. Orthogonality of characters 67
(i) Prove that w,- acts on as the identity if j = 2', and as the
null map ifj yé i.
(ii) Prove that zp, are idempotents; i.e., wf = zp, for any i, and
w,-wj = O for any 2' aéj.
Hint: In (i), notice that w, commutes with any element of k [G]
and thus acts on as an intertwining operator. Corollary 2.3.10 thus
yields that w,- acts on as a scalar. Compute this scalar by taking
its trace in
Remark 4.5.3. We see that characters of irreducible complex rep­
resentations of G can be defined without mentioning irreducible rep­
resentations. Namely, equip the space F (G, (C) of complex-valued
functions on G with the convolution product
(f * g)(z) = Z f(Iv)9(y)­
:n,y€G':a:y=z

This product turns F (G, (C) into an associative algebra, with unit
56 (the characteristic function of the unit 6 E G’), and the space
of class functions FC(G',(C) is a commutative subalgebra. Then one
can define renormalized characters )2, E FC(G', (C) to be the primitive
idempotents in this algebra, i.e., solutions of the equation f * f = f
which cannot be decomposed into a sum of other nonzero solutions.
Then one can define the characters by the formula

X.-(9) = ‘/ )Z|%1|)>Z.-(9)

(check it!). This is, essentially, how Frobenius defined characters (see
[Cu], equation Note that Frobenius defined representations at
approximately the same time, but for some time it was not clear
that there is a simple relation between irreducible representations
and characters (namely, that irreducible characters are simply traces
of group elements in irreducible representations). Even today, many
group theorists sometimes talk of irreducible characters of a finite
group rather than irreducible representations.

Here is another “orthogonality formula” for characters, in which

summation is taken over irreducible representations rather than group
elements.
68 4. Representations of finite groups: Basic results
Theorem 4.5.4. Let g, h E G, and let Z9 denote the centralizer of g
in G. Then
G _ |Zg|, if g is conjugate to h,
gm/(g)XV(h) _ { 0, otherwise,
where the summation is taken over all irreducible representations of
G.

Proof. As noted above, X‘/(h) = Xv: (h), so the left-hand side equals
(using Maschke’s theorem):

2 XV(9)Xv-= (h) = T? leav v®v- (9 8’ (h*)—1)

v
= T? laav Endx/(517 '—* 95191-1) = T? |<c[G](517 '—* 9517h_1)­

done. E]
If g and h are not conjugate, this trace is clearly zero, since the matrix
of the operator as »—> g:1:h‘1 in the basis of group elements has zero
diagonal entries. On the other hand, if g and h are in the same
conjugacy class, the trace is equal to the number of elements :1: such
that :1: -—= g:1:h‘1, i.e., the order of the centralizer Z9 of g. We are

Remark 4.5.5. Another proof of this result is as follows. Consider

the matrix U whose rows are labeled by irreducible representations of
G and whose columns are labeled by conjugacy classes, with entries
Uv,g = XV(g)/‘/ |Zg|. Note that the conjugacy class ofg is G/Z9; thus
|G| / |Zg| is the number of elements conjugate to G. Thus, by Theorem
4.5.1, the rows of the matrix U are orthonormal. This means that U
is unitary and hence its columns are also orthonormal, which implies
the statement.

4.6. Unitary representations. Another proof of

Maschke’s theorem for complex
representations
Definition 4.6.1. A unitary finite dimensional representation of
a group G is a representation of G on a complex finite dimensional
vector space V over (C equipped with a G-invariant positive definite
4.6. Unitary representations 69
Hermitian forml ( , ), i.e., such that pV(g) are unitary operators:
(pv(g)'v. pv(g)'w) = (v. w)­

Theorem 4.6.2. If G is finite, then any finite dimensional repre­

sentation of G has a unitary structure. If the representation is irre­
ducible, this structure is unique up to scaling by a positive real number.

Proof. Take any positive definite form B on V and define another

form B as follows:

B(v7w) : Z B(pV(g)v9pV(g)w)'
g€G
Then B is a positive definite Hermitian form on V, and pV(g) are
unitary operators. If V is an irreducible representation and B1, B2
are two positive definite Hermitian forms on V, then B1('u,w) =
B2(A'u,w) for some homomorphism A : V —> V (since any posi­
tive definite Hermitian form is nondegenerate). By Schur’s lemma,
A = )\Id, and clearly A > 0. D
Theorem 4.6.2 implies that if V is a finite dimensional representa­
tion of a finite group G, then the complex conjugate representa­
tion V (i.e., the same space V with the same addition and the same
action of G, but complex conjugate action of scalars) is isomorphic to
the dual representation V*. Indeed, a homomorphism of representa­
tions V —> V* is obviously the same thing as an invariant sesquilinear
form on V (i.e., a form additive on both arguments which is linear on
the first one and antilinear on the second one), and an isomorphism is
the same thing as a nondegenerate invariant sesquilinear form. So one
can use a unitary structure on V to define an isomorphism V —> V*.
Theorem 4.6.3. A finite dimensional unitary representation V of
any group G is completely reducible.

Proof. Let W be a subrepresentation of V. Let Wi be the orthogo­

nal complement of W in V under the Hermitian inner product. Then
Wi is a subrepresentation of W, and V = W 63 Wi. This implies
that V is completely reducible. D
1We agree that Hermitian forms are linear in the first argument and antilinear in
the second one.
70 4. Representations of finite groups: Basic results
Theorems 4.6.2 and 4.6.3 imply Maschke’s theorem for complex
representations (Theorem 4.1.1). Thus, we have obtained a new proof
of this theorem over the field of complex numbers.

Remark 4.6.4. Theorem 4.6.3 shows that for infinite groups G, a

finite dimensional representation may fail to admit a unitary structure
(as there exist finite dimensional representations, e.g., for G = Z,
which are indecomposable but not irreducible).

4.7. Orthogonality of matrix elements

Let V be an irreducible representation of a finite group G, and let
211,212, . . . ,u,, be an orthonormal basis of V under the invariant Her­
mitian form. The matrix elements of V are = (,oV(a:)'u,-, 21,-).
Proposition 4.7.1. ( Matria: elements of nonisomorphic irreducible
representations are orthogonal in F (G, (C) under the form (f, g) =
|%| 236G
One has (t,l§,tXj,) = 6,-,-,6,-J-I - #7.
Thus, matrix elements of irreducible representations of G form
an orthogonal basis of F(G, (C).

Proof. Let V and W be two irreducible representations of G. Take

{vi} to be an orthonormal basis of V and {w,-} to be an orthonormal
basis of W under their positive definite invariant Hermitian forms.
Let w;-" E W* be the linear function on W defined by taking the
inner product with w,-: = (u, w,-). Then for as E G we have
(azwf, = (asw,-, wj). Therefore, putting P = fl Exec :13, we have

(t,¥;,t,.V,V,.,) = |G'|‘1 Z(:vu,-,'uj)(:rw,-r,wjr)

:c€G

=|G|_1 Z(:ru,-,uj)(:rw;",,w;-‘,) = (P('u,- ®w;",),uj ®w;-3).

:cEG

If V 75 W, this is zero, since P projects to the trivial representation,

which does not occur in V (8 W*. If V = W, we need to consider
4.8. Character tables, examples 71
(P(’U¢ ® 'u;'‘,), 11,- 69 'U;‘,). We have a G’-invariant decomposition

V as V* = (C as L,

(C = span(Z 21;, (8 vi),

L = spanmzk akk=0(§ akyuk ® vf),
k,l

and P projects to the first summand along the second one. The
projection of 21,- ® 21;“, to (C C (C 69 L is thus
5w ,..
dim V Z 7”‘ 8’ '”"°
This shows that
5..,5. .,
(p (7)
1. =3; . ='.‘, =.___'‘ 99
®vz)avJ®vy)
which finishes the proof of and (ii). The last statement follows
immediately from the sum of squares formula. C]

4.8. Character tables, examples

The characters of all the irreducible representations of a finite group
can be arranged into a character table, with conjugacy classes of
elements as the columns and characters as the rows. More specifically,
the first row in a character table lists representatives of conjugacy
classes, the second one lists the numbers of elements in the conjugacy
classes, and the other rows list the values of the characters on the
conjugacy classes. Due to Theorems 4.5.1 and 4.5.4, the rows and
columns of a character table are orthonormal with respect to the
appropriate inner products.
Note that in any character table, the row corresponding to the
trivial representation consists of ones, and the column corresponding
to the neutral element consists of the dimensions of the representa­
tions.
Here is, for example, the character table of S3 :
72 4. Representations of finite groups: Basic results

#132
S3 Id (12) (123)
<C+ 1 1 1
(C- 1 -1 1
(C2 2 0 -1
It is obtained by explicitly computing traces in the irreducible repre­
sentations.
For another example consider A4, the group of even permutations
of four items. There are three 1-dimensional representations (as A4
has a normal subgroup Z2 69 Z2 and A4 / (Z2 69 Z2) = Z3). Since
there are four conjugacy classes in total, there is one more irreducible
representation of dimension 3. Finally, the character table is

#(C1141 41 31
A4 Id (123) (132) (12)(34)

(Ce 1 e 62 1
(C62
(C3 31 06206-1
1
where e = exp(2T"'-).
The last row can be computed using the orthogonality of rows.
Another way to compute the last row is to note that (C3 is the repre­
sentation of A4 by rotations of the regular tetrahedron: in this case
(123), (132) are the rotations by 1200 and 2400 around a perpendicu­
lar to a face of the tetrahedron, while (12)(34) is the rotation by 1800
around an axis perpendicular to two opposite edges.
Example 4.8.1. The following three character tables are of Q3, S4,
and A5, respectively:
Q3
# 1 11 -1
2 2] 2k
(C++ 1 1 1 1 1
<C+_ 1 1 1 -1 -1
(C_+ 1 1 -1 1 -1
C__ 1 1 -1 -1 1
(C2 2 -2 0 0
4.8. Character tables, examples 73
#16386
84 Id (12) (12)(34) (123) (1234)
<C+1 1-111111-1
1
C-
C2 2 0 2 -1 0
(C1 33-11 -1
<C3_ -1 00 -1
1
#<1:
1 20
1 15
1 112112
A5 Id (123) (12)(34) (12345) (13245)
1
(Ci 3 O _1 1+-25 1-25
C4
(C5 4 1 -1
5 -1 1 0-1
(C3_ 3 _1 1-2x/3 lizx/5-)
0
Indeed, the computation of the characters of the 1—dimensional
representations is straightforward.
The character of the 2-dimensional representation of Q8 is ob­
tained from the explicit formula (4.3.1) for this representation, or by
using orthogonality.
For S4, the 2-dimensional irreducible representation is obtained
from the 2—dimensional irreducible representation of S3 via the surjec—
tive homomorphism S4 —> S3, which allows one to obtain its character
from the character table of S3.
The character of the 3—dimensional representation (C1 is com­
puted from its geometric realization by rotations of the cube. Namely,
by rotating the cube, S4 permutes the main diagonals. Thus ( 12) is
the rotation by 1800 around an axis that is perpendicular to two op­
posite edges, (12)(34) is the rotation by 1800 around an axis that
is perpendicular to two opposite faces, (123) is the rotation around
a main diagonal by 1200, and (1234) is the rotation by 900 around
an axis that is perpendicular to two opposite faces; this allows us
to compute the traces easily, using the fact that the trace of a rota­
tion by the angle ¢ in R3 is 1 + 2cos ¢. Now the character of (C5: is
found by multiplying the character of (C1 by the character of the sign
representation.
74 4. Representations of finite groups: Basic results
Finally, we explain how to obtain the character table of A5 (even
permutations of five items). The group A5 is the group of rotations
of the regular icosahedron. Thus it has a 3-dimensional “rotation
representation” (C1, in which (12)(34) is the rotation by 1800 around
an axis perpendicular to two opposite edges, (123) is the rotation by
1200 around an axis perpendicular to two opposite faces, and (12345),
(13254) are the rotations by 720, respectively 1440, around axes going
through two opposite vertices. The character of this representation
is computed from this description in a straightforward way.
Another representation of A5, which is also 3-dimensional, is (Ci
twisted by the automorphism of A5 given by conjugation by (12) in­
side S5. This representation is denoted by (C51. It has the same char­
acter as (Ci, except that the conjugacy classes (12345) and (13245)
are interchanged.
There are two remaining irreducible representations, and by the
sum of squares formula their dimensions are 4 and 5. So we call them
(C4 and (C5.
The representation (C4 is realized on the space of functions on
the set {1, 2, 3,4, 5} with zero sum of values, where A5 acts by per­
mutations (check that it is irreduciblel). The character of this repre­
sentation is equal to the character of the 5—dimensional permutation
representation minus the character of the 1-dimensional trivial repre­
sentation (constant functions). The former at an element g is equal
to the number of items among 1, 2, 3, 4, 5 which are fixed by g.
The representation (C5 is realized on the space of functions on
pairs of opposite vertices of the icosahedron which has zero sum of
values (check that it is irreducible!). The character of this repre­
sentation is computed similarly to the character of (C4, or from the
orthogonality formula.

4.9. Computing tensor product multiplicities

using character tables
Character tables allow us to compute the tensor product multiplicities
N,-'3 using
V. 69 V7’ = ZN.’;Vk, N5; = (Xz'XjaXk)­
4.10. Frobenius determinant 75
Example 4.9.1. The following tables represent computed tensor
product multiplicities of irreducible representations of S3, S4, and A5,
respectively:
S3 C+ (C. (C2
<C+ C+ (C. (C2
(C. C+ (C2
(C2 (IL, 69 C- ea C2

54 C+ C- (C2 C1 C9;
<C+ C+ C. C2 C1 (C3.
C- C+ C2 C3 C1
(C2 <C+EB<C_€B<C2 C1eBC‘1 CE’;eaC‘i
CE’; C+eaC2eBC1eBC‘i C_e3C2eC1eC3.
Ci C+eaC2eaCE*;eBC‘i
A5 C CE’; C9;
C C C5; Cl C4 (C5 C4 C5
CE’; Ce9C5eaCE’; C4eaC5 C‘ie3C4a9C5 C~1e9C‘ie3C4eBC5
C9; CeaC5saCE*, CE‘,e>C4e9C5 CE1e9C3_eaC4e9C5
C4 C£‘,e9C3_e9C CE’;eaC‘ie92C5e3C4
(C5 C ea C1 ea C3. EBC4 ea C5

ea2C4 ea 2C5

4.10. Frobenius determinant

Enumerate the elements of a finite group G as follows: g1, g2, . . . , gn.
Introduce n variables indexed with the elements of G :
:rg,,:rg2,...,:rgn.
Definition 4.10.1. Consider the matrix XG with entries a,-j = crgigj.
The determinant of X g is some polynomial of degree n of $91, $92, . . . ,
azgn that is called the Frobenius determinant, or group deter­
minant.
76 4. Representations of finite groups: Basic results
The following theorem, discovered by Dedekind and proved by
Frobenius, became the starting point for creation of representation
theory (see [Cu] and Section 4.11).
Theorem 4.10.2.

det XG = H P]-(x)degP3'
j=1
for some pairwise nonproportional irreducible polynomials Pj (x),
where r is the number of conjugacy classes of G.

We will need the following simple lemma.

Lemma 4.10.3. Let Y be an n X n matrix with entries yij. Then

det Y is an irreducible polynomial of {yz-j}.

Proof. Let X = t -Id+§::l=1:1:z-Em-+1, where i + 1 is computed

modulo n, and Em are the elementary matrices. Then det(X) =
t" — (—1)”a:1 . ..a:n, which is obviously irreducible. Hence det(Y) is
irreducible (since it is so when Y is specialized to X, and since ir­
reducible factors of a homogeneous polynomial are homogeneous, so
cannot specialize to nonzero constants). C]
Now we are ready to proceed to the proof of Theorem 4.10.2.

Proof. Let V = (C[G] be the regular representation of G. Consider

the operator-valued polynomial

L(x) = 2 9399(9):
g€G'

where p(g) E EndV is induced by g. The action of L(x) on an element

h E G is

L(x)h = Z a:gp(g)h = Z xggh = Z a:_,_h—1z.

gEG' gEG' z€G'
So the matrix of the linear operator L(x) in the basis g1, g2, . . . , gn is
XG with permuted columns and hence has the same determinant up
to sign.
4.11. Historical interlude: Georg Frobenius 77
Further, by Maschke’s theorem, we have

detVL(x) = H(detV,L(x))dimW,
¢=1

where V, are the irreducible representations of G. We set P, =

detV,L(x). Let {em} be bases of V, and let E,-,3-k E EndV,- be the
matrix units in these bases. Then {Em-;,} is a basis of (C[G] and

L(X)|v.- = Z 3/z'.jkEz'.jk»
Jkk

where y,-,3-k are new coordinates on (C[G’] related to mg by a linear

transformation. Then
P, (x) = det V.-L(x) = det(yz',jk)'
Hence, P, are irreducible (by Lemma 4.10.3) and not proportional to
The theorem is proved. E]
each other (as they depend on different collections of variables y,-,3-k).

4.11. Historical interlude: Georg Frobenius’s

“Principle of Horse Trade”
Ferdinand Georg Frobenius (1849-1917) studied at the famous Berlin
University under both Karl Weierstrass and Leopold Kronecker, two
great mathematicians who later became bitter opponents. Weier­
strass considered Frobenius one of his brightest doctoral students and
greased the wheels of his career by securing him a full professorship at
the Zurich Polytechnikum. In Ziirich Frobenius quickly got married,
but the joys of happy matrimony did not prevent him from continu­
ing productive research. Frobenius earned a high reputation for his
studies of elliptic and theta functions, determinant and matrix the­
ory, and bilinear forms. Some fifteen years later, Kronecker’s passing
opened a vacant slot in Berlin, and Weierstrass got posthumous re­
venge on his old opponent by hiring his own student, Frobenius, for
Kronecker’s chair. In 1892 Frobenius left for Berlin, just four years
before Einstein enrolled as a student in the Zurich Polytechnikum.
Praised by his colleagues as “a first-rate stylist”, who “writes
clearly and understandably without ever attempting to delude the
78 4. Representations of finite groups: Basic results
reader with empty phrases”, Frobenius was soon elected to the presti­
gious Prussian Academy of Sciences [11, p. 38]. At this point, Frobe­
nius began reevaluating his research interests. As he explained in
his inaugural speech, the “labyrinth of formulas” in the theory of
theta functions was having “a withering effect upon the mathemat­
ical imagination”. He intended “to overcome this paralysis of the
mathematical creative powers by time and again seeking renewal at
the fountain of youth of arithmetic”, i.e., number theory (quoted in
[24, p. 220]). In his 40s, Frobenius indeed found this “fountain of
youth” in “arithmetic”, that is, in the theory of finite groups linked
to the theory of numbers by Galois theory.
In 1896 Frobenius’s comfortable life in Berlin was unceremoni­
ously disrupted by several letters from Richard Dedekind, his old ac­
quaintance and a predecessor at the Zurich Polytechnikum, now the
dean of abstract algebra in Germany. After dealing with epistolary
niceties and thanking Frobenius for brightening “the African dark­
ness of the theory of groups”, Dedekind shared some of his recent
results on group theory, including his concept of the group deter­
minant and the statement of his theorem about its factorization for
abelian groups. Since Dedekind had not bothered to publish his work
on the topic, Frobenius had never even heard of the group determi­
nant, but he quickly grasped this concept and never let it go again.
In his reply, Frobenius mildly chastised Dedekind for keeping back
these “extremely beautiful results from your friends and admirers”,
and asked for more details. Dedekind then stumped his colleague
with a “conjectured theorem” that the number of linear factors in
the group determinant is equal to the index of the commutator sub­
group and admitted, “I distinctly feel that I will not achieve anything
here” (quoted in [24, pp. 223, 224]).
Dedekind’s challenge was perfectly timed for the spring break,
and Frobenius, having some free time on his hands, immediately took
the bait. Employing military metaphors to describe his own work,
he waged a war against the group determinant. He invented gen­
eralized group characters and assaulted Dedekind’s conjecture with
their help. Frobenius’s initial results seemed unsatisfactory, and he
reported them to Dedekind with a disclaimer that “my conclusions
4.11. Historical interlude: Georg Frobenius 79
are so complicated that I myself do not rightly know where the main
point of the proof is, and in fact I am still slightly mistrustful of it”
(quoted in [24, pp. 225, 230]).
A few days later a jubilant Frobenius wrote to Dedekind that he
finally saw the way to a solution. Citing his former colleague Friedrich
Schottky, Frobenius now looked at his own earlier frustration as the
harbinger of a forthcoming breakthrough: “If in an investigation,
after difficult mental exertion, the feeling arises that nothing will be
achieved on the matter in question, then one can rejoice for he is
standing before the solution” (quoted in [24, p. 230]). Within ten
days, Frobenius proved all the main theorems of the theory of group
determinants, except for one most important result.
The missing theorem stated, in modern language, that an irre­
ducible representation occurs in the regular representation as often as
its degree and was called by Frobenius “the Fundamental Theorem of
the theory of group determinants”. “It would be wonderful” if it were
true, he wrote, “for then my theory would supply everything need­
ful” for the determination of prime factors (quoted in [24, p. 235]).
The proof took five months, during which Frobenius repeatedly used
a new, effective technique, which he kept secret and shared privately
only with Dedekind. The new method drew on the well-known bar­
gaining strategy: “At the market, the desired horse is ignored as much
as possible and at last is allowed to be formally recognized” (quoted
in [24, p. 236]). In Frobenius’s interpretation, in order to solve a
mathematical problem, one had to preoccupy oneself with activities
totally unrelated to mathematics. He applied this “Principle of the
Horse Trade” widely, taking his wife to trade exhibits and art shows,
reading fiction, and clearing his garden of caterpillars. Frobenius im­
plored Dedekind to keep this ingenious method a secret, promising to
disclose it in a posthumous volume, On the Methods of Mathematical
Research, with an appendix on caterpillar catching [24, p. 237]. Un­
fortunately, this volume never came Out, which was a huge setback
for the science of caterpillar-catching.
Miraculously, taking a break from research, combined with a
healthy dose of working desk disorder, did help Frobenius refresh his
80 4. Representations of finite groups: Basic results
thoughts and find a new approach. After returning home from vaca­
tion, Frobenius failed to find his earlier proof of one particular case
of the missing theorem among his “highly scattered and disorganized
papers”. After “much torment”, however, he discovered a new proof
and recognized the crucial possibility of generalization [24, p. 237].
The same year Frobenius announced his results to the world in
a series of papers. This was a wise move, since most of his corre­
spondence with Dedekind eventually ended up in the hands of an
American lawyer, who kept it in his drawer for thirty odd years and
parted with it only after his retirement by giving it to a mathematics
professor in exchange for $25 [27].
Wielding the group determinant as his main weapon, within a
few years Frobenius demolished a whole range of targets in repre­
sentation theory. Although his proofs have now been supplanted by
easier modern versions, his skills as a group determinant virtuoso
remain unsurpassed, as the tool itself went out of use.
Frobenius labored mightily to keep things at the University of
Berlin just the way they were in the glorious days of his student youth,
and he accused the advocates of applied mathematics of trying to re­
duce this venerable institution to the rank of a technical school. His
personality, which has been described as “occasionally choleric, quar­
relsome, and given to invectives” [21], alienated him from many of his
colleagues. Despite (or maybe partly due to?) Frobenius’s uncom­
promising stance, the numbers of doctorates and teaching staff at the
University gradually declined. Frobenius advised nineteen doctoral
students, and only one of them wrote a dissertation on representation
theory. Training that one student, however, proved fortuitous, for
that student was the brilliant mathematician Issai Schur, who kept
the Frobenius tradition alive for decades to come.
In the meantime, Berlin was losing its reputation as the leading
center of German mathematics to Gottingen. This did not endear
G<'5ttingen—style mathematics to Frobenius’s heart, and he disaffec­
tionately called it “a school, in which one amuses oneself more with
rosy images than hard ideas” (quoted in [11, p. 47]). His aversion to
the Gottingen patriarch Felix Klein and to Sophus Lie, according to
one commentator, “knew no limits” (Biermann in [11, p. 47]).
4.12. Problems 81
Despite Frobenius’s best efforts to keep the subject pure, repre­
sentation theory has since been irretrievably polluted by applications
in quantum physics and chemistry, crystallography, spectroscopy, and
even virology. Worse still, thanks to the efforts of Hermann Weyl
and Claude Chevalley, his beloved representation theory eventually
merged with the theory of Lie groups. Frobenius thus was eventually
reconciled with Lie, if only in the Platonic world of eternal mathe­
matical objects.

4.12. Problems
Problem 4.12.1. Let G be the group of symmetries of a regular
N -gon (it has 2N elements).
(a) Describe all irreducible complex representations of this group
(consider the cases of odd and even N
(b) Let V be the 2-dimensional complex representation of G ob­
tained by complexification of the standard representation on the real
plane (the plane of the polygon). Find the decomposition of V <8) V
in a direct sum of irreducible representations.
Problem 4.12.2. Let p be a prime. Let G be the group of 3 x 3 matri­
ces over IFP which are upper triangular and have 1’s on the diagonal,
under multiplication (its order is p3). It is called the Heisenberg
group. For any complex number z such that zp = 1, we define a
representation of G on the space V of complex functions on IFP by
1 10
(/9 01 0 f)(=v)=f(:r-1),
0 01
100
0 01 11f)(2=)=z"’f(:r)
(/9 (0
(note that 2"’ makes sense since zp = 1).
(a) Show that such a representation exists and is unique, and
compute p(g) for all g E G.
(b) Denote this representation by Rz. Show that Rz is irreducible
if and only if z aé 1.
82 4. Representations of finite groups: Basic results
(c) Classify all 1-dimensional representations of G. Show that R1
decomposes into a direct sum of 1-dimensional representations, where
each of them occurs exactly once.
(d) Use (a)——(c) and the “sum of squares” formula to classify all
irreducible representations of G.
Problem 4.12.3. Let V be a finite dimensional complex vector
space, and let GL(V) be the group of invertible linear transforma­
tions of V. Then S'”V and AmV (m g dim(V)) are representations
of GL(V) in a natural way. Show that they are irreducible represen­
tations.
Hint: Choose a basis {e,-} in V. Find a diagonal element H of
GL(V) such that p(H) has distinct eigenvalues (where p is one of the
above representations). This shows that if W is a subrepresentation,
then it is spanned by a subset S’ of a basis of eigenvectors of p(H
Use the invariance of W under the operators p(1 + E,-j) (where E,-j is
defined by E,-J-ck = 631.8,) for all 2' aé j to show that if the subset S’ is
nonempty, it is necessarily the entire basis.
Problem 4.12.4. Recall that the adjacency matrix of a graph I‘
(without multiple edges) is the matrix in which the ijth entry is 1 if
the vertices 2' and j are connected with an edge, and zero otherwise.
Let I‘ be a finite graph whose automorphism group is nonabelian.
Show that the adjacency matrix of I‘ must have repeated eigenvalues.
Problem 4.12.5. Let I be the set of vertices of a regular icosahedron
(|I| = 12). Let F (I) be the space of complex functions on I. Recall
that the group G = A5 of even permutations of five items acts on
the icosahedron, so we have a 12-dimensional representation of G on
F(I
(a) Decompose this representation in a direct sum of irreducible
representations (i.e., find the multiplicities of occurrence of all irre­
ducible representations).
(b) Do the same for the representation of G on the space of func­
tions on the set of faces and the set of edges of the icosahedron.
Problem 4.12.6. Let Fq be a finite field with q elements, and let G
be the group of nonconstant inhomogeneous linear transformations,
4.12. Problems 83
cc —> a:c+b, over Fq (i.e., a 6 IF; , b E Fq). Find all irreducible complex
representations of G, and compute their characters. Compute the
tensor products of irreducible representations.
Hint: Let V be the representation of G on the space of func­
tions on Fq with sum of all values equal to zero. Show that V is an
irreducible representation of G.
Problem 4.12.7. Let G = S’ U (2) (the group of unitary 2 X 2 matrices
with determinant 1), and let V = (C2 be the standard 2-dimensional
representation of SU We regard V as a real representation, so it
is 4—din1ensional.
(a) Show that V is irreducible (as a real representation).
(b) Let H be the subspace of End1R(V) consisting of endomor­
phisms of V as a real representation. Show that 11-11 is 4—dimensional
and closed under multiplication. Show that every nonzero element in
11-11 is invertible, i.e., 11-11 is an algebra with division.
(c) Find a basis 1, 2', j, k of H such that 1 is the unit and
z'2=j2=k2=—1,ij=—ji=k,jk=—kj=i, kz'=—z'k=j.
Thus we have that Q3 is a subgroup of the group 11-11)‘ of invertible
elements of 11-11 under multiplication.
The algebra 11-11 is called the quaternion algebra, and its el­
ements are called quaternions. Quaternions were discovered by
W. R. Hamilton in 1843 (see Section 4.13).
(d) For q = a,+bz'+cj+dk, a,b,c,d E R, let (I: a.—bz'—cj—dk
and ||q||2 = qr} = a2 + b2 + c2 + dz. Show that E = cjgcjl and
||q1q2|| = ||€11|| ° ||€12||­
(e) Let G be the group of quaternions of norm 1. Show that
this group is isomorphic to SU (Thus geometrically S’ U (2) is the
3-dimensional sphere.)
(f) Consider the action of G on the space V C 11-11 spanned by 2', j, k,
by cc —> q:cq"1, q E G, :3 E V. Since this action preserves the norm
on V, we have a homomorphism h : S'U(2) —> S'O(3), where S'O(3)
is the group of rotations of the 3-dimensional Euclidean space. Show
that this homomorphism is surjective and that its kernel is {1, -1}.
84 4. Representations of finite groups: Basic results
Problem 4.12.8. It is known that the classification of finite sub­
groups of 50(3) is as follows:
1) the cyclic group Z/'n,Z, n 2 1, generated by a rotation by 27r/n
around an axis;
2) the dihedral group D,, of order 277,, n 2 2 (the group of rota­
tional symmetries in 3—space of a plane containing a regular 'n,-gon2;
3) the group of rotations of a regular tetrahedron (A4);
4) the group of rotations of a cube or regular octahedron (S4);
5) the group of rotations of a regular dodecahedron or icosahedron
(A5)­
(a) Derive this classification.
Hint: Let G be a finite subgroup of 50(3). Consider the action
of G on the unit sphere. A point of the sphere preserved by some
nontrivial element of G is called a pole. Show that every nontriv­
ial element of G fixes a unique pair of opposite poles and that the
subgroup of G fixing a particular pole P is cyclic, of some order m
(called the order of P). Thus the orbit of P has n / m elements, where
n = |G Now let P1, . . . , Pk be a collection of poles representing all

1 1
the orbits of G on the set of poles (one representative per orbit), and
let ml, . . . , mk be their orders. By counting nontrivial elements of G,
showthat
21—— = 1——.
( .> :( m.)
Then find all possible mi and n that can satisfy this equation and
classify the corresponding groups.
(b) Using this classification, classify finite subgroups of SU (2)
(use the homomorphism S'U(2) —> S'O(3)).
Problem 4.12.9. Find the characters and tensor products of irre­
ducible complex representations of the Heisenberg group from Prob­
lem 4.12.2.

Problem 4.12.10. Let G be a finite group and let V be a complex

representation of G which is faithful, i.e., the corresponding map G —>

2A regular 2-gon is just a line segment.

4.12. Problems 85
GL(V) is injective. Show that any irreducible representation of G
occurs inside S'"V (and hence inside V®") for some n.

Hint: Show that there exists a vector u E V* whose stabilizer in

G is 1. Now define the map SV —+ F (G, C) sending a polynomial f
on V* to the function fu on G given by f.,(g) = f Show that
this map is surjective and use this to deduce the desired result.

Problem 4.12.11. This problem is about an application of repre­

sentation theory to physics (elasticity theory). We first describe the
physical motivation and then state the mathematical problem.
Imagine a material which occupies a certain region U in the phys­
ical space V = R3 (a space with a positive definite inner product).
Suppose the material is deformed. This means, we have applied a dif­
feomorphism (= change of coordinates) g : U —+ U’. The question in
elasticity theory is how much stress in the material this deformation
will cause.
For every point P E U, let Ap : V —+ V be defined by Ap =
dg(P). Here Ap is nondegenerate, so it has a polar decomposition
Ap = DpOp, where Op is orthogonal and Dp is symmetric. The
matrix Op characterizes the rotation part of A p (which clearly pro­
duces no stress), and D p is the distortion part, which actually causes
stress. If the deformation is small, Dp is close to 1, so Dp = 1 + dp,
where dp is a small symmetric matrix, i.e., an element of S'2V. This
matrix is called the deformation tensor at P.
Now we define the stress tensor, which characterizes stress. Let 1)
be a small nonzero vector in V, and let 0 be a small disk perpendicular
to v centered at P of area Let F” be the force with which the
part of the material on the v-side of 0 acts on the part on the opposite
side. It is easy to deduce from NeWton’s laws that F” is linear in 12,
so there exists a linear operator Sp : V —> V such that F” = S pv. It
is called the stress tensor.
An elasticity law is an equation Sp = f (dp), Where f is a func­
tion. The simplest such law is a linear law (Hooke’s law): f : S'2V —+
End(V) is a linear function. In general, such a function is defined by
9 - 6 = 54 parameters, but we will show there are actually only two
86 4. Representations of finite groups: Basic results
essential ones — the compression modulus K and the shearing
modulus h. For this purpose we will use representation theory.
Recall that the group S'O(3) of rotations acts on V, so S'2V,
End(V) are representations of this group. The laws of physics must
be invariant under this group (Galileo transformations), so f must be
a homomorphism of representations.
(a) Show that End(V) admits a decomposition IR 69 V69 W, where
IR is the trivial representation, V is the standard 3-dimensional rep­
resentation, and W is a 5-dimensional representation of SO(3). Show
that SZV = IR 69 W.
(b) Show that V and W are irreducible, even after complexifica­
tion. Deduce using Schur’s lemma that Sp is always symmetric, and
for as E R, y E W one has f(a: + y) = Kat + uy for some real numbers
K, p.
In fact, it is clear from physics that K, p are positive. Physically,
the compression modulus K characterizes resistance of the material to
compression or dilation, while the shearing modulus ,u characterizes
its resistance to changing the shape of the object without changing
its volume. For instance, clay (used for sculpting) has a large com­
pression modulus but a small shearing modulus.

4.13. Historical interlude: William Rowan

Hamilton’s quaternion of geometry,
algebra, metaphysics, and poetry
At age 17, William Rowan Hamilton’s interest in mathematics was
sparked by his discovery of an error in Laplace’s Celestial Mechan­
ics. The Royal Astronomer of Ireland was so impressed that in a few
years he secured for Hamilton the appointment as Royal Astronomer
at Dunsinsk Observatory, where Hamilton was stuck for the rest of his
life. Hamilton ploughed through university studies by winning every
conceivable honor and took his job at the Observatory even before
graduation, but as a practical astronomer he proved to be a failure.
Tedious observations did not appeal to him; he found theoretical sub­
jects much more exciting.
4.13. Historical interlude: William Rowan Hamilton 87
Hamilton’s social circle included major Romantic poets and philo­
sophers. Hamilton himself harbored poetic aspirations, but William
Wordsworth gently channeled his creative energies back to mathe­
matics. Hamilton immersed himself in the reading of Kant, absorbed
the Kantian notions of space and time as pure intuitions, and be­
came intrigued by Kant’s casual remark that just as geometry was
the “science of space”, algebra could be thought of as the “science
of pure time” [61]. Later Hamilton insisted that Kant merely con­
firmed his own ideas and reading him was more “recognizing” than
“discovering” [17 , 2:96, 2:98].
Hamilton made his reputation as a mathematician by his studies
in optics and mechanics, based on his notion of the characteristic
function. He saw his greatest achievement in reducing the solution of
3n ordinary differential equations of the second order to the solution of
two partial differential equations of the first order and second degree.
It was not obvious that this represented any progress toward actually
solving the problem, but Hamilton was convinced that even if “no
practical facility is gained” from his method, the reduction of all
complex calculations to the study of one characteristic function should
give one “an intellectual pleasure” (quoted in [18, p. 89]).
Hamilton divided all algebraists into three schools of thought:
the practical, the philological, and the theoretical. The practical
school viewed algebra as an art and was interested in computation;
the philological school viewed it as a language, a set of symbols to be
manipulated according to some rules; and only the theoretical school,
to which Hamilton modestly assigned himself, treated algebra as a
science, “strict, pure, and independent, deduced by valid reasonings
from its own intuitive principles” [17 , 2:48].
“I am never satisfied unless I think that I can look beyond or
through the signs to the things signified”, wrote Hamilton. He was
frustrated by the inability to find “things signified” behind the notions
of negative and imaginary numbers, which therefore looked absurd to
him. By 1828 he became “greatly dissatisfied with the phrases, if
not the reasonings, of even very eminent analysts”. He believed that
in order to “clear away the metaphysical stumbling-blocks that beset
the entrance of analysis”, one needed either to discard negative and
88 4. Representations of finite groups: Basic results
imaginary numbers or to explain their “true sense” [17, 2:143, 1:304].
Hamilton soon learned of the so-called Argand diagram representing
the complex number as a point on a plane, with its real and imaginary
parts plotted on two rectangular axes. Inspired by this geometrical
representation, he began looking for an algebraic representation of
complex numbers for which all valid operations could be defined and
soon developed the concept of algebraic couples. Echoing the Kan­
tian vision of algebra as a science of time, Hamilton viewed these
number couples as “steps” in time, rather than magnitudes, and he
interpreted negative signs as reversals of temporality [43, p. 281]. He
sought a new foundation of algebra in the intuitive notion of pure or
mathematical time: the moment was to algebra what the point was
to geometry, time intervals were finite straight lines, and an indefinite
straight line represented Time itself. Being awfully busy prevented
Hamilton from pursuing this promising line of reasoning further. As
he remarked, “Time is needed, with all its gross reality of hours and
days, even to write upon Pure Time” [17 , 2:144].
The 2-dimensional representation of complex numbers also in­
spired Hamilton to seek hypercomplex numbers related to the “real”
3-dimensional space, or “triplets”, as he called them. “Triplets” fit
nicely with his philosophical interest in the idealistic triad “thesis—
antithesis-synthesis” [20]. Adding triplets was easy, but they stub­
bornly refused to multiply. They did it for a good reason, for, as
Ffrobenius would prove only after Hamilton’s death, no such algebra
was possible. In desperation, Hamilton tried ordered sets of four num­
bers, or “quaternions”. They agreed to multiply only if the commu­
tative law was lifted. This realization came to Hamilton on October
16th, 1843, as he was walking with his wife along the Royal Canal in
Dublin. “I then and there felt the galvanic circuit of thought close;
and the sparks which fell from it were the fundamental equations”
of quaternions, he later recalled. In a less poetic metaphor, he com­
pared his satisfaction upon solving the problem to “an intellectual
want relieved” [17 , 2:435—436]. Hamilton immediately pulled out
his notebook and jotted down the fundamental quaternion formula.
Anxious to check the consistency of his new algebra, he continued
scribbling in a carriage on the way to a meeting of the Council of the
Royal Irish Academy and later while presiding over that meeting as
4.13. Historical interlude: Willia1n Rowan Hamilton 89

Academy President. Neither chatting with his wife nor chairing an

academic meeting apparently interfered with the train of his mathe­
matical thought.
Hamilton felt that he might explore the ramifications of the quater­
nion theory for 10-15 years — as long as he had spent trying to work
it out. He speculated that the scalar part of the quaternion might
express quantity (say, of electricity), while the imaginary parts might
determine direction and intensity, for example, electrical polarity. He
even gave the calculus of quaternions a “semi—metaphysical” inter­
pretation as a “calculus of polarities”, echoing his earlier interest in
idealist philosophy [17 , 2:436, 2:440]. In an even stronger metaphys­
ical vein, he came to view quaternions as a natural algebra of space
and time, in which the three dimensions of space were joined by the
fourth dimension of time. “The quaternion (was) born, as a curious
offspring of a quaternion of parents, say of geometry, algebra, meta­
physics, and poetry”, he wrote and quoted his own sonnet addressed
to Sir John Herschel as the clearest expression of the quaternion idea:
“And how the one of Time, of Space the Three,
Might in the Chain of Symbols girdled be”
(quoted in [20, p. 176]).
Hamilton did everything he could to make his theory unpalat­
able to the reader, choosing from the outset a “metaphysical style
of expression”. Herschel implored him to make his ideas “clear and
familiar down to the level of ordinary unmetaphysical apprehension”
and to “introduce the new phrases as strong meat gradually given to
babes” [17 , 2:633], but to no avail. Hamilton loaded his 800-page-long
Lectures on Quaternions (1853) with new impenetrable terminology,
such as vector, vehend, vection, vectum, revector, revehend, revec—
tion, revectum, provector, transvector, factor, profactor, versor, and
quandrantal versor [10, p. 36]. His attempt at a more basic introduc­
tion, Elements of Quaternions, ended up being even longer than the
Lectures.
Quaternions were popularized by Hamilton’s disciple Peter Tait,
who fought a protracted battle with Josiah Willard Gibbs and Oliver
Heaviside, promoters of the rival vector analysis. If quaternion multi­
plication lacked commutativity, the dot product and the cross product
90 4. Representations of finite groups: Basic results
of vectors seemed to have even greater problems, and Tait branded
vector analysis a “hermaphrodite monster” [10, p. 185].
Telling his great discovery story to his son 23 years after the
fact, Hamilton added that at his eureka moment he could not “resist
the impulse — unphilosophical as it may have been — to cut (the
quaternion formula) with a knife on a stone” of a nearby bridge [17 ,
2:435]. Whether this was true or Hamilton simply could not resist
the impulse to embellish his story, we will never know, as Time has
long erased the etching off the bridge stone — but not off the annals
of mathematics. Today the site of Hamilton’s vandalism is marked by
a plaque which reads, “Here as he walked by on the 16th of October
1843 Sir William Rowan Hamilton in a flash of genius discovered the
fundamental formula for quaternion multiplication 2'2 = j2 = k2 =
ijk = -1 & cut it on a stone of this bridge”. The bridge is now a
pilgrimage site for mathematicians seeking to ignite their imagination
off the sparks of the original flash.
Chapter 5

Representations of finite
groups: Further results

5.1. Frobenius-Schur indicator

Suppose that G is a finite group and V is an irreducible representation
of G over (C.

Definition 5.1.1. We say that V is

- of complex type if V 3.5 V*,
- of real type if V has a nondegenerate symmetric form in­
variant under G,
- of quaternionic type if V has a nondegenerate skew form
invariant under G.

Problem 5.1.2. (a) Show that Endmg] V is (C for V of complex type,

Mat; (R) for V of real type, and H for V of quaternionic type, which
motivates the names above.
Hint: Show that the complexification Vc of V decomposes as
V EB V*. Use this to compute the dimension of Endmg] V in all three
cases. Using the fact that (C C Endmg] V, prove the result in the
complex case. In the remaining two cases, let B be the invariant
bilinear form on V and let ( , ) be the invariant positive Hermitian
form (they are defined up to a nonzero complex scalar and a positive

91
92 5. Representations of finite groups: Further results
real scalar, respectively). Define the operator j : V —> V such that
B(u, w) = (u,jw). Show that j is complex antilinear (ji = —ij), and
3'2 = /\ - Id, where /\ is a real number, positive in the real case and
negative in the quaternionic case (if B is renormalized, j multiplies by
a nonzero complex number 2 and j2 by 22, as j is antilinear). Thus
j can be normalized so that j2 = 1 in the real case and j2 = -1 in
the quaternionic case. Deduce the claim from this.
(b) Show that V is of real type if and only if V is the complexi­
fication of a representation VR over the field of real numbers.
Example 5.1.3. For Z/nZ all irreducible representations are of com­
plex type except the trivial one and, if n is even, the “sign” representa­
tion, m -—> (—1)"‘, which are of real type. For S3 all three irreducible
representations (C+, (C_, (C2 are of real type. For S4 there are five ir­
reducible representations (C+, ((3., (C2, (C1, (C5: , which are all of real
type. Similarly, all five irreducible representations of A5 — (C, (C1,
(C3; , (C4, (C5 — are of real type. As for Q3, its 1-dimensional represen­
tations are of real type, and the 2-dimensional one is of quaternionic
type.
Definition 5.1.4. The Frobenius-Schur indicator FS(V) of an
irreducible representation V is 0 if it is of complex type, 1 if it is of
real type, and -1 if it is of quaternionic type.

Theorem 5.1.5 (Frobenius-Schur). The number of inuolutions (=

elements of order 3 2) in G is equal to EV dim(V)FS(V), i.e., the
sum of dimensions of all representations of G of real type minus the
sum of dimensions of its representations of quaternionic type.

Proof. Let A : V —> V have eigenvalues /\1, /\2, . . . , x\,,. We have

Tr |s2v(A <3 A) = Z /\z'/\j,
iSj
Tr|A2V(A®A) = Z:/\.,;/\,-.
i<j
Thus,
TT|s2v(A®A) — TrlA2V(A®/1): Z X? = Tr(A2)­
lfiign
5.2. Algebraic numbers and algebraic integers 93
Thus for g E G we have
Xv(92) = xs2v(g) - XA2V(9)°
Therefore, setting P = |G|‘1 Zgea g, we get
IGr1xv<Z 92> = Xs2v(P) — XA2v(P) = dim<s2v>G — dim<A2v>G
g€G

1 if V is of real type,
= -1 if V is of quaternionic type,
0 if V is of complex type.
Finally, the number of involutions in G equals

g€ G realZ V
|—g|Zv:dimVXV(Zg2)= quat.ZVdimV.
dimV—
[3

Corollary 5.1.6. Assume that all representations of a finite group G

are defined over real numbers (i.e., all complea: representations of G
are obtained by compleazifying real representations Then the sum of
the dimensions of all the irreducible representations of G equals the
number of involutions in G.
Exercise 5.1.7. Show that any nontrivial finite group of odd order
has an irreducible representation which is not defined over IR (i.e., not
realizable by real matrices).

5.2. Algebraic numbers and algebraic integers

We are now passing to deeper results in the representation theory of
finite groups. These results require the theory of algebraic numbers,
which we will now briefly review.
Definition 5.2.1. 2 E (C is an algebraic number (respectively, an
algebraic integer ) if z is a root of a monic polynomial with rational
(respectively, integer) coefficients.
Definition 5.2.2. 2 E (C is an algebraic number, (respectively, an
algebraic integer), if z is an eigenvalue of a matrix with rational
(respectively, integer) entries.
94 5. Representations of finite groups: Further results
Proposition 5.2.3. Definitions (5.21) and are equivalent.
Proof. To show that the condition of Definition 5.2.2 implies the
condition of Definition 5.2.1, notice that z is a root of the charac­
teristic polynomial of the matrix (a monic polynomial with rational,
respectively integer, coefficients). To establish the converse, suppose
z is a root of
p(:I:) = 51:" + a1:I:"“1 + . . . + a,,_1:I: + an.
Then the characteristic polynomial of the following matrix (called the
companion matrix) is p(:I:):

0 O O 0 —an
1 0 0 0 —an_1
0 1 0 0 —-an_2
0 0 0 1 —a1
Since 2 is a root of the characteristic polynomial of this matrix, it is
an eigenvalue of this matrix. C]
The set of algebraic numbers is denoted by Q, and the set of
algebraic integers is denoted by A.
Proposition 5.2.4. A is a ring.
@ is a field. Namely, it is an algebraic closure of the field of
rational numbers.

Proof. We will be using Definition 5.2.2. Let a be an eigenvalue of

A 6 Mat” (C)
with eigenvector v, and let 3 be an eigenvalue of
B E Matm (C)
with eigenvector 21). Then a :1: fl is an eigenvalue of
A <8) Idm :1: Id” (813,
5.2. Algebraic numbers and algebraic integers 95
and 0:3 is an eigenvalue of
A <8) B.

The corresponding eigenvector is in both cases v®w. This shows that

both A and Q are rings. To show that the latter is a field, it suffices
to note that if a gé 0 is a root of a polynomial p(:I:) of degree d, then
a‘1 is a root of £I3dp(l./ The last statement is easy, since a number
a is algebraic if and only if it defines a finite extension of Q. E]
Proposition 5.2.5. A D Q = Z.
Proof. We will be using Definition 5.2.1. Let 2 be a root of
p(:c) = 51:" + a1:r:"”1 + . . . + a,,_1:I: + an,
and suppose
z= 3 ea gcd(p,q>=1.
Notice that the leading term of p(:I:) will have q" in the denominator,
whereas all the other terms will have a lower power of q there. Thus,
if q gé :l:1, then p(z) ¢ Z, a contradiction. Thus, 2 E A H Q => 2 E Z.
The reverse inclusion follows because n E Z is a root of :1: — n. Cl

Every algebraic number a has a minimal polynomial p(:I:)

which is the monic polynomial with rational coefficients of the small­
est degree such that p(a) = 0. Any other polynomial q(:I:) with ra­
tional coefficients such that q(a) = 0 is divisible by Roots of
p(:I:) are called the algebraic conjugates of a; they are roots of any
polynomial q with rational coefficients such that q(a) = 0.
Note that any algebraic conjugate of an algebraic integer is obvi­
ously also an algebraic integer. Therefore, by the Vieta theorem, the
minimal polynomial of an algebraic integer has integer coefficients.
Below we will need the following lemma:
Lemma 5.2.6. If a1, . . . ,am are algebraic numbers, then all alge­
braic conjugates to a1 +- - -+am are of the form 0/1 +- - -+a;,,, where
oz; are some algebraic conjugates of a,-.

Proof. It suffices to prove this for two summands. If 0:, are eigenval­
ues of rational matrices A, of smallest size (i.e., their characteristic
96 5. Representations of finite groups: Further results
polynomials are the minimal polynomials of a,-), then 0:1 + 0:2 is an

so we are done. El
eigenvalue of A := A1 <8) Id +Id ®A2. Therefore, so is any algebraic
conjugate to a1 +a2. But all eigenvalues of A are of the form a'1+a§,

Problem 5.2.7. (a) Show that for any finite group G there exists a
finite Galois extension K C (C of Q such that any finite dimensional
complex representation of G has a basis in which the matrices of the
group elements have entries in K.
Hint: Consider the representations of G over the field Q of alge­
braic numbers.
(b) Show that if V is an irreducible complex representation of a
finite group G of dimension > 1, then there exists g E G such that
Xv(9) = 0­
Hint: Assume the contrary. Use orthonormality of characters to
show that the arithmetic mean of the numbers |XV(g)|2 for g ;£ 1 is
< 1. Deduce that their product fl satisfies O < 3 < 1. Show that all
conjugates of 3 satisfy the same inequalities (consider the Galois con­
jugates of the representation V, i.e., representations obtained from V
by the action of the Galois group of K over Q on the matrices of group
elements in the basis from part (a)). Then derive a contradiction.

Remark 5.2.8. Here is a modification of this argument, which does

not use (a). Let N = For any 0 < j < N coprime to N,
show that the map g »—> gj is a bijection G —> G. Deduce that
]—[g¢1|XV(gj)|2 = 3. Then show that fl E K := Q((), C = e27”./N,
and that it does not change under the automorphism of K given by
C »—> (3. Deduce that fl is an integer, and derive a contradiction.

5.3. Frobenius divisibility

Theorem 5.3.1. Let G be a finite group, and let V be an irreducible
representation of G over (C. Then

dim V divides |G
5.3. Frobenius divisibility 97
Proof. Let C1, C2, . . . , C” be the conjugacy classes of G’. Let go, be
representatives of C,-. Set
C2’
M = Xv(9c.-)#­
Proposition 5.3.2. The numbers A, are algebraic integers for all 2'.

Proof. Let C be a conjugacy class in G, and let P = Zhec h. Then

P is a central element of Z[G], so it acts on V by some scalar A, which
is an algebraic integer (indeed, since Z[G] is a finitely generated Z­
module, any element of Z[G] is integral over Z, i.e., satisfies a monic

A = 121122. E]
polynomial equation with integer coefficients). On the other hand,
taking the trace of P in V, we get |C|XV(g) = /\dim V, g E C, so
dim V

Now, consider

2 )\z'XV(9C¢)'
This is an algebraic integer, since:
(i) A, are algebraic integers by Proposition 5.3.2,
(ii) X1/(gci) is a sum of roots of unity (it is the sum of eigenvalues
of the matrix of p(gC,), and since glgl = e in G, the eigenvalues of
p(gC1.) are roots of unity), and
(iii) A is a ring (Proposition 5.2.4).
On the other hand, from the definition of A,-,

Z:)\iXV(gCi) = Z: lCilXV(gCi)XV(gCi).
dim V
Ci 2'

Recalling that XV is a class function, this is equal to

dimV dimV °
Z XV(9)XV(9) = lGl(XVaXV)
gEG

—G
Since V is an irreducible representation, (X1/,XV) = 1, so
98 5. Representations of finite groups: Further results

d,',%',, e Z. III
Since dJi%LV E Q and E01. /\,-X1/(gci) E A, by Proposition 5.2.5,

5.4. Burnside’s theorem

Definition 5.4.1. A group G is called solvable if there exists a series
of nested normal subgroups
{€}=G1 <1 G2 <1 <1 Gn=G
Where G,-+1/G1 is abelian for all 1 3 i 3 n — 1.
Remark 5.4.2. Such groups are called solvable because they first
arose as Galois groups of polynomial equations which are solvable in
radicals.
Theorem 5.4.3 (Burnside). Any group G of order paqb, where p and
q are primes and a, b 2 0, is solvable.
This famous result in group theory was proved by the British
mathematician William Burnside in the early 20th century, using
representation theory (see Section 5.5 and [Cu]). Here is this proof,
presented in modern language.
Before proving Burnside’s theorem, we will prove several other
results which are of independent interest.
Theorem 5.4.4. Let V be an irreducible representation of a finite
group G and let G be a conjugacy class of G with gcd(|C|, dim(V)) =
1. Then for any g E C, either X1/(g) = 0 or g acts as a scalar on V.
The proof will be based on the following lemma.
Lemma 5.4.5. If 51,52, . . . ,5” are roots of unity such that
1
;(51 +52 +- - -+5”) is an algebraic integer, then either 51 = - - - = 5n
or51+---+5., =0.
Proof. Let a = %(51 + - - - +5”). If not all 5, are equal, then |a| < 1.
Moreover, since any algebraic conjugate of a root of unity is also a
root of unity, |a’| 3 1 for any algebraic conjugate a’ of a. But the
product of all algebraic conjugates of a is an integer. Since it has
absolute value < 1, it must equal zero. Therefore, a = O. D
5.4. Burnside’s theorem 99
Proof of Theorem 5.4.4. Let dim V = n. Let 51,52, . . . ,5” be the
eigenvalues of pv They are roots of unity, so XV (g) is an algebraic
integer. Also, by Proposition 5.3.2, %|C'|xv (g) is an algebraic integer.
Since gcd(n, |C = 1, there exist integers a, b such that a|C I +bn = 1.
This implies that

T +bXV(9)
‘1lC|XV(9)
- _E(51‘l‘°°°"l"5n)
: Xv(9) 1

is an algebraic integer. Thus, by Lemma 5.4.5, we get that either

51 = = 5,, or 51 +---+8.. = X1/(g) = 0. In the first case,
since pv (g) is diagonalizable, it must be scalar. In the second case,
X1/(g) = 0. The theorem is proved. D
Theorem 5.4.6. Let G be a finite group, and let C be a conjugacy
class in G of order pk where p is a prime and k > 0. Then G has a
proper nontrivial normal subgroup (i.e., G is not simple

Proof. Choose an element g E G. Since g 79 e, by orthogonality of

columns of the character table,

(5.4.1) 2 dim VXV(g) = 0.

VEIrrG

We can divide IrrG into three parts:

(1) the trivial representation,
(2) D, the set of irreducible representations whose dimension is
divisible by p, and
(3) N, the set of nontrivial irreducible representations whose
dimension is not divisible by p.

Lemma 5.4.7. There exists V E N such that X1/(g) 79 0.

Proof. If V E D, the number % dim(V)XV (g) is an algebraic integer,

so

a = 2:1-1)dim(V)Xv(9)
VED
is an algebraic integer.
100 5. Representations of finite groups: Further results
Now, by (5.4.1), we have

VED VEN
0 = X<c(9) + Z dim VXv(9) + Z dim VX1/(9)

=1+pa + Z dim VX1/(g).

VEN
This means that the last summand is nonzero. Cl
Now pick V E N such that X1/(g) 79 0; it exists by Lemma 5.4.7.
Theorem 5.4.4 implies that g (and hence any element of C) acts by a
scalar in V. Now let H be the subgroup of G generated by elements
ab’1, a,b E C. It is normal and acts trivially in V, so H 75 G, as V
is nontrivial. Also H 75 1, since |C'| > 1. D
Proof of Burnside’s theorem. Assume Burnside’s theorem is false.
Then there exists a nonsolvable group G of order paqb. Let G be the
smallest such group. Then G is simple, and by Theorem 5.4.6, it
cannot have a conjugacy class of order pk or qk, k 2 1. So the or­
der of any conjugacy class in G either equals 1 or is divisible by pq.
Adding the orders of conjugacy classes and equating the sum to paqb,

contradiction. E]
we see that there has to be more than one conjugacy class consist­
ing just of one element. So G has a nontrivial center, which gives a

5.5. Historical interlude: William Burnside and

intellectual harmony in mathematics
While at Cambridge, William Burnside (1852-1927) distinguished
himself in rowing; his other achievements include emerging from the
1875 Mathematical T‘ripos as Second Wrangler and then beating First
Wrangler in an even more grueling mathematical competition for the
Smith Prize. Afterwards, he taught at Cambridge as a mathematics
lecturer and a coach for both the Math Tripos and for the rowing
crews. In 1885, true to his aquatic interests, Burnside accepted the
position of professor of mathematics in the Royal Naval College at
Greenwich, where he taught until retirement. When his enthusiasm
for rowing subsided, fishing became Burnside’s favorite hobby. Even
5.5. Historical interlude: William Burnside 101
in mathematical research he stayed close to water, making major con­
tributions to hydrodynamics.
On the strength of his contributions to mathematical physics and
complex function theory, Burnside was elected to the Royal Society
in 1893. Once this worthy goal was achieved, however, he abandoned
such trifle subjects and dipped into the theory of groups. Four years
later Burnside published the first English textbook on the subject,
Theory of Groups of Finite Order. He was apparently delighted to
take a break from applied studies and to immerse himself in an ab­
stract theory, for he wrote in the preface: “The present treatise is
intended to introduce to the reader the main outlines of the the­
ory of groups of finite order apart from any applications”. Burnside
noted that group theory was not yet particularly popular in England.
“It will afford me much satisfaction”, he remarked, “if, by means of
this book, I shall succeed in arousing interest among English mathe­
maticians in a branch of pure mathematics which becomes the more
fascinating the more it is studied” (quoted in [11, pp. 88-89]).
The interest of English mathematicians, however, proved not to
be easily aroused, and ten years later Burnside bitterly remarked in
his retiring Presidential address to the London Mathematical Soci­
ety: “It is undoubtedly the fact that the theory of groups of finite
order has failed, so far, to arouse the interest of any but a very small
number of English mathematicians”. Burnside cited the proliferation
of courses on group theory in France, the United States, and espe­
cially Germany (attended by thirty students in Gottingenl), and he
lamented the total indifference of British students toward the sub­
ject. His explanation was that group theory was treated in a highly
abstract manner, “one which the young mind grasps with difficulty,
if at all”. As an example, Burnside cited a formal statement about
the properties of the icosahedral group and claimed that “a proposal
to verify the statement appears equivalent to proposing a series of
conundrums. There would be nothing here to attract the student or
to suggest anything but the driest formalism utterly divorced from
any of his previous mathematical studies” [8, pp. 1, 3, 5].
Burnside himself, however, had done much to establish the cul­
ture of purely abstract reasoning in group theory. In the preface to
102 5. Representations of finite groups: Further results
his 1897 book he wrote that his university teacher Arthur Cayley’s
“dictum that ‘a group is defined by means of the laws of combina­
tion of its symbols’ would imply that, in dealing with the theory of
groups, no more concrete mode of representation should be used than
is absolutely necessary” (quoted in [11, p. 90]). No wonder Burnside
omitted any applications from his book, while his 1899 article on the
simple group of order 504 mentioned a concrete example only in the
last paragraph [1, p. 13]. He cultivated conciseness as a highest virtue.
When a friend once asked for a more expanded treatment of certain
topics from Theory of Groups, Burnside responded by “a declaration
of regret that he had been unable to effect further condensation” [15,
p. 70]. Burnside’s ideal lived on in the tradition of Bourbaki, causing
the wrath of the champions of “mathematics with a human face”,
led by Vladimir Arnold: “Algebraists usually define groups as sets
with operations that satisfy a long list of hard-to—remember axioms.
I think one cannot understand such a definition. I believe the alge­
braists set up such obstacles in the path of students to make it harder
for the uninitiated to penetrate their field. Perhaps their goal, if only
subconscious, is to boost the reputation of their field” [2, p. 118].
While pursuing the condensation ideal, Burnside decided to omit
any discussion of linear substitution groups from the 1897 edition of
Theory of Groups. “It would be difficult to find a result that could be
most directly obtained by the consideration of groups of linear trans­
formations”, he wrote, justifying the exclusion of this useless subject
(quoted in [11, p. 90]). Within a few months, however, Burnside had
to reevaluate the wisdom of his decision, as he came across Frobenius’s
articles on group characters. Frobenius’s results proved highly rele­
vant to Burnside’s own research on finite groups, and Burnside set out
to reformulate them in his own language. Unlike Frobenius, Burnside
felt at ease with Sophus Lie’s apparatus of continuous groups of trans­
formations, and he was able to derive all of Frobenius’s main results
on characters and on the group determinant by using the methods
of Lie groups and Lie algebras. Burnside published his research with
a modest disclaimer that his paper was “not original, as the results
arrived at are, with one or two slight exceptions, due to Herr Probe­
nius. The modes of proof, however, are in general quite distinct from
those used by Herr Frobenius” (quoted in [11, p. 106]).
5.5. Historical interlude: William Burnside 103
Herr Frobenius was hardly impressed by what he saw as a lame
excuse for stepping on his toes. Richard Dedekind was similarly out­
raged, when he recognized in one of Burnside’s papers his own the­
orem on the factorization of the group determinant of an abelian
group. Frobenius consoled his friend by recounting his own losses:
“This is the same Herr Burnside who annoyed me several years ago
by quickly rediscovering all the theorems I had published on the the­
ory of groups, in the same order and without exception: first my proof
of Sylow’s Theorems, then the theorem on groups with square-free or­
ders, on groups of order paq, on groups whose order is a product of
four or five prime numbers, etc., etc. In any case, a very remark­
able and amazing example of intellectual harmony, probably as is
possible only in England and perhaps America” (quoted in [24, p.
242]). Burnside, for his part, began stressing that he had “obtained
independently the chief results of Prof. Frobenius’ earlier memoirs”
(quoted in [23, p. 278]). Herren Burnside and Frobenius never cor­
responded to straighten things out, leaving the matter to historians,
who somewhat qualified Burnside’s claim of independence. Burnside
was clearly inspired by Frobenius’s work, although he did not know
all of the relevant Frobenius papers, which left him enough space to
explore on his own [23, p. 278].
Burnside and Frobenius worked neck and neck on the solvability
of paqfi groups. While Sylow (1872), Frobenius (1895), and Jordan
(1898) proved some special cases, Burnside succeeded in proving the
general case in 1904. Burnside’s character theoretic proof has been
described as “so easy and pleasant” that later group-theoretic proofs
would not even come close to its “compelling simplicity” and “strik­
ing beauty” [30, p. 469]. As Walter Feit suggested, “[T]he elegance
of both the statement and the proof have attracted more people to
the study of characters than any other result in the subject” [14, p.
4]. In particular, they attracted Feit, who in 1962 proved (with J.
G. Thompson) Burnside’s conjecture that every group of odd order
is solvable. Another seminal conjecture, Burnside’s Problem, related
to the structure of torsion groups, has preoccupied group theory and
representation theory specialists for over a century, yielding the 1994
Fields Medal to Efim Zelmanov for the solution of its restricted ver­
sion.
104 5. Representations of finite groups: Further results
The appreciation of the beauty of Burnside’s work took quite a
while. His employment at the Royal Naval College separated him
geographically from his university colleagues. He apparently had no
“extensive direct contacts with other mathematicians interested in
the subject (of group theory). It appears that he worked in isolation,
possibly even more so than was normal for his times, with little op­
portunity (or, perhaps, inclination) to discuss his ideas with others”,
according to his biographer [41, p. 32]. Burnside taught several gen­
erations of navy officers but created no mathematical school of his
own.
In December 1925 Burnside suffered a stroke, and his doctor for­
bade him, among other unhealthy activities, from doing mathematics.
Burnside naturally disobeyed and did not live very long. His obitu­
ary in the London Evening News barely mentioned his mathematical
studies but reported that “rowing men will regret to hear of the death
of W. Burnside, one of the best known Cambridge athletes of his day”
(quoted in [11, p. 96]).
Shortly before his death Burnside answered a query from a young
mathematician named Philip Hall, who asked for advice on topics of
group theory. Burnside sent him a postcard listing a few problems
worth investigating. This message in a bottle, thrown into the sea,
miraculously found a perfect addressee. The same volume of the Jour­
nal of the London Mathematical Society that contained Burnside’s
obituary featured Hall’s “Note on Soluble Groups”, which marked
the beginning of his lifetime career in this field. Hall eventually suc­
ceeded Burnside as the chief promoter of group theory in England.
“The aim of my researches”, he later wrote, “has been to a very con­
siderable extent that of extending and completing in certain directions
the work of Burnside” (quoted in [11, p. 96]).

5.6. Representations of products

Theorem 5.6.1. Let G, H be finite groups, let be the irreducible
representations of G over a field lc (of any characteristic), and let
{W3-} be the irreducible representations of H over lc. Then the irre­
ducible representations of G X H over lc are {V,- (8 W]-}.
5.8. Induced representations 105
Proof. This follows from Theorem 3.10.2. CI
5.7. Virtual representations
Definition 5.7.1. A virtual representation of a finite group G
is an integer linear combination of irreducible representations of G,
V = Zn,-V,-, n,- E Z (i.e., n,- are not assumed to be nonnegative).
The character of V is XV := Z n,-Xvi.
The following lemma is often very useful (and will be used several
times below).
Lemma 5.7.2. Let V be a virtual representation with character XV.
If (X1/,XV) = 1 and X1/(1) > 0, then xv is a character of an irre­
ducible representation of G.

Proof. Let V1,V2, . . .,Vm be the irreducible representations of G,

and let V = Z n,-V,-. Then by orthonormality of characters, (xv, xv) =
2,. So = 1, meaning that n,- = :|:1 for exactly one i and
nj = O for j 7é i. But X1/(1) > 0, so n,- = +1 and we are done. E]

5.8. Induced representations

Given a representation V of a group G and a subgroup H C G, there
is a natural way to construct a representation of H. The restriction
of V to H, ResgV is the representation given by the vector space V,
and the action pReSgV = pV| H.
There is also a natural, but less trivial, way to construct a rep­
resentation of a group G given a representation V of its subgroup
H.
Definition 5.8.1. If G is a group, H C G, and V is a representation
of H, then the induced representation Ind,G,,V, is the representa­
tion of G with
Indfiv = {f : G —> V|f(ha:) = pV(h)f(a:) Va: e G,h e H}
and the action = f(:cg) \7’g E G.
Remark 5.8.2. In fact, Ind,G,,V is naturally isomorphic to the repre­
sentation HomH (k[G], V).
106 5. Representations of finite groups: Further results
Let us check that Indg V is well defined as a representation. In­
deed, we have
9(f)(h5'«‘) = f(h5'«‘9) = pv(h)f(939) = pv(h)9(f)(5'3),
and
9(9’(f))(:v) = 9’(f)(:v9) = f (:v9g’) = (9g’)(f)(:v)
for any g,g’,:I: E G and h E H.
Remark 5.8.3. Notice that if we choose a representative cc, from ev­
ery right H -coset 0 of G, then any f E Ind,G,V is uniquely determined
by {f (:va)}­
Because of this,

dim(1nd,‘;‘,V) = dim V ­
Problem 5.8.4. Check that if K C H C G are groups and if V is a
representation of K, then Indfi Indfi V is isomorphic to Indfi V.
Exercise 5.8.5. Let K C G be finite groups, and let X : K —> (C*
be a homomorphism. Let (CX be the corresponding 1-dimensional
representation of K. Let

6x = If; 2gEK
x(9)‘19 E <C[K]
be the idempotent corresponding to X. Show that the G-representa­
tion Ind,§(CX is naturally isomorphic to (C[G]eX (with G acting by left
multiplication).

5.9. The Frobenius formula for the character of

an induced representation
Let us now compute the character X of Indg V. In each right coset
0 E H \G, choose a representative :00.
Theorem 5.9.1. One has
x(9) = Z xv(:va9:v; 1)­
aEH\G::raga:;1EH
This formula is called the Frobenius formula.
 1_
5.10. Frobenius reciprocity 107
Remark 5.9.2. If the characteristic of the ground field k is relatively
prime to |H|, then this formula can be written as

x(9) = rm 2 xv(:v9:v 1)­

a:EG:a:ga:‘1€H

Proof. For a right H -coset o of G, let us define

V. = {f e1nd§vIf<g> = 0 vg ¢ 0}.
Then one has
Indfiv = Q3 Va,
and so
x(9) = Z xa(9).
where X, (g) is the trace of the diagonal block of p(g) corresponding
to V0.
Since g(o) = og is a right H-coset for any right H-coset o,
x,,(g) = 0 if 0 aé og.
Now assume that o = og. Then mag = hard where h = ccagccgl E
H. Consider the map oz : V, —> V defined by a( f ) = f Since
f 6 Va is uniquely determined by f (13,), or is an isomorphism. We
have

a(9f) = 9(f)(5'«‘a) = f(5'«‘a9) = f(h5'«‘a) = /1v(h)f(5'«‘a) = h0t(f),

and gf = oz‘1hoz(f). This means that x.,(g) = X1/(h). Therefore
X(9) = Z Xv(1‘a9$;1)­
o'€H\G,o'g=o

5.10. Frobenius reciprocity

A very important result about induced representations is the Frobe­
nius reciprocity theorem which connects the operations Ind and Res.
Theorem 5.10.1 (Frobenius reciprocity). Let H C G be groups, V
a representation of G and W a representation of H. Then the space
HomG(V, Indfi W) is naturally isomorphic to HomH (Resgl/,
108 5. Representations of finite groups: Further results
Proof. Let E = HomG(V,Ind,‘§,W) and E’ = HomH(Res§V, W).
Define F : E —> E’ and F’ : E’ —> E as follows: F(a)v = (av)(e) for
any oz 6 E and (F’(fi)v)(:c) = fi(:cv) for any 5 E E’.
In order to check that F and F’ are well defined and inverse to
each other, we need to check the following five statements.
Letoz€E,fi€E’,v€V,anda:,g€G.
(a) F(a) is an H-homomorphism; i.e., F(a)hv = hF(a)v.
Indeed, F(a)hv = (ahv)(e) = (hav)(e) = (av)(he) = (av)(eh) =
h - (av)(e) = hF(a)v.
(b) F’(fi)v E Ind§W; i.e., (F’(fi)v)(ha:) = h(F’(fi)v)(:c).
Indeed. (F ’(5)v)(h$) = fi(h€W) = h5(5W) = h(F’(fi)v)(5'«‘)­
(c) F’(fi) is a G-homomorphism; i.e. F’(fi)gv = g(F’(fi)v).
Indeed, (F’(5)9v)(5'~‘) = M509”) = (F’(5)”)(5v9) = (9(F’(5)v))(5'~‘)­
(d) Fo F’ = ME’.
This holds since F(F’(fi))v = (F’(fi)1))(e) = 5(1)).
(e) F’ o F = Idg; i.e., (F’(F(a))v)(:c) = (ozv)(:c).
and we are done. [3
Indeed, (F’(F(Oz))’U)(:C) = F(oz:cv) = (a:cv)(e) = (:cav)(e) = (ozv)(a:),

Problem 5.10.2. The purpose of this problem is to understand the

notions of restricted and induced representations as part of a more
advanced framework. This framework is the notion of tensor products
over k-algebras. In particular, this understanding will lead us to a
new proof of the Frobenius reciprocity and to some analogies between
induction and restriction.
Throughout this exercise, we will use the notation and results of
Problem 2.11.6.
Let G be a finite group and H C G a subgroup. We consider
is [G] as a (k [H] ,k [G])-bimodule (both module structures are given
by multiplication inside is We denote this bimodule by k [G]1.
On the other hand, we can also consider is [G] as a (k [G] ,k [H])­
bimodule (again, both module structures are given by multiplication).
We denote this bimodule by k [G]2.
5. 10. Frobenius reciprocity 109
(a) Let V be a representation of G. Then, V is a left is [G]­
module. Thus, the tensor product is: [G], ®k[G] V is a left is [H ]—module.
Prove that this tensor product is isomorphic to Res§V as a left is [H ]­
module. The isomorphism

Res§V —> /6 [all ®k[G] V

is given by v »—> 1 ®k[G] v for every 12 E Resfi V.

(b) Let W be a representation of H. Then W is a left is [H ]­
module. According to Remark 5.8.2, Ind?,W E’ HomH (k [G] , In
other words, we have IndgW E’ Homk[H] (k [G]1 ,W). Now use part
(b) of Problem 2.11.6 to conclude Theorem 5.10.1.
(c) Let V be a representation of G . Then, V is a left is [G]—module.
Prove that not only k[G]1 ®k[G] V but also Homk[G] (k [G]2,V) is
isomorphic to ResgV as a left is: [H ]-module. The isomorphism

Homk[G] (k [G]2 , V) —> Resfiv

is given by f I-—> f (1) for every f E Homk[G] (k [G]2 , V).

(d) Let W be a representation of H. Then, W is a left is [H ]­
module. Show that Ind§W is isomorphic to k [G]2 ®k[H] W. The
isomorphism Homk[H] (k [G]1 ,W) —> k [G]2 ®k[H] W is given by f I-—>
ZgEPg‘1 ®k[H] f (g) for every f E Homkm] (k[G]1,W), where P is
a set of distinct representatives for the right H-cosets in G. (This
isomorphism is independent of the choice of representatives.)
(e) Let V be a representation of G and let W be a representation
of H. Use (b) to prove that Homg (lndg W, V) is naturally isomorphic
to Hom H (W, Resfi V).
(f) Let V be a representation of H. Prove that Ind,G,, (V*) §
(lndg V) as representations of G. [Hintz Write Ind§V as k [G]2®k[H]
V and write Indfi (V*) as Homk[H] (k [G]1 ,V*). Prove that the map
Homk[H] (k [G]1 ,V*) x (Ind,G, (V*)) —> is given by (f, (cc ®k[H] 12)) »—>
(f (v) is a nondegenerate G-invariant bilinear form, where S :
k [G] —> k [G] is the linear map defined by Sg = g‘1 for every g E G.]
110 5. Representations of finite groups: Further results
5. 1 1. Examples
Here are some examples of induced representations (we use the nota­
tion for representations from the character tables).
(1) Let G = 33, H = Z2. Using the Frobenius reciprocity, we
obtain Ind§C+ = C2 613 C+ and Ind,G,,C_ = C2 EB C_.
(2) Let G = 33, H = Z3. Then we obtain Ind§C+ = C+ EBC_,
Ind§C€ = Ind§C€2 = C2.
(3) Let G = 54,11 = 33. Then 1nd,‘;‘,<c+ = c+eac=:,1nd,‘;‘,c_ =
c- eaci, 1nd,‘;‘,<c2 = c2 ea c3_ ea cg.
Problem 5.11.1. Compute the decomposition into irreducibles of all
the representations of A5 induced from the irreducible representations
of

5.12. Representations of 3,,

In this subsection we give a description of the representations of the
symmetric group Sn for any n.
Definition 5.12.1. A partition A of n is a representation of n in
the form n = A1 + A2 + - - - + Ap, where A, are positive integers and
M Z )\i+1­

To such A we will attach a Young diagram Y)‘, which is the

union of rectangles -2’ 3 y g -2’ + 1, 0 § cc -3 A, in the coordinate
plane, for 2' = 1,. . . , p. Clearly, Y,\ is a collection of n unit squares. A
Young tableau corresponding to Y,\ is the result of filling the num­
bers 1, . . . ,n into the squares of Y,\ in some way (without repetitions).
For example, we will consider the Young tableau T,\ obtained by filling
in the numbers in increasing order, left to right, top to bottom.
We can define two subgroups of Sn corresponding to T,\:
5.12. Representations of Sn 111
1. The row subgroup P): the subgroup which maps every element
of {1, . . . ,n} into an element standing in the same row in T).
2. The column subgroup Q): the subgroup which maps every
element of {1, . . . , n} into an element standing in the same column in
T).
Clearly, P) H Q) =
Define the Young projectors
1

aA-:1-P5"\_|Zgv
9€P,\
1

5). 3= m 2QEQA
(—1)ggv
where (—-1)9 denotes the sign of the permutation g. Set c) = a )b).
Since P) H Q) = {1}, this element is nonzero.
The irreducible representations of Sn are described by the follow­
ing theorem.
Theorem 5.12.2. The subspace V) := C[Sn]c) of C[Sn] is an ir­
reducible representation of Sn under left multiplication. Every irre­
ducible representation of Sn is isomorphic to V) for a unique A.

The modules V) are called the Specht modules.

The proof of this theorem is given in the next subsection.
Example 5.12.3. For the partition A = (n), P) = Sn, Q) = {1}, so
c) is the symmetrizer, and hence V) is the trivial representation.
For the partition A = (1,. . .,1), QA = Sn, P)‘ = {1}, so c) is the
antisymmetrizer, and hence V) is the sign representation.
n = 3. For A = (2,1), V) = C2.
n = 4. For A = (2,2), V) = C2; for A = (3,1), V) = CE; for
A = (2,1,1), V) = C1.
Corollary 5.12.4. All irreducible representations of Sn can be given
by matrices with rational entries.
Problem 5.12.5. Find the sum of dimensions of all irreducible rep­
resentations of the symmetric group Sn.
112 5. Representations of finite groups: Further results
Hint: Show that all irreducible representations of Sn are real,
i.e., admit a nondegenerate invariant symmetric form. Then use the
Frobenius-Schur theorem.

5.13. Proof of the classification theorem for

representations of Sn
Lemma 5.13.1. Leta: E C[S,,]. Then a,\a:b,\ = €,\(:c)c,\, where 6;‘ is
a linear function.

Proof. If g 6 P,\Q,\, then g has a unique representation as pq,

p E P,\,q E Q,\, so a,\gb,\ = (—1)‘7c,\. Thus, to prove the required
statement, we need to show that if g is a permutation which is not in
P,\Q,\, then a,\gb,\ = 0.
To show this, it is sufficient to find a transposition t such that
t E P)‘ and g‘1tg E Q,\; then
a,\9bA = axtgbx = a,\9(9_1t9)b,\ = —a,\9b,\,
so a,\gb,\ = O. In other words, we have to find two elements i, j
standing in the same row in the tableau T = T ,\ and in the same
column in the tableau T’ = gT (where gT is the tableau of the same
shape as T obtained by permuting the entries of T by the permutation
g). Thus, it suffices to show that if such a pair does not exist, then
g E P,\Q,\, i.e., there exists p E P,\, q’ E Q; := gQ,\g‘1 such that
pT = q’T’ (S0 that 9 = pq‘1.q = 9‘1q’9 6 Ch)­
Any two elements in the first row of T must be in different
columns of T’, so there exists qj E Q; which moves all these ele­
ments to the first row. So there is pl 6 P,\ such that p1T and q’1T’
have the same first row. Now do the same procedure with the second
row, finding elements pg, qg such that p2p1T and q§q’1 T’ have the same

p, q’. The lemma is proved. B

first two rows. Continuing so, we will construct the desired elements

Let us introduce the lexicographic ordering on partitions:

A > ,u if the first nonvanishing A, —— ,u,- is positive.

Lemma 5.13.2. If)\ > ,u, then a,\C[Sn]b,, = 0.

5.13. Proof of the classification theorem for Sn 113
Proof. Similarly to the previous lemma, it suffices to show that for
any g 6 Sn there exists a transposition t E P,\ such that g‘1tg E Qn.
Let T = T,\ and T’ = gTn. We claim that there are two integers
which are in the same row of T and the same column of T’. Indeed, if
A1 > ,u1, this is clear by the pigeonhole principle (already for the first
row). Otherwise, if A1 = ,u1, as in the proof of the previous lemma, we
can find elements pl 6 P,\, q’1 E gQng‘1 such that p1T and q’1T’ have
the same first row and repeat the argument for the second row, and so
on. Eventually, having done i—-1 such steps, we’ll have A, > ,u,-, which
means that some two elements of the ith row of the first tableau are
in the same column of the second tableau, completing the proof. Cl
Lemma 5.13.3. c,\ is proportional to an idempotent. Namely, Ci =
d+!v:C«\­

statement. D
Proof. Lemma 5.13.1 implies that oi is proportional to c,\. Also, it
is easy to see that the trace of c,\ in the regular representation is n!
(as the coefficient of the identity element in c,\ is 1). This implies the

Lemma 5.13.4. Let A be an algebra and let e be an idempotent in

A. Then for any left A-module M, one has HomA(Ae,M) E eM
(namely, a: E eM corresponds to f,, : Ae —> M given by f,,(a) = am,
a E Ae).

Proof. Note that 1 — e is also an idempotent in A. Thus the state­

ment immediately follows from the fact that HomA (A, M) E M and
the decomposition A = Ae EB A(1 —— e). D
Now we are ready to prove Theorem 5.12.2. Let A 2 ,u. Then by
Lemmas 5.13.3 and 5.13.4
Hom5n(V,\, Vn) = Homsn (C[Sn]c,\, C[Sn]cn) = c,\C[Sn]cn.
The latter space is zero for A > ,u by Lemma 5.13.2 and 1-dimensional
if A = ,u by Lemmas 5.13.1 and 5.13.3. Therefore, V)‘ are irreducible,
and VA is not isomorphic to Vn if A 7é ,u. Since the number of par­
titions equals the number of conjugacy classes in Sn, the represen­
tations V,\ exhaust all the irreducible representations of Sn. The
theorem is proved.
114 5. Representations of finite groups: Further results
5.14. Induced representations for Sn
Denote by U ,\ the representation Ind‘;:C. It is easy to see that U ,\
can be alternatively defined as U ,\ = C[Sn]a,\.
Proposition 5.14.1. We have Hom(U,\,Vn) = O for ,u < A and
dimHom(U,\,V,\) = 1. Thus, U,\ = G3#Z’\Kn,\Vn, where KM are
nonnegative integers and K A A = 1.
Definition 5.14.2. The integers K n A are called the Kostka num­
bers.

Proof. By Lemmas 5.13.3 and 5.13.4,

Hom(U,\, Vn) = Hom(C[Sn]a,\, C[Sn]anbn) = a,\C[Sn]anbn,
and the result follows from Lemmas 5.13.1 and 5.132. D
Now let us compute the character of U ,\. Let Ci be the conjugacy
class in Sn having il cycles of length l for all l 2 1 (here i is a shorthand
notation for (i1, . . . , il, . . . Also let 3:1, . . . ,a:N be variables, and let
Hm = Z i
be the power sum polynomials.
Theorem 5.14.3. Let N 2 p (where p is the number of parts of A).
Then XUA (Ci) is the coeflicientl of 2:" := in the polynomial
H Hn(a;)im.
mzl
Proof. The proof is obtained easily from the Frobenius formula.
Namely, XUA (Ci) is the number of elements as 6 Sn such that a:ga:‘1 E
P,\ (for a representative g 6 Ci), divided by |P,\|. The order of P,\ is
Hz. A,-!, and the number of elements as such that :cga:‘1 E P,\ is the
number of elements in P,\ conjugate to g (i.e., |C’i fl P,\|) times the
order of the centralizer Z9 of g (which is n! / |C’i|). Thus,

XU,\(Ci) =|Zg|
|Ci 0 PAI­
H3‘ )‘.7'!

llf j > p, we define /\j to be zero.

5.15. The Frobenius character formula 115
Now, it is easy to see that the centralizer Z9 of g is isomorphic to

H S,-m x (Z/mz)*'m,
m
so

IZQI : Hmimim!a
m
and we get
Hm mimim!
XU,\(Ci) = ’i 0 PAI­
Now, since P,\ = .S,\.,
J .7 we have
A-!
1‘ 3'21 o
|Ci Fl PA] : Z 11 Hm>1 ,n;~jm,,.jm,a

where 7‘ = (rjm) runs over all collections of nonnegative integers such

that
m .1
ZTTI/l"jm=)\j, Z7‘jm=’im.
Indeed, an element of Ci that is in P,\ would define an ordered parti­
tion of each A_,- into parts (namely, cycle lengths), with m occurring
rjm times, such that the total (over all j) number of times each part
m occurs is im. Thus we get
'!
XUA(Ci) =
But this is exactly the coefficient of 33* in

ll(x§”+-~~+w';’v‘>im
mzl
(71,-m is the number of times we take D
5.15. The Frobenius character formula
Let A(:::) = ]—[19<jSN(a:,;—a:j). Let p = (N—1,N—2,...,0) 6 (IN.
The following theorem, due to Frobenius, gives a character formula
for the Specht modules V,\.
116 5. Representations of finite groups: Further results
Theorem 5.15.1. Let N 2 p. Then X1/,(C’i) is the coeflicient of
a:’\+P := 1—[a:;‘j+N_‘7 in the polynomial

A(:::) H H,,,(a;)im.
mzl
Remark 5.15.2. Here is an equivalent formulation of Theorem 5.15.1:
xv, (Ci) is the coefficient of 33* in the (Laurent) polynomial

i<j
H( Hmzl
Hm<w>e~
Proof. For brevity denote xv, by X)‘. Let us denote the class func­
tion defined in the theorem by 0;. We claim that this function has
the property 6; = ZHZA L,,,\x,,, where L,” are integers and L,\,\ = 1.
Indeed, from Theorem 5.14.3 we have
0)‘ : E (—1)0XU>~+p—U(p)’
UESN

where if the vector A+p—a(p) has a negative entry, the corresponding

term is dropped, and if it has nonnegative entries which fail to be
nonincreasing, then the entries should be reordered in nonincreasing
order, making a partition that we’ll denote by (A + p — a(p)) (i.e., we
agree that U,\+p_,,(p) := U<,\+p_,,(p))). Now note that ,u = (A + p —
a(p)) is obtained from A by adding vectors of the form e,- — e_,-, 2' < j ,
which implies that ,u > A or ,u = A, and the case ,u = A arises only if
0 = 1, as desired.
Therefore, to show that 0; = x,\, by Lemma 5.7.2, it suffices to
ShOW that (0,\,0,\)=1.
We have 1
(49,\,9,\) = — |Ci|0A(Ci)2­
Using that
o. = n! i

| I Hm mzmiml
v

we conclude that (0,\, 6;) is the coefficient of a:’\+'°y’\+P in the series

R03’?/) = A(93)A(3/)5(93,Z/), Where

mim im! im!

303, y) : H (2, as;-")w(Zk yinlzm : H (23-,k x",7”Z/IT/m)zm.
5.15. The Frobenius character formula 117
Summing over i and m, we get

m j,k
S(w, 2/) = 1—Iexp(Zw§-”y;’;”/m)

= €XP(* Z10g(1* 33j3/k)) = H(1— 333'?/k)_1°

j,k j,k
Thus,
Rm, y) : l_I.<_,—(a:.- — w;;)(y.- — Z/j)
Hi,j(1_ xi?/.7’)
Now we need the following lemma.
Lemma 5.15.3.

)— yj) : det i ya‘) °

Proof. Multiply both sides by — yj). Then the right—hand
side must Vanish on the hyperplanes z,- = z_,- and y,~ :2 yj (i.e., must
be divisible by A(z)A(y)) and is a homogeneous polynomial of degree
N (N — 1). This implies that the right-hand side and the left-hand
side are proportional. The proportionality coefficient (which is equal
then setting zN = 3/N. C]
to 1) is found by induction by multiplying both sides by zN — my and

Z 2 1N .
Now setting 21,- = 1/33, in the lemma, we get
Corollary 5.15.4 (Cauchy identity).

UESN Hj=1(1_ $.71/‘7(.7'))

monomial..1
Corollary 5.15.4 easily implies that the coefficient of a:’\+'°y’\+'° is
1. Indeed, if 0 yé 1 is a permutation in SN, the coefficient of this
in T­ is
1_[j(1‘$jya(j))
obviously zero. Cl
Remark 5.15.5. For partitions A and ,u of n, let us say that A j ,u
or ,u : A if ,u — A is a sum of Vectors of the form 42,- — e_,-, 2' < j (called
positive roots). This is a partial order, and ,u 1- A implies ,u 2 A. It
follows from Theorem 5.15.1 and its proof that
XA = Z I?,.,\xu,.,
u>_'«\
118 5. Representations of finite groups: Further results
where ,\,,) is the matrix inverse to the matrix of Kostka numbers
(K ,\,,). This implies that the Kostka numbers KM vanish unless
,u : A.

5. 16. Problems
In the following problems, we do not make a distinction between
Young diagrams and partitions.

Problem 5.16.1. For a Young diagram ,u, let A(u) be the set of
Young diagrams obtained by adding a square to ,u, and let R(/1) be
the set of Young diagrams obtained by removing a square from ,u.
(3.) Show that Res§:_1V,, = e9,\€R(,,, V,\.

(b) Show that Ind§:_1 V,, = e[9,\€A(,,, VA.

Problem 5.16.2. The content c(A) of a Young diagram A is the
sum 2]. — j). Let C =2 Zi<j(ij) E (C[S.,,] be the sum of
all transpositions. Show that C acts on the Specht module V,\ by
multiplication by c(A).

Problem 5.16.3. (a) Let V be any finite dimensional representation

of Sn. Show that the element E := (12) + - - - + (1n) is diagonalizable
and has integer eigenvalues on V which are between 1 — n and n — 1.
Hint: Represent E as Cn — C.,,_1, where Cn = C is the element
from Problem 5.16.2.
(b) Show that the element (12) +- - -+(1n) acts on V,\ by a scalar
if and only if A is a rectangular Young diagram, and compute this
scalar.

5.17. The hook length formula

Let us use the Frobenius character formula to compute the dimension
of VA. According to the character formula, dim V,\ is the coefficient
of a:’\+'° in A(a:)(a:1 + - - - + am)”. Let l_,- = A_,- + N — j. Then, using
the determinant formula for A(a:) and expanding the determinant as
 ZJL
!.
5.18. Schur-Weyl duality for g[(V) 119
, _ _ S n!
a sum over permutations, we get

‘hm V" " ......§N-.(.>( 1) M — N + s<a'>>!

z1!...zN!SESN J,
= —2fidet(l_,-(l_,-
1-1;; .2­
—1)...(z_,,- — N+z+1)).
Using column reduction and the Vandermonde determinant formula,
we see from this expression that
n!
(5.17.1) dim V,\ = det(z_;j"—*')=l "f!_'
. H (1,--1,,-)
Hj la"! H3 -7 1gz'<jgN
(where N 2 p).
In this formula, there are many cancellations. After making some
of these cancellations, we obtain the hook length formula. Namely,
for a square (z', j) in a Young diagram A (z', j 2 1,2’ g )\_,-), define the
hook of (z',j) to be the set of all squares (z",j’) in A with 2" Z 2', j’ = j
or 2" = 2', j’ 2 j. Let h(z',j) be the length of the hook of z',j, i.e., the
number of squares in it.

,d1mn!
Theorem 5.17.1 (The hook length formula). One has

V,\ —
ll! = k.
Proof. The formula follows from formula (5.17.1). Namely, note that

l—l1<jsN(l-1 " l-7') 1gkgz1,k;ez1 -1,­

It is easy to see that the factors in this product are exactly the
proceed by induction. Cl
hook lengths h(z',1). Now delete the first row of the diagram and

5.18. Schur-Weyl duality for g[(V)

We start with a simple result which is called the Double Centralizer
Theorem.
120 5. Representations of finite groups: Further results
Theorem 5.18.1. Let A, B be two subalgebras of the algebra EndE
of endomorphisms of a finite dimensional vector space E, such that
A is semisimple and B = End A E . Then:
A = EndB E (i.e., the centralizer of the centralizer of A is
A
B is semisimple.
( As a representation of A 69 B, E decomposes as
E = 69 V. e W.-,
iEI
where V, are all the irreducible representations of A and W, are all
the irreducible representations of B. In particular, we have a natural
bijection between irreducible representations of A and B.

Proof. Since A is semisimple, we have a natural decomposition E =

®,,EI V,-®W,-, where W, := HomA(V,-, E) and A = End V,-. There­
fore, by Schur’s lemma, B :2 End A(E) is naturally identified with
Q31. End(W,-). This implies all the statements of the theorem. CI

We will now apply Theorem 5.18.1 to the following situation:

E = V®”, where V is a finite dimensional vector space over a field
of characteristic zero and A is the image of <C[Sn] in End E. Let us
now characterize the algebra B. Let g[(V) be End V regarded as a
Lie algebra with operation ab — ba.
Theorem 5.18.2. The algebra B = EndAE is the image of the
universal enveloping algebra Z1 (g[(V)) under its natural action on E .
In other words, B is generated by elements of the form
An(b) :-_-b®1®---®1+1®b®---®1+---+1®1®---®b,
b E g[(V).

Proof. Clearly, the image of L{(g[(V)) is contained in B, so we just

need to show that any element of B is contained in the image of
L{(g[(V)). By definition, B = S” End V, so the result follows from
part (ii) of the following lemma.
Lemma 5.18.3. Let k be a field of characteristic zero.
5.18. Schur-Weyl duality for g[(V) 121
For any finite dimensional vector space U over is, the space
S"U is spanned by elements of the form u 69 - - - ® u, u E U.
For any algebra A over 19, the algebra S"A is generated by
elements An(a), a E A.

Proof. The space S"U is an irreducible representation of GL(U)

(Problem 4.12.3). The subspace spanned by u (3 ° ° ° (8) u is a nonzero
subrepresentation, so it must be everything.
(ii) By the fundamental theorem on symmetric functions, there
exists a polynomial P with rational coefficients such that
P(H1(a:), . . .,Hn(a:)) = :r1...a:n
(where 3: = (331, . . . ,:I:,,)). Then

P(An(a), An(a2), . . . , An(a”)) = a ® - - - ® a.

The rest follows from Cl
This completes the proof of the theorem. Cl
Now, the algebra A is semisimple by Maschke’s theorem, so the
double centralizer theorem applies, and we get the following result,
which goes under the name “Schur—Weyl duality” (as it was discovered
by Schur and popularized by Weyl in his books The theory of groups
and quantum mechanics and Classical groups; see Section 5.20).
Theorem 5.18.4. The image A of (C[Sn] and the image B of
L{(g[(V)) in End(V®”) are centralizers of each other.
Both A and B are semisimple. In particular, V®" is a
semisimple g[(V)—module.
( We have a decomposition of (A (29 B)-modules

V® " = {B V,\ <3 L A,

A

where the summation is taken over partitions of n, V)‘ are Specht

modules for Sn, and L A are some distinct irreducible representations
of g[(V) or zero.
122 5. Representations of finite groups: Further results
5.19. Schur-Weyl duality for GL(V)
The Schur-Weyl duality for the Lie algebra g[(V) implies a similar
statement for the group GL(V).
Proposition 5.19.1. The image of GL(V) in End(V®") spans B.

Proof. Denote the span of g®", g E GL(V), by B’. Let b E EndV

be any element.
We claim that B’ contains b®”. Indeed, for all values of t but
finitely many, t - Id +b is invertible, so (t - Id +b)®" belongs to B’.
This implies that this is true for all t, in particular for t = 0, since
(t - Id +b)®" is a polynomial in t.
The rest follows from Lemma 5.18.3. Cl
Corollary 5.19.2. As a representation of Sn x GL(V), V®” de­
composes as ®A V)‘ (3 L A, where L A = Homgn (VA, V®") are distinct
irreducible representations of GL(V) or zero.

Exarnple 5.19.3. If A = (n), then L)‘ = S"V, and if A = (1") (n

copies of 1), then L)‘ = /\"V. It was shown in Problem 4.12.3 that
these representations are indeed irreducible (except that /\"V is zero
if n > dim V).

5.20. Historical interlude: Hermann Weyl at the

intersection of limitation and freedom
Hermann Weyl (1885-1955) received his doctorate at the University
of Gottingen under the guidance of David Hilbert, whom he later
paid a rather dubious compliment by calling him “the Pied Piper
. . .seducing so many rats to follow him into the deep river of mathe­
matics”. Hilbert called mathematical physics “a vital nerve” of math­
ematics, and Weyl inherited the interest in cross—fertilization of math­
ematics and physics from his teacher (quoted in [53, pp. 357, 358]).
The willingness of Gottingen mathematicians to get their formulas
dirty by engaging physical problems set them apart from the obses­
sive purism of the Berlin mathematical school.
5.20. Historical interlude: Hermann Weyl 123
In 1913 Weyl was offered a professorship at the ETH in Zurich,
despite a somewhat lukewarm endorsement from Frobenius, who re­
garded the work of all candidates who came from the Géttingen school
as “very general, very deep, so abysmally deep that a shortsighted per­
son like me finds it difficult to recognize new ideas” (quoted in [22,
p. 421]). Weyl accepted and found himself in the company of Albert
Einstein, who was teaching at the ETH at the time. Yet it took a
world war to make Weyl pay attention to what Einstein was doing.
The German government initially judged Weyl unfit to fight in World
War I, but as losses were mounting, it reevaluated the concept of phys­
ical fitness and dragged him onto the battlefield. Weyl’s encounter
with the German army brought little satisfaction to either side. In
1916, at the request of the Swiss government, Weyl was discharged
and allowed to resume his work at the ETH. “My mathematical mind
was as blank as any veteran’s, and I did not know what to do”, he
later recalled. “I began to study algebraic surfaces; but before I had
gotten far, Einstein’s memoir came into my hands and set me afire”
(quoted in [52, p. 62]). The inflaming memoir was Einstein’s account
of general relativity.
In 1917 Weyl gave a lecture course on general relativity at the
ETH and soon turned it into the widely read book Space—Tz'me—
Matter, which went through four different editions within five years
and was admired by Einstein himself as a “symphonic masterpiece”
(quoted in [46, p. 65]). Weyl, however, could hardly resist the im­
pulse to improve on Einstein’s belabored mathematics. He was con­
vinced that Riemannian geometry, on which Einstein based his theory,
was not a consistently infinitesimal geometry and set out to create a
“purely infinitesimal geometry”. When he explained to a student that
in Riemannian geometry the direction of the transported vector de­
pended on the path, the student innocently asked whether the vector
length changed as well. “Of course I gave him the orthodox answer
at that moment, but in my bosom gnawed the doubt”, Weyl recalled
(quoted in [52, p. 154]). The orthodox answer was no, but he was
tempted to see what would happen if the length indeed depended on
the path. Weyl questioned the assumption of a fixed distance scale,
or “gauge”, implicit in Riemannian geometry, and arrived at a more
general geometry, in which the gauge factor varied from point to point
124 5. Representations of finite groups: Further results
in space-time, just as railway gauge varied from country to country
in the early 20th century [42, p. 3].
Weyl aspired at a unified field theory, bringing together the grav­
itational and the electromagnetic fields. Einstein was delighted to
count Weyl among the supporters of general relativity but was less
than enthusiastic to see him compete in the construction of physical
theories. Einstein praised Weyl’s theory as a “stroke of genius of the
highest order” and concluded that “except for the agreement with
reality it is in any case a grandiose achievement of thought” (quoted
in [52, pp. 163—164]). As a true mathematician, Weyl could hardly
share Einstein’s obsession with “reality”. As he later argued, “[I]t
becomes evident that now the words ‘in reality’ must be put between
quotation marks; we have a symbolic construction, but nothing which
we could seriously pretend to be the true real world” (quoted in [42,
p. 15]). In his reply to Einstein, Weyl pointedly wrote: “It must be
emphasized that the geometry that has been developed here is, from
the mathematical viewpoint, the true local geometry. It would be
strange if, instead of this true (geometry), a partial and inconsistent
local geometry with the electromagnetic field glued to it were real­
ized in Nature”. Weyl easily matched Einstein in the degree of his
sarcasm: “If you are right with regard to the real world, then I regret
having to point out a mathematical inconsistency to the dear Lord”
(quoted in [22, p. 434]).
Recasting the tensor language of Einstein’s general relativity in
the mold of the Gottingen school, Weyl focused on the development of
tensor algebra, particularly on tensors with specific symmetry prop­
erties relevant to physical applications. He aspired to obtain a math­
ematical overview of possible symmetry types and found an appro­
priate vantage point in Frobenius’s theory of finite group represen­
tations. Gradually Weyl turned further away from relativity and
towards purely mathematical questions of group theory, trying to de­
velop a group—theoretical foundation of the tensor calculus. Drawing
on the work of Elie Cartan and Issai Schur, Weyl delved into the rep­
resentation theory of finite groups, which he called “one of the most
wonderful theories to be found in mathematics”. Weyl exchanged
5.20. Historical interlude: Hermann Weyl 125
letters with Schur, who wrote that “it would be of considerable in­
terest to me to see how my latest results on the number of variables
and characters could be derived on the basis of Cartan’s methods”
(quoted in [22, pp. 456, 473]). Weyl obliged and quickly arranged
the marriage of Elie Cartan’s infinitesimal methods with Issai Schur’s
integral procedure. He did not limit himself, however, to rederiving
Schur’s results but went somewhat further and developed an entire
theory of the representations of semisimple Lie groups, including ex­
plicit formulas for the irreducible characters and for the degrees of
the corresponding representations.
The rise of quantum mechanics provided an occasion for Weyl to
mount an attack on the hegemonic status of the mathematical con­
tinuum. He believed that different branches of mathematics might
have different concepts of the number and supported the view that
“each object which is offered to mathematical analysis carries its own
kind of numbers to be defined in terms of that object and its intrinsic
constituents, instead of approaching every object by the same uni­
versal number system developed a priori and independently of the
applications” (quoted in [52, p. 240]). In his 1928 book, The Theory
of Groups and Quantum Mechanics, Weyl argued that the essence of
symmetries central to quantum mechanics was to be found not in the
continuum of real numbers but in the concepts of group theory, partic­
ularly in the “reciprocity” between the representations of symmetric
permutation groups and complete linear groups [60, p. vii].
The physics community welcomed the help of a mathematician
with cries of outrage and disgust. The American theoretical physicist
John Slater christened the approach of Weyl and his followers “Grup­
penpest”, or “the pest of group theory”. “The authors of the ‘Grup­
penpest’ wrote papers which were incomprehensible to those like me
who had not studied group theory”, he later confessed. Their results
appeared “negligible” to him, and he widely shared his “frustrating
experience, worthy of the name of a pest”, with other physicists, who
were, as he claimed, “as disgusted as I had been with the group­
theoretical approach to the problem” (quoted in [55, p. As late
126 5. Representations of finite groups: Further results
as 1975, he still believed that he had “slain the ‘Gruppenfest’ ”, bliss­
fully unaware of the essential and now routine applications of group
theory in elementary particle physics.
More sympathetic physicists also had difficulty grappling with
Weyl’s mathematics, but this only heightened their admiration for
him. Julian Schwinger had to admit that he had not “ever —— not
even to this day —— fully mastered” Weyl’s Theory of Groups and
Quantum Mechanics. Yet he called Weyl “one of my gods”, explain­
ing that “the ways of gods are mysterious, inscrutable, and beyond
the comprehension of ordinary mortals”. When asked if he ever met
a scholar he could not understand, Paul Dirac unhesitatingly replied,
“Weyl”. Yet he found Weyl’s — purely mathematical — argument
that electrons and antielectrons must have the same mass so disarm­
ingly convincing that he reportedly derived from this encounter his
famous maxim, “[I]t is more important to have beauty in one’s equa­
tions than to have them fit experiment” (quoted in [42, p. 11, 19,
12]). Weyl expressed a similar sentiment with respect to mathemat­
ics: “My work has always tried to unite the true with the beautiful,
and when I had to choose one or the other, I usually chose the beau­
tiful” (quoted in [45, p. 161]).
Weyl’s return to Gottingen in 1930 to take up Hilbert’s chair was
badly timed to coincide with the Nazis’ rise to power. One contem­
porary privately remarked, “Prof. Weyl is a peculiar race mixture, at
least seven parts Holstein and one part Jewish blood with the par­
ticular vasomotor irritability that one encounters relatively often in
people from Holstein and Friesland” (quoted in [49, p. 52]). Whatever
vasomotor dysfunction may have resulted from the German compo­
nents of his blood, it clearly paled in comparison with the trouble
that his Jewish ancestry could cause him in Nazi Germany. Although
Weyl’s low Jewish blood count exempted him from the Nazis’ direct
attacks, his wife and children would have become targets of anti­
Semitic measures. Having barely returned to his homeland, Weyl
had to face the prospect of immigration again. The decision was dif­
ficult. At first he rejected an invitation to join the Princeton Institute
for Advanced Study, unable to overcome “the love that binds me with
every string of my heart to the German language” (quoted in [52, p.
5.20. Historical interlude: Hermann Weyl 127
271]). But when Hitler came to power, Weyl’s head prevailed over his
heart, and his family left for America.
Upon his arrival at Princeton, Weyl plunged into a mathematical
culture quite inhospitable to the speculative philosophy so dear to
him. Weyl perceived the “danger of a too thorough specialization and
technicalization” of American mathematical research that produced
“a mode of writing which must give the reader the impression of
being shut up in a brightly illuminated cell where every detail sticks
out with the same dazzling clarity, but without relief”. Personally
Weyl preferred “the open landscape under a clear sky with its depth
of perspective, where the wealth of sharply defined nearby details
gradually fades away toward the horizon” [57, p. viii].
In 1939, struggling with the “yoke of foreign language”, imposed
upon his writing, Weyl summed up his results on group invariants and
representations in the book The Classical Groups [57, p. viii]. As Sir
Michael Atiyah has noted, this volume “is not a linear book with a
beginning, middle, and end. It is more like an elaborate painting that
has to be studied from different angles and in different lights. It is the
despair of the student and the delight of the professor” [3, p. 328].
Weyl viewed his mathematical writings as works of art, as much as
science. “My own mathematical works are always quite unsystematic,
without mode or connection”, he admitted. “Expression and shape
are almost more to me than knowledge itself. But I believe that,
leaving aside my own peculiar nature, there is in mathematics itself,
in contrast to the experimental disciplines, a character which is nearer
to that of free creative art” (quoted in [3, p. 323]). Weyl believed
that “ ‘mathematizing’ may well be a creative activity of man, like
language or music, of primary originality, whose historical decisions
defy complete objective rationalization” [58, p. 550].
In the artistry of mathematical creativity, in the beauty of formu­
las, in the eternal truths of mathematics sought Weyl a refuge from
the horrors of the world wars and from the destruction of European
culture [53, p. 365]. Yet after Hiroshima he realized that even pretty
formulas may have deadly uses. “To what extent shall and can the
theorist take responsibility for the practical consequences of his dis­
coveries?” he asked. “What a beautiful theoretical edifice is quantum
128 5. Representations of finite groups: Further results
physics — and what a terrible thing is the atomic bomb!”. Weyl dis­
agreed with G. H. Hardy’s defense of pure mathematics as a useless
and harmless pursuit of beauty. Weyl insisted that the Very meaning
and Value of mathematics were questioned “by the deadly menace
of its misuse” (quoted in [49, p. 67-68]). He saw in mathematics
the same moral choices that one faced in real life. His earlier reflec­
tions on the “metaphysical implications” of knowledge now acquired
a broader meaning. “Mathematics is not the rigid and uninspiring
schematism which the layman is so apt to see in it”, he had written.
“On the contrary, we stand in mathematics precisely at that point
of intersection of limitation and freedom which is the essence of man
himself” [59, p. 68].

5.21 . Schur polynomials

Let A = (A1, . . . , AP) be a partition of n, and let N 2 p. Let
N
DA<z> = Z <—1>3H z:z;;”"' = det<a.-.W—3>.
SESN j=1
Define the polynomials
Sz\('T)
:= D,\(IE)

(clearly D0 is just It is easy to see that these are indeed

polynomials, as D,\ is antisymmetric and therefore must be divisible
by A. The polynomials S ,\ are called the Schur polynomials.
Proposition 5.21.1.
H<za"+~~~+zn>im = Z Xz\(Ci)S«\($)­
Xp§N'

Proof. The identity follows from the Frobenius character formula

and the antisymmetry of
A(x) + - - - +
Cl

Certain special Values of Schur polynomials are of importance.

Namely, we have
5.22. The characters of L,\ 129
Proposition 5.21.2.

S,\(1,z,z2,...,zN'1) =H
1gz'<jgN rt _ Z 3
Z/\,-—i _ z,\,--3‘

Therefore,
X — X '—'
S,\(1,...,1)= H ' 9+,’ '.
1gi<jgN 7 ‘'2
Proof. The first identity is obtained from the definition using the
one by setting 2 = 1. D
Vandermonde determinant. The second identity follows from the first

5.22. The characters of LA

Proposition 5.21.1 allows us to calculate the characters of the repre­
sentations L ,\.
Namely, let dimV = N, let g E GL(V), and let x1,...,xN be
the eigenvalues of g on V. To compute the character XL, (g), let us
calculate 'I‘rV®,.(g®"s), where s 6 Sn. If s 6 Ci, we easily get that
this trace equals

H T1‘(gm)’7'1 = H Hm(x)i"‘.

On the other hand, by the Schur-Weyl duality

TFv® n (9® "S) = Z X.\(Ci)'IYLi (9)­

/\

Comparing this to Proposition 5.21.1 and using linear independence

of columns of the character table of Sn, we obtain
Theorem 5.22.1 (Weyl character formula). The representation L ,\
is zero if and only if N < p, where p is the number of parts of A. If
N 2 p, the character of L ,\ is the Schur polynomial S,\ Therefore,
the dimension of L,\ is given by the formula
, )\,'—)\j+j-‘l
d1mL,\ = H j_i .
1gz'<jgN
130 5. Representations of finite groups: Further results
This shows that irreducible representations of GL(V) which occur
in V8” for some n are labeled by Young diagrams with any number
of squares but at most N = dim V rows.
Proposition 5.22.2. The representation L,\+1~ (where 1N = (1, 1,
...,1) E ZN) is isomorphic to L,\ (8 /\NV.
Proof. Indeed, L,\ (8 /\NV C V8” (8) /\NV C V® "+N, and the only
component of V®"+N that has the same character as L,\ (8 /\N V is
L ,\+1~. This implies the statement. Cl
5.23. Algebraic representations of GL(V)
Definition 5.23.1. We say that a finite dimensional representation
Y of G'L(V) is algebraic (or rational, or polynomial) if its matrix
elements are polynomial functions of the entries of g, g‘1, g E GL(V)
(i.e., belong to k[g,-J-][1/det(g)]).

For example, V®" and hence all L ,\ are algebraic. Also define
L,\_.,._1~ := L,\ (8 (/\NV*)®"‘ (this definition makes sense by Propo­
sition 5.22.2). This is also an algebraic representation. Thus we
have attached a unique irreducible algebraic representation L,\ of
GL(V) = GLN to any sequence ()\1,...,)\N) of integers (not nec­
essarily positive) such that A1 2 2 AN. This sequence is called
the highest weight of L ,\.
Theorem 5.23.2. Every finite dimensional algebraic representa­
tion of GL(V) is completely reducible, and decomposes into summands
of the form L ,\ (which are pairwise nonisomorphic).
(The Peter- Weyl theorem for GL(V)) Let R be the alge­
bra of polynomial functions on GL(V). Then as a representation
of GL(V) >< GL(V) (with action (p(g,h)qb)(z) = qb(g"1xh), g,h,x E
GL(V), (15 E R), R decomposes as
R = {B L: <29 L/\a
A

where the summation runs over all A.

Proof. Let Y be an algebraic representation of GL(V). We have

an embedding § : Y —> Y (X) R given by (u,§(u))(g) := u(gu), u E Y*.
5.24. Problems 131
It is easy to see that 5 is a homomorphism of representations (where
the action of GL(V) on the first component of Y®R is trivial). Thus,
it suffices to prove the theorem for a subrepresentation Y C Rm. Now,
every element of R is a polynomial of g,-j times a nonpositive power of
det Thus, R is a quotient of a direct sum of representations of the
form S''’(V®V*)®(/\NV*)® 3. So we may assume that Y is contained
in a quotient of a (finite) direct sum of such representations. As V* =
/\N“1V (X) /\N V*, Y is contained in a direct sum of representations of
the form V®" ® (/\NV*)®3, and we are done.
(ii) Let Y be an algebraic representation of GL(V), and let us
regard R as a representation of GL(V) via (p(h)q5)(x) = q5(xh).
Then HomGL(V)(Y, R) is the space of polynomial functions f on
GL(V) with values in Y* which are GL(V)-equivariant (i.e., such
that f (gm) = gf This space is naturally identified with Y*. Tak­
ing into account the proof of (i), we deduce that R has the required
decomposition, which is compatible with the second action of GL(V)
(by left multiplications). This implies the statement. CI
Remark 5.23.3. Since the Lie algebra 5[(V) of traceless operators
on V is a quotient of g[(V) by scalars, the above results extend in a
straightforward manner to representations of the Lie algebra 5[(V).
Similarly, the results for GL(V) extend to the case of the group
S'L(V) of operators with determinant 1. The only difference is that
in this case the representations L,\ and L,\+1m are isomorphic, so
the irreducible representations are parametrized by integer sequences
A1 2 - - - 2 AN up to a simultaneous shift by a constant.
In particular, one can show that any finite dimensional represen­
tation of 5[(V) is completely reducible and any irreducible represen­
tation is of the form L ,\ (we will not do this here). For dim V = 2 one
then recovers the representation theory of 5[(2) studied in Problem
2.15.1.

5.24. Problems
Problem 5.24.1. (a) Show that the Sn-representation
V): 2:‘ (C[S'n]b,\a,\
is isomorphic to V,\.
132 5. Representations of finite groups: Further results
Hint: Define Sn-homomorphisms f : V,\ —> V( and g : V( —> V,\
by the formulas f = .’E0.,\ and g(y) = yb,\, and show that they are
inverse to each other up to a nonzero scalar.
(b) Let (:5 : (C[S'n] —> (C[S'n] be the automorphism sending s to
(—1)3s for any permutation 3. Show that ab maps any representation
V of Sn to V (X) (C_. Show also that q5((C[S'n]a) = (C[S'n]q5(a), for
a E (C[S'n]. Use (a) to deduce that V,\ (X) (C- = V,\-, where X“ is the
conjugate partition to A, obtained by reflecting the Young diagram
of A.

Problem 5.24.2. Let Rk,N be the algebra of polynomials on the

space of k-tuples of complex N x N matrices X1, . . . ,Xk, invariant
under simultaneous conjugation. An example of an element of Rk,N
is the function Tu, := Tr(w(X1, . . . ,Xk)), where w is any finite word
on a k—letter alphabet. Show that Rk,N is generated by the elements
Tw.
Hint: Consider invariant functions that are of degree d,- in each
X,-, and realize this space as a tensor product Sd‘ (V (8) V*). Then
embed this tensor product into (V (X) V*)® N = End(V)®", and use
the Schur-Weyl duality to get the result.

5.25. Representations of GL2(IFq)

5.25.1. Conjugacy classes in G'L2(lFq). Let lFq be a finite field of
size q of characteristic other than 2 and G = GL2 (lFq). Then

I0! = (42 -1)(q2 - <1),

since the first column of an invertible 2 x 2 matrix must be nonzero and

the second column may not be a multiple of the first one. Factoring,

|GL2(1Fq)| = q(q +1)(q -1)2­

The goal of this section is to describe the irreducible representations

of G.
5.25. Representations of GL2 (E1) 133
To begin, let us find the conjugacy classes in GL2 (IFq).
Number of ele­
Representatives ments in a conju- Number of classes
gacy class

1 :1: 0 1 (this is a central q—1 (one for ev­

Sea at (0 9") element) ery nonzero x)
q2 — 1 (elements
that commute with q—1 (one for ev­
Parabolic ("5 this one are of the ery nonzero z)
form (5%), tgé 0)

q2 + q (elements
Hyperbolic that commute with %(q — 1)(q — 2)
:30,y#w
this one are of the (x,y 76 0 and
(Dy) form (33), t,u 75 zyéy)
0)

Elliptic (§i?),z e
%q(q—1) (matri­
lFq, y E IF;‘, 8 E q2 — q (the reason ces with y and
lFq\lF3 (characteris- will be described —y are conju­
tic polynomial over below)
gate)
lFq is irreducible)
Let us explain the structure of the conjugacy class of elliptic
matrices in more detail. These are the matrices whose character­
istic polynomial is irreducible over lFq and which therefore don’t have
eigenvalues in lFq. Let A be such a matrix, and consider a quadratic
extension of lFq, namely, lFq(\/E), where 8 6 IR, \ F3. Over this field,
A will have eigenvalues
6! = 01 + \/E02
and
6 = a1 — \/Ea?)
with corresponding eigenvectors
12,11 (Av =av, Afi=a—v).
134 5. Representations of finite groups: Further results
Choose a basis
{cl = 12 +5, e2 = \/3(1) —
In this basis, the matrix A will have the form
a1 sag
(02 01 ) ’
justifying the description of representative elements of this conjugacy
class.
In the basis {vfl}, matrices that commute with A will have the
form
AO
0 X’
for all
AGE},
so the number of such matrices is q2 — 1.

5.25.2. 1-dimensional representations. First, we describe the 1­

dimensional representations of G.
Proposition 5.25.1. [G, G] = S'L2(]Fq).
Proof. Clearly,
det(xyx"1y'1) = 1,
so
[G, G] Q SL2(]Fq).
To show the converse, it suflices to show that the matrices

(5 1)» (3 .2)» (1 2’)

are commutators (as such matrices generate S'L2(]Fq)). Clearly, by
using transposition, it suflices to show that only the first two matrices
are commutators. But it is easy to see that the matrix

(3. 1)
is the commutator of the matrices

A42. 142). B H /'\ Or—t

lo l—*
\_/
5.25. Representations of GL2 (E1) 135
while the matrix
a0
O a’1

=aB O O=1.
is the commutator of the matrices

A (0 1)» (1 o)
This completes the proof. Cl
Therefore,
G/[G, G] '5 IF; via g v—> det(g).
The 1-dimensional representations of G thus have the form
0(9) = 6 (det(9)),
where € is a homomorphism
§ : F; —> C)‘;
so there are q — 1 such representations, denoted (C5.

5.25.3. Principal series representations. Let

BcG,{Q
B = *Q}
*
(the set of upper triangular matrices); then

Lam=U={Q DL
and
B/[B, B] 2 F; x F;
(the isomorphism maps an element of B to its two diagonal entries).
Let
A : B -—> C)‘

be a homomorphism defined by

,\ (3 = )\1(a))\2(c)
136 5. Representations of finite groups: Further results
for some pair of homomorphisms A1, A2 : 117;‘ —> C’‘. Define
V)(1,)(2 = Indg<C;1,
where (C )1 is the 1-dimensional representation of B in which B acts
by A. We have
I0 I
dim(V;11,;12)= 1,71 = q+ 1.
Theorem 5.25.2. (1) A1 75 A2 ¢ VX1112 is irreducible.
(2) A1 = A2 = p ¢ VX1112 = (C11 EB W11, where W11 is a q­
dimensional irreducible representation of G.

Trv;.1.x1(9) =aEG,
131
1_
(3) W11 E’ W1, if and only ifp = 1/; VX1112 E VX3113 if and only if
{A1, A2} = {,\’1, A11} (in the second case, ,\1 75 ,\1, X1 75 ,\i1 ).

Proof. From the Frobenius formula, we have

2 M090 1)­
aga‘1 EB
If
_ :3 0
g—0m3
the expression on the right evaluates to
IGI
)\(9)E,—1 = )\1(5F))\2(-'5) (<1 + 1)­
If
_ :3 1
gi0$7
the expression evaluates to

since here aga‘1 E B => a E B.

If
_ :3 0
g—Oy7
the expression evaluates to
()\1(<'«‘))\2(?J) + )\1(3/))\2(9«‘)) '1,
5.25. Representations of GL2 (IFq) 137
since here aga’1 E B implies that a E B or a is an element of B
multiplied by the transposition matrix.
If
9=(y,w)w#m
:3 53/

the expression on the right evaluates to 0 because matrices of this

type do not have eigenvalues over IR, (and thus cannot be conjugated
into B From the definition, for 2' = 1, 2 is a root of unity, so
|G|<XVA1,A2 a XVA1,A2> = (q + 1)2(q _ 1) + ((12 — 1)(q — 1)

+ 2<q2 + q) (4 ’ 1;” ' 2) + (q? + q) 2 A1<x>x2<y>AT<y'>A‘2<x>.

93#y

The last two summands come from the expansion

|a + b|2 = |a|2 + |b|2 + a5 + ab.
If
*1 = A2 = M,
the last term is equal to
(e12 + q)(q - 2)(q - 1),
and the total in this case is
(q+1)(q—1)[(q+1)+(q-1)+2q(q—2)] = (q+1)(q-1)2q(q-1) = 2|G|,
so
<XVA1,A2 v XVA1,A2> : 2'
Clearly,
c,, g 1nd§§<(:,,,,,,
since

HomG(<C,,,Indg C,,,,,) = HomB((C,,,<C,,) = (C (Theorem 5.10.1).

Therefore, Indg (CW, = <C,,EBW,,; W” is irreducible; and the character
of W“ is different for distinct values of p, proving that W“ are distinct.
If A1 aé A2, let 2: = :ry‘1. Then the last term of the summation is

scaaéy :z:;z:,£1 z-7&1

<q2+q> Z A1<z>T<.z> = <q2+q> Z :—;<z> = <q2+q><q—1>Z :—;<z>.
138 5. Representations of finite groups: Further results
Since

Z:—:z>=0,
26F?
because the sum of all roots of unity of a given order m > 1 is zero,
the last term becomes
A

—<q2 +q><q — 1>;§<1>= —<q2 +q><q — 1).

The difference between this case and the case of A1 = A2 is equal to
-(q2 + q)[(q — 2)(q - 1) + (q - 1)] = IGI.
so this is an irreducible representation by Lemma 5.7.2.
To prove the third assertion of the theorem, we look at the char­
acters on hyperbolic elements and note that the function
)\1(9«‘))\2(3/) + )\1(3/))\2(3«‘)
determines A1, A2 up to permutation. D
5.25.4. Complementary series representations. Let IFq2 D lFq
be a quadratic extension IFq(\/E),€ E ]Fq \ 1173. We regard this as a
2-dimensional vector space over lFq; then G is the group of linear
transformations of ]Fq2 over IFq. Let K C G be the cyclic group of
multiplications by elements of 117:2,

K={(”” f§’)}. |K|=q"’-1.

3/

For 1/ : K —> C’‘ a homomorphism, let

Y,, = Ind? (CV.
This representation, of course, is reducible. Let us compute its char­
acter, using the Frobenius formula. We get
:3 0
x (0 3:) — q(q - 1)!/(Ir),
x(A) = 0 for A parabolic or hyperbolic,

y :r y :2 y :r
X(:r ey)=l/(:2 5y)+I/(:2 ey)q.
5.25. Representations of GL2(lFq) 139
The last assertion holds because if we regard the matrix as an element
of IFq2, conjugation is an automorphism of IB‘q2 over lFq, but the only
nontrivial automorphism of ]Fq2 over lFq is the qth power map.
We thus have
Ind? (Cue: E‘ Ind? (CV
because they have the same character. Therefore, for 1/‘? 75 1/ we get
%q(q — 1) representations.
Next, we look at the tensor product
W1 ® Va,1,

where 1 is the trivial character and W1 is defined as in the previous

section. The character of this representation is
:3 0
x (0 9:) — q(q + 1)a(rr).
X(A) = 0 for A parabolic or elliptic,

x (3 = a<a:> + a<y>.
Thus the “virtual representation”
W1 <29 Va,1 — Va,1 —1nd§ <c,,,

where oz is the restriction of 1/ to scalars, has the character

X C; 0) = (q - 1)a(Ir),
IL‘

X (3 = -a(rr),

3/ 3 3/ 3 3/ 3
X(:r ey)=_l/(:2 5y)_Vq(:r 5y).
In all that follows, we will have V‘? 7é V.
The following two lemmas will establish that the inner product
of this character with itself is equal to 1 and that its value at 1 is
140 5. Representations of finite groups: Further results
positive. As we know from Lemma 5.7.2, these two properties imply
that it is the character of an irreducible representation of G.
Lemma 5.25.3. Let X be the character of the virtual representation
defined above. Then
Wm§=1
and
x(1) > 0.
Proof.
X0)=dq+D-%m+D-qw-1%=¢-1>0
We now compute the inner product (X, X). Since a is a root of unity,
this will be equal to
1

W-1Pdq+D Kq—D-@—1V-L+@-llllf-4)
2-1 . .
+ ‘“"" ’ ~ 2 Mo + uq<<>><u<<> + u«<<>>].
C elliptic
Because 1/ is also a root of unity, the last term of the expression
evaluates to
Z <2 + uq-1<<> + ul-4(4)).
C elliptic
Let’s evaluate the last summand.
Since F‘; is cyclic and V‘? 75 V,
Z vq"1<<> = Z v1‘q<<> =0.
CelP‘:2 (e1I«‘:2
Therefore,

C2 (vi-1<<>
elliptic CEF:
+u1-4(0) = — Z <uq—1<<> +u1-q<<>> = —2<q— 1)

since Ff; is cyclic of order q — 1. Therefore,

1

W‘) 2 <q — 1>2q<q + 1)

Uq-D~@-1V~P+@-llllf-4)
+ q—‘§—”~<2<q2-q> —2<q— 1») = 1.
5.26. Artin’s theorem 141
We have now shown that for any 1/ with 1/‘? 7E 1/ the representation
Y,, with the same character as
W1 ® V0,, — V0,, — Indfi <C,,
exists and is irreducible. These characters are distinct for distinct
pairs (a, 1/) (up to switching 1/ —> 1/4), so there are gal) such
representations, each of dimension q — 1.
We have thus found q — 1 1-dimensional representations of G,
gq—;1—) principal series representations, and gal) complementary se­
ries representations, for a total of q2 — 1 representations, i.e., the
number of conjugacy classes in G. This implies that we have in fact
found all irreducible representations of GL2 (lFq).

5.26. Artin’s theorem

Theorem 5.26.1. Let X be a conjugation-invariant system of sub­
groups of a finite group G. Then two conditions are equivalent:
Any element of G belongs to a subgroup H E X.
The character of any irreducible representation of G belongs
to the Q-span of characters of induced representations Indg V, where
H E X and V is an irreducible representation of H.
Remark 5.26.2. Statement (ii) of Theorem 5.26.1 is equivalent to
the same statement with Q-span replaced by (C-span. Indeed, consider
the matrix whose columns consist of the coefficients of the decompo­
sition of Indg V (for various H, V) with respect to the irreducible
representations of G. Then both statements are equivalent to the
condition that the rows of this matrix are linearly independent.
Proof. Proof that implies Assume that g E G does not
belong to any of the subgroups H E X. Then, since X is conjugation
invariant, it cannot be conjugated into such a subgroup. Hence by
the Frobenius formula, xIndg(V)(g) = O for all H E X and V. So
by (ii), for any irreducible representation W of G, xw (g) = 0. But
irreducible characters span the space of class functions, so any class
function vanishes on g, which is a contradiction.
Proof that ( implies (ii Let U be a virtual representation of G
over (C (i.e., a linear combination of irreducible representations with
142 5. Representations of finite groups: Further results

nonzero integer coefficients) such that (xy, xlndg V) = O for all H, V.

So by Frobenius reciprocity, (XUIH, xv) = 0. This means that XU
vanishes on H for any H E X. Hence by (i), XU is identically zero.
This implies (ii) (because of Remark 5.26.2). D
Corollary 5.26.3. Any irreducible character of a finite group is a
rational linear combination of induced characters from its cyclic sub­
groups.

5.27. Representations of semidirect products

Let G,A be groups and let qb : G —> Aut(A) be a homomorphism.
For a E A, denote q5(g)a by g(a). The semidirect product G t>< A
is defined to be the product A x G with multiplication law
(a1,91)(a2,92) = (a191(a2)a9192)­
Clearly, G and A are subgroups of G t>< A in a natural way.
We would like to study irreducible complex representations of
G [X A. For simplicity, let us do it when A is abelian.
In this case, irreducible representations of A are 1-dimensional
and form the character group AV, which carries an action of G. Let
0 be an orbit of this action, :1: E O a chosen element, and G9; the
stabilizer of 1: in G. Let U be an irreducible representation of G3.
Then we define a representation V(0,U) of G t>< A as follows.
As a representation of G, we set
V(O,a:,U) = Indg, U = {f = G + U|f(hg) = mg), /2 e 0.}.
Next, we introduce an additional action of A on this space by (a f ) ( g) =
:I:(g(a)) f Then it is easy to check that these two actions combine
into an action of G t>< A. Also, it is clear that this representation
does not really depend on the choice of :13, in the following sense. Let
13,31 6 O and g E G be such that ga: = y, and let g(U) be the rep­
resentation of Gy obtained from the representation U of G,; by the
action of g. Then l/(0,941) is (naturally) isomorphic to V(0,y,g(U)).
Thus we will denote l/(0,,,d.,U) by V(0,U) (remembering, however, that
:1: has been fixed).

Theorem 5.27.1. The representations V(0,U) are irreducible.

5.27. Representations of semidirect products 143
They are pairwise nonisomorphic.
They form a complete set of irreducible representations of
G [X A.

(iv) The character of V = V(0,U) is given by the Frobenius-type

formula
1
XV(a,9) =I h€G-h
T 2) :c<h<a>>xU<hgh-1).
ml - h-1
9 6G:
Proof. Let us decompose V = V(0,U) as an A-module. Then we
get
V = Q9 Vy,
yEO
where Vy = {'0 E V(0,U)|av = (y,a)v,a E A}. (Equivalently, Vy =
{v E V(0,U)|v(g) = 0 unless gy = So if W C V is a subrepresen­
tation, then W = Qyeo Wy, where Wy C Vy. Now, Vy is a represen­
tation of Gy, which goes to U under any isomorphism Gy ——> Gm de­
termined by g E G mapping :1: to y. Hence, Vy is irreducible over Gy,
so Wy = O or Wy = Vy for each y. Also, if hy = z, then my = W,,,
so either Wy = O for all y or Wy = Vy for all y, as desired.
(ii) The orbit O is determined by the A-module structure of V,
and the representation U is determined by the structure of V1. as a
G9,.-module.
(iii) We have

O O 0
formula. C]
U,O U,O
2 dim I/(30, = Z |O|2(dim U)2 =

Z |O|2|Ga:| = Z: |O||G/Gmllaxl = |G| Z: |0| = |G||AV| = l0 °< Al­

(iv) The proof is essentially the same as that of the Frobenius

Exercise 5.27.2. Redo Problems 4.12.1(a), 4.12.2, and 4.12.6 using

Theorem 5.27.1.

Exercise 5.27.3. Deduce parts (i)—(iii) of Theorem 5.27.1 from part

(iv).
Chapter 6

Quiver representations

6. 1. Problems
Problem 6.1.1. Field embeddings. Recall that k(y1, . . . , ym) de­
notes the field of rational functions of y1,. . . ,ym over a field k. Let
f : k[:c1, . . . ,:cn] —> k(y1, . . . ,ym) be an injective k—algebra homomor­
phism. Show that m 2 n. (Look at the growth of dimensions of the
spaces WN of polynomials of degree N in 33,- and their images under
f as N -+ oo.) Deduce that iff : k(:c1,...,:cn) -—> k(3/1,...,ym) is a
k—linear field embedding, then m 2 n.

Problem 6.1.2. Some algebraic geometry. Let k be an alge­

braically closed field, and let G = GLn(k). Let V be an algebraic
representation of G. Show that if G has finitely many orbits on V,
then dim(V) 3 n2. Namely:
(a) Let 3:1, . . . ,:cN be linear coordinates on V. Let us say that
a subset X of V is Zariski dense if any polynomial f (:31, . . . ,:cN)
which vanishes on X is zero (coefficientwise). Show that if G has
finitely many orbits on V, then G has at least one Zariski dense orbit
on V.
(b) Use (a) to construct a field embedding k(:I:1, . . . ,:cN) -+ k(g,,q).
Then use Problem 6.1.1.
(c) Generalize the result of this problem to the case when G =
GLn1(k) x >< GLnm(k).

145
146 6. Quiver representations
Problem 6.1.3. Dynkin diagrams. Let I‘ be a graph, i.e., a finite
set of points (vertices) connected with a certain number of edges (we
allow multiple edges). We assume that I‘ is connected (any vertex can
be connected to any other by a path of edges) and has no self—loops
(edges from a vertex to itself Suppose the vertices of I‘ are labeled
by integers 1,. ..,N. Then one can assign to I‘ an N X N matrix
Rp = (r,-j), where r,-,- is the number of edges connecting vertices 2' and
j. This matrix is obviously symmetric and is called the adjacency
matrix. Define the matrix Ar = 2I — Rp, where I is the identity
matrix.
Definition 6.1.4. I‘ is said to be a Dynkin diagram if the quadratic
form on RN with matrix AF is positive definite.

Dynkin diagrams appear in many areas of mathematics (singu­

larity theory, Lie algebras, representation theory, algebraic geometry,
mathematical physics, etc.). In this problem you will get a complete
classification of Dynkin diagrams. Namely, you will prove
Theorem. I‘ is a Dynkin diagram if and only if it is one of the
following graphs:
0 AN:
o:o:o-----o-—o:o
O DNZ
0:0-:0-----O-:0
O

0 E5
O——O:O:OiO
O
6.1. Problems 147
0 E7
o-—o:o:o:o:o
O

0 E8
OjOjOjOjOjOjO
O

(a) Compute the determinant of Ar where I‘ = A N, DN. (Use the

row decomposition rule, and write down a recursive equation for it.)
Deduce by Sylvester criterion that A N, DN are Dynkin diagrams.1
(b) Compute the determinants of Ar for E6, E7, E3 (use row de­
composition and reduce to (a)). Show they are Dynkin diagrams.
(c) Show that if I‘ is a Dynkin diagram, it cannot have cycles.
For this, show that det(Ap) = 0 for a graph I‘ below2:

1111 1

(Show that the sum of rows is 0.) Thus I‘ has to be a tree.

(d) Show that if I‘ is a Dynkin diagram, it cannot have vertices
with four or more incoming edges and that I‘ can have no more than
one vertex with three incoming edges. For this, show that det(Ap) = 0
for a graph I‘ below:

(e) Show that det(Ap) = 0 for all graphs I‘ below:

1The Sylvester criterion says that a symmetric bilinear form ( , ) on RN is positive
definite if and only if for any k 3 N, detls,-__,-Sk(e,-, e_,-) > 0.
2Please ignore the numerical labels; they will be relevant for Problem 6.1.6 below.
148 6. Quiver representations
1

12321 2

O
1 2f 3O4V3 92 51 3

°1 5 3 3 E '6 Z 3
(f) Deduce from (a)—(e) the classification theorem for Dynkin
diagrams.
(g) A (simply laced) affine Dynkin diagram is a connected
graph without self-loops such that the quadratic form defined by AF
is positive semidefinite. Classify affine Dynkin diagrams. (Show that
they are exactly the forbidden diagrams from (c)——(e).)

Problem 6.1.5. Let Q be a quiver with a set of vertices D. We

say that Q is of finite type if it has finitely many indecomposable
representations. Let b,~,- be the number of edges from i to j in Q
(i, j e D).
We have the following remarkable theorem, proved by P. Gabriel
in early 1970s.
Theorem. A connected quiver Q is of finite type if and only
if the corresponding unoriented graph (i.e., with directions of arrows
forgotten) is a Dynkin diagram (see Theorem 6. 5.2 below).
In this problem you will prove the “only if” direction of this
theorem (i.e., why other quivers are NOT of finite type).
6.1. Problems 149
(a) Show that if Q is of finite type, then for any rational numbers
33,- _>_ 0 which are not simultaneously zero, one has q(a:1, . . . ,a:N) > 0,
where
1
q(a:1, . . . ,mN) := — 5 Z b,-J-:13,-:13,-.
1l€D 1I,j€D
Hint: It suffices to check the result for integers: 33,- = 72,-. First
assume that n,- _>_ 0, and consider the space W of representations V of
Q such that dimV,- = n,-. Show that the group Hz. GLn,(k) acts with
finitely many orbits on W ED is, and use Problem 6.1.2 to derive the
inequality. Then deduce the result in the case when n,- are arbitrary
integers.
(b) Deduce that q is a positive definite quadratic form.
Hint: Use the fact that Q is dense in R.
(c) Show that a quiver of finite type can have no self-loops. Then,
using Problem 6.1.3, deduce the theorem.

Problem 6.1.6. Let G 79 {1} be a finite subgroup of SU(2) and let V

be the 2-dimensional representation of G coming from its embedding
into SU Let V,-, 13 E I, be all the irreducible representations of G.
Let 1",-j be the multiplicity of V, in V <8)
(a) Show that 1",-j = 7",,-.
(b) The McKay graph of G’, M (G), is the graph whose vertices
are labeled by 13 E I, and 13 is connected to j by nj edges. Show that
M(G) is connected. (Use Problem 4.12.10.)
(c) Show that M (G) is an affine Dynkin diagram (one of the
“forbidden” graphs in Problem 6.1.3). For this, show that the matrix
a,-,- = 26,-, — 133- is positive semidefinite but not definite, and use
Problem 6.1.3.
Hint: Let f = Z: 3:,-Xvi, where xv, are the characters of V,-. Show
directly that ((2 — X1/)f, f) 2 0. When is it equal to 0? Next, show
that M (G) has no self-loops by using the fact that if G is not cyclic,
then G contains the central element — Id 6 SU
(d) Which groups from Problem 4.12.8 correspond to which dia­
grams?
150 6. Quiver representations
(e) Using the McKay graph, find the dimensions of irreducible
representations of all finite G’ C S U (2) (namely, show that they are
the numbers labeling the vertices of the affine Dynkin diagrams on
our pictures). Compare with the results on subgroups of S0(3) we
obtained in Problem 4.12.8.

6.2. Indecomposable representations of the

quivers A1, A2, A3
We have seen that a central question about representations of quivers
is whether a certain connected quiver has only finitely many indecom­
posable representations. In the previous subsection it is shown that
only those quivers whose underlying undirected graph is a Dynkin
diagram may have this property. To see if they actually do have this
property, we first explicitly decompose representations of certain easy
quivers.

Remark 6.2.1. By an object of the type 1 :>0 we mean a map

from a 1-dimensional vector space to the zero space. Similarly, an
object of the type 0 :>1 is a map from the zero space into a 1­
dimensional space. The object 1 :>1 means an isomorphism from
a 1-dimensional to another 1-dimensional space. The numbers in such
diagrams always mean the dimension of the attached spaces and the
maps are the canonical maps (unless specified otherwise).
Example 6.2.2 (A1). The quiver A1 consists of a single vertex and
has no edges. Since a representation of this quiver is just a single
vector space, the only indecomposable representation is the ground
field (= a 1-dimensional space).
Example 6.2.3 (A2). The quiver A2 consists of two vertices con­
nected by a single edge:
O———>O

A representation of this quiver consists of two vector spaces V, W and

an operator A : V —> W:
0 0W,
V A
————>
6.2. Indecomposable representations of A1, A2, A3 151

A0N0
To decompose this representation, we first let V’ be a complement to
the kernel of A in V and let W’ be a complement to the image of A
in W. Then we can decompose the representation as follows:
A
V W 2o—>o
o——>o kerAQ300/—>o
V ImA 0 W’.
{B o—>o
The first summand is a multiple of the object 1 {>0 , the second
a multiple of 1 —>1 , and the third of 0 —>1 . We see that the
quiver A2 has three indecomposable representations, namely
1:>0 1—>1, and 0—>1.
7

Note that this statement is just the Gauss elimination theorem

for matrices.

Example 6.2.4 (A3). The quiver A3 consists of three vertices and

two connections between them. So we have to choose between two

>0 Or 0}.
possible orientations:

(1) We first look at the orientation

o>>o>>o.
A B
Then a representation of this quiver looks like

V W>0Y.
0 >0
0 0
As in Example 6.2.3 we first split away

ker A >0
0 >0 0 0.
This object is a multiple of 1 —>0 —>O . Next, let Y’ be a com­
0 0
plement of ImB in Y. Then we can also split away

0 0 Y’
0 >0 >0

which is a multiple of the object 0 :>O —>1 . This results in a

situation where the map A is injective and the map B is surjective
152 6. Quiver representations
(we rename the spaces to simplify notation):

VWY
oC————A:->0 ——L>>o .

Next, let X = ker(B o A) and let X’ be a complement of X in V. Let

W’ be a complement of A(X) in W such that A(X’) C W’. Then we
get

VWY oC—Aj>o ——L>>o

X A(X)
0 A >0 0 X’ W’
B >0 EB OQ———>o Y
——L>>o.

The first of these summands is a multiple of 1 ;>1 —>0 . Look­

ing at the second summand, we now have a situation where A is in­
jective, B is surjective, and furthermore ker(B o A) = 0. To simplify
notation, we redefine
V=X’, W=W’.
Next we let X = Im(B o A) and let X’ be a complement of X in Y.
Furthermore, let W’ = B‘1(X’ Then W’ is a complement of A(V)

VWY A B
oC:’i———>o B 0
B
in W. This yields the decomposition

V A(V) X 0 W’ X’
= O "“’ >0 "" >0 G3 0 0 ————>>O .

Here, the first summand is a multiple of 1 #>1 #>1 . By

splitting away the kernel of B, the second summand can be decom­
posed into multiples of 0 —>1 #>1 and 0 —>1 :>0 . So,
on the whole, this quiver has six indecomposable representations:
1:>0:>0, 0Z>0—>1, 1#>1Z>0,
1#>1;>1, 0—>1;>1, 0—>12>0.
(2) Now we look at the orientation
o—>o<—o.
6.2. Indecomposable representations of A1, A2, A3 153
Very similarly to the other orientation, we can split away objects of
the type
1Z>0<?0, 02>0<—1,
which results in a situation where both A and B are injective:

V W< BY
0( A >0 30 .
By identifying V and Y as subspaces of W, this leads to the problem of
classifying pairs of subspaces of a given space W up to isomorphism
(the pair of subspaces problem). To do so, we first choose a

V Y
complement W’ of V H Y in W and set V’ = W’ D V, Y’ = W’ H Y.
Then we can decompose the representation as follows:
0C >0 < 30
= 0C >0 < 30 0 e ~ VnY
<——— X .
V’ W’ Y’ Vr°wY Vr°wY
The second summand is a multiple of the object 1 A>1<i 1 . We
go on decomposing the first summand. Again, to simplify notation,
we let
V=V’, W=W’, Y=Y’.
We can now assume that V 0 Y = 0. Next, let W’ be a complement
of V EB Y in W. Then we get

V
0C W Y
;0< 30
=V VEBY Y 0{BW’
0Cj———>0<e30 0
0 «AL
The second of these summands is a multiple of the indecomposable
object 0 2>1<: 0 . The first summand can be further decom­

0C 70
posed as follows:

V V?) 7/ Y
VV00YY
= 0 0<j—0 {B 0:——>0<—-at-:0,
These summands are multiples of
12>1<Z0, 0j>1<—1
154 6. Quiver representations
So — as in the other orientation — we get six indecomposable repre­
sentations of A3:
1—>0<—0, 0—>0<—1, 1;>1<L1,
0—>1<—0, 1—>1<:0, 0—>1<Z1.
6.3. Indecomposable representations of the
quiver D4
As the last — slightly more complicated — example we consider the
quiver D4.
Example 6.3.1 (D4). We restrict ourselves to the orientation
O?%OéjO_
1

So a representation of this quiver looks like

O A1
> CV A3
<O,
V1 I V3 A2

V2

The first thing we can do is — as usual — split away the kernels of the
maps A1, A2, A3. More precisely, we split away the representations
O 0>O<
0 O,
kerA1 I 0
00
O >O< O, 0
O
0

0
6.3. Indecomposable representations of the quiver D4 155
O> jg 0 0.
O ker A3

OO——>OO

These representations are multiples of the indecomposable objects

0 0 0,
0 >o<
1T0
00 0 >o< 0,
0
0
0

0
1

0 o o_
0 >o<
0T1 0
0

So we get to a situation where all of the maps A1, A2, A3 are injective:

As in Example 6.2.4, we can then identify the spaces V1, V2, V3 with
subspaces -of V. So we get to the triple of subspaces problem of
classifying triples of subspaces of a given space V.
156 6. Quiver representations
The next step is to split away a multiple of

V A
on on

oo——-—>o+—:

to reach a situation where

V1 + V2 + V3 = V.

By letting Y = V1 V1 V2 fl V3, choosing a complement V’ of Y in V, and

0 0
setting V,-’ = V’ (WV,-, i = 1, 2, 3, we can decompose this representation
into

V1’ V3’
oC————> 0 «:30 Y YO ,
0>0<

V; Y
1 ea ]2
The last summand is a multiple of the indecomposable representation

+—t HO

~A

HO—————>O+—t

So — considering the first summand and renaming the spaces to sim­

plify notation — we are in a situation where

V=V1+V2-i-V3, V1flV2flV3=0.

As a next step, we let Y = V1 fl V2 and we choose a complement V’

of Y in V such that V3 C V’ and set V1’ = V’ flV1,V§ = V’fl V2. This
6.3. Indecomposable representations of the quiver D4 157
yields the decomposition

oC————>‘o/<—:3o

V1 L V3
V2

V1, V3 Y 0 o 0 I
OC———>‘: «:30 0 ~ > Y < 0 .

V; Y
The second summand is a multiple of the indecomposable object

A 0
r—t O

FNJ

F-*O——->Or—*

In the resulting situation we have V1 F1 V2 = 0. Similarly we can split

away multiples of

A O V A
r—t r—t O0 HO

and

OO——-->Or—t +aO——-—>O+—t

to reach a situation where the spaces V1, V2, V3 do not intersect pair­
wise:

V1flV2 = V10‘/3 = V2flV3 =

If V1 Q V2G3V3, we let Y = V1 fl(V2 69 V3). We let V1’ be a complement

of Y in V1. Since then V1’ V1 (V2 69 V3) = 0, we can select a complement
158 6. Quiver representations

oC—————>’o/<——3o

V1 L V3
V1’
V2

0
2 T 69 Y V3
IQ 0I
V’ of V1’ in V which contains V2 GBV3. This gives us the decomposition

Q N’ >‘:l< Q .C————J: <4. .

0 V2
The first of these summands is a multiple of
C> &
4 O.
X
1 0

0>000 0O
©O——-->Ol—*

By splitting these away, we get to a situation where V1 Q V2 69 V3.

Similarly, we can split away objects of the type
1
{fi V A 0

T2 and
r—t

O
1 $0--—>Or—*

to reach a situation in which the following conditions hold:

(1) V1+V2+V3=V.
(2) V1flV2=0, V1flV3=0, V2flV3=0.
(3) V1 Q V2G9V3, V2 Q V1G9V3, V3 Q V1G9V2­
But this implies that
V1G9V2=V1G9V3=V2G9V3=V­
So we get
dim V1 7- V2 = dim V3 = 7?,
6.3. Indecomposable representations of the quiver D4 159
and
dim V = 277,.
Since V3 C_Z V1 69 V2, we can write every element of V3 in the form
:1: 6 V3, :1: = (:1:1,:1:2), 51:1 6 V1, 51:2 6 V2.
We then can define the projections
B1 : V3 ——> V1, (:1:1,:1:2) I——> 5131,

B2 : V3 ——> V2, (:1:1,:1:2) +—> 51:2.

Since V3 fl V1 = 0 and V3 V1 V2 = 0, these maps have to be injective
and therefore are isomorphisms. We then define the isomorphism
A=B2oB1‘1:V1—+V2.
Let e1,...,en be abasis for V1. Then we get
V1 =(Ce1G9(Ce2G9---G9(Ce,,,,

V2
V2 =(CAe1G9(CAe2G9---G9(CAen,
V3=C(€1+A€1)G9C(€2+A€2)G9---G9(C(en+A6n).

O0
So we can think of V3 as the graph of an isomorphism A : V1 —-> V2.
From this we obtain the decomposition
oC————> ‘Q/«:30 oC——>(Cg +40 .

V1 L V3 é C(1,
j=1 0) L C(1, 1)

These correspond to the indecomposable object

2

O > O2 < O ,
1I1 O
1

Thus the quiver D4 with the selected orientation has 12 indecompos­

able objects. If one were to explicitly decompose representations for
the other possible orientations, one would also find 12 indecomposable
objects.
160 6. Quiver representations
It appears as if the number of indecomposable representations
does not depend on the orientation of the edges — and indeed Gabriel’s
theorem will generalize this observation.

6.4. Roots
From now on, let I‘ be a fixed graph of type An, Dn, E5, E7, E3. We
denote the adjacency matrix of P by Rp.
Definition 6.4.1 (Cartan matrix). We define the Cartan matrix
of I‘ as
Ar = 2 Id —Rp.

On the lattice Z" (or the space R") we then define an inner
product
B(rv. y) = IvTAry
corresponding to the graph I‘.
Lemma 6.4.2. (1) B is positive definite.
(2) B(:1:,:1:) takes only even values for :3 E Z".

Proof. ( 1) This follows by definition, since I‘ is a Dynkin diagram.

(2) By the definition of the Cartan matrix we get

1,] iaj. 1751

B(:1:,:c)=:1:TAp:1:=E:1:,-a,-J-zztj=2E:z:§+ Z £lI,'a,'j:lIj

i ’i<j
which is even. El
=2 E £lI§+2- E 0.,'j£lI,'£lI_7',

Definition 6.4.3. A root with respect to a certain positive inner

product is a shortest (with respect to this inner product) nonzero
vector in Z".

So for the inner product B, a root is a nonzero vector :1: E Z"

such that
B(:1:, :11) = 2.

Remark 6.4.4. There can be only finitely many roots, since all of
them have to lie in some ball.
6.4. Roots 161
Definition 6.4.5. We call vectors of the form
ith

simple roots.
The 0:; naturally form a basis of the lattice Z".
Lemma 6.4.6. Let a be a root, a = ELI k,-a,-. Then either k,- 2 0
for all 2' or k,- 3 0 for all 2'.

i .7
Proof. Assume the contrary, i.e., k,- > 0, kj < 0. Without loss of
generality, we can also assume that ks = 0 for all s between 2' and 3'.

0000000,
We can identify the indices 2', j with vertices of the graph I‘:
6Z -1

Next, let 6 be the edge connecting 2' with the next vertex towards j
and let 2" be the vertex on the other end of e. We then let P1, P2 be
the graphs obtained from I‘ by removing 5. Since I‘ is supposed to
be a Dynkin diagram — and therefore has no cycles or loops — both
P1 and P2 will be connected graphs which are not connected to each
other:

F1

F2

Then we have 2' 6 P1, 3' 6 P2. We define

B =mEF1
Z kmam,m€I‘2
7 = Z kmam.
With this choice we get
a=B+y
Since k,- > 0, kj < 0, we know that fl ;£ 0, 7 ;£ 0 and therefore
B(B,B) 2 2. B(7,7) 2 2.
162 6. Quiver representations
Furthermore,
B(3/Y) = ‘k2’/W
since P1, P2 are only connected at 5. But this has to be a nonnegative
number, since k,- > 0 and 19,-: 3 O. This yields

22 20 22
B(a.a) = B(B + 7.3 + 7) = B(B.fi) +2 B(B.7) +B(7.7) 2 4.

But this is a contradiction, since a was assumed to be a root. El

Definition 6.4.7. We call a root a = k,-ai a positive root if all
k,- 2 0. A root for which k,- 3 0 for all 2' is called a negative root.

Remark 6.4.8. Lemma 6.4.6 states that every root is either positive
or negative.
Example 6.4.9. (1) Let I‘ be of the type AN_1. Then the lattice
L = ZN"1 can be realized as a subgroup of the lattice ZN by letting
L Q ZN be the subgroup of all vectors (331, . . . ,:1:N) such that

The vectors
C21 = (1,-1,0,...,0),
a2=(0,1,—1,0,...,0),

C¥N_1 = (0, . . .,0, 1, -1)

naturally form a basis of L. Furthermore, the standard inner product
(177 y) = 2 €132‘?/2'

on ZN restricts to the inner product B given by I‘ on L, since it takes

the same values on the basis vectors:
(05127 0571) = 27

(G.
2’a_)
J __ 0,
—1, z',j are adjacent,
otherwise.
This means that vectors of the form
(0,...,0,1,0,...,0,—1,0,...,0)=CIi+C)!¢+1+-°-+O!j_1
6.5. Gabriel’s theorem 163
and

(O,...,0,—1,0,...,0,1,0,...,0) = —(0.’,; +C¥,;+1 + +C¥_7'_1)

are the roots of L. Therefore the number of positive roots in L equals
N(N— 1)
—_2‘_‘
(2) As a fact, we also state the number of positive roots in the
other Dynkin diagrams:
DN: N(N -1) roots,
E6: 36 roots,
E7: 63 roots,
E3: 120 roots.
Definition 6.4.10. Let a E Z" be a positive root. The reflection 3,,
is defined by the formula
sa('u) = v — B('v, a)a.

We denote sat. by s,- and call these simple reflections.

Remark 6.4.11. As a linear operator of IR", 3,, fixes any vector
orthogonal to a and
sa(a) = —a.
Therefore 3,, is the reflection at the hyperplane orthogonal to a and
in particular fixes B. The s,- generate a subgroup W Q 0(lR"), which
is called the Weyl group of I‘. Since for every 10 E W, w(a,-) is a
root, and since there are only finitely many roots, W has to be finite.

6.5. Gabriel’s theorem

Definition 6.5.1. Let Q be a quiver with any labeling 1, . . . ,n of
the vertices. Let V = (V1, . . . , Vn) be a representation of Q. We then
call
d(V) = (dim V1, . . . , dim Vn)
the dimension vector of this representation.

We are now able to formulate Gabriel’s theorem using roots.

164 6. Quiver representations
Theorem 6.5.2 (Gabriel’s theorem). Let Q be a quiver of type An,
Dn, E6, E7, E3. Then Q has finitely many indecomposable represen­
tations. Namely, the dimension vector of any indecomposable repre­
sentation is a positive root (with respect to Br) and for any positive
root a there is exactly one indecomposable representation with dimen­
sion vector a.

6.6. Reflection functors

Definition 6.6.1. Let Q be any quiver. We call a vertex i E Q a
sink if all edges connected to i point towards i:
——>o<-—.

We call a vertex i E Q a source if all edges connected to i point away

from i:
10g. i

l
Definition 6.6.2. Let Q be any quiver and let i E Q be a sink
(respectively, a source). Then we let be the quiver obtained from
Q by reversing all arrows pointing into (respectively, pointing out of)
i.

We are now able to define the reflection functors (also called

Coxeter functors).
Definition 6.6.3. Let Q be a quiver, and let i E Q be a sink. Let V
be a representation of Q. Then we define the reflection functor
F,-+ : RepQ —> Rep—Q—,­
by the rule
F,-+(V)k = Vk if k 79 i,

F,+(V),- =ker cp : QBV, —+ V,­

j—+i'
6.6. Reflection functors 165
Also, all maps stay the same except those now pointing out of i; these
are replaced by compositions of the inclusion of kercp into G3
j—+i'
with the projections $3. V,-_,,- —> V,-.
Definition 6.6.4. Let Q be a quiver, and let i E Q be a source. Let
V be a representation of Q. Let «b be the canonical map
it : V,- —> QB V,-.
i'—+j

Then we define the reflection functor

F2.’ : RepQ —> Repm
by the rule
F.-(V)k = Vk if/<3 75 2'.

F;(V),- = Coker (ii) = €BV,- /Imp.

i'—+j

Again, all maps stay the same except those now pointing into i; these
are replaced by the compositions of the inclusions Vk —>
i'—+jG)

with the natural map @314]. —> @143. V}/Imib.

Proposition 6.6.5. Let Q be a quiver and V be an indecomposable
representation of Q.
(1) Let i E Q be a sink. Then either dimV,- = 1, dimVj = O for
j 75 i or
so I G} Vj —> V.­
]—-)7.

is surjective.
(2) Let i E Q be a source. Then either dimV,- = 1, diml/3 = O
for j aé i or
10 : V. —> {B V,­
’i—-) j

is injective.

Proof. ( 1) Choose a complement W of Im cp. Then we get

0%}.
W

V = 0 T 0 ea V’. C
0
166 6. Quiver representations
Since V is indecomposable, one of these summands has to be zero. If
the first summand is zero, then cp has to be surjective. If the second
summand is zero, then the first one has to be of the desired form,
because else we could write it as a direct sum of several objects of the
type

O0 G0

O0—>0+—t

which is impossible since V was supposed to be indecomposable.

(2) This follows similarly by splitting away the kernel of ab. Cl
Proposition 6.6.6. Let Q be a quiver, and let V be a representation
of Q.
(1) If
so = G} V; —> V.­
3-»
is surjective, then
F; F,.+V = V.
(2) If
10 2 V. —> {B V,­
*+j
is injective, then
Fi+F2-“V = V.

Proof. In the following proof, by i —> j we will always mean that

i points into j in the original quiver Q. We only establish the first
statement and we also restrict ourselves to showing that the spaces of
V and F1-“F,-+V are the same. It is enough to do so for the ith space.
Let

s0=Q}V,-—>V'.­

be surjective and let

K = ker cp.
6.6. Reflection functors 167
When applying F2-+, the space V,- gets replaced by K. Furthermore,
let
¢;K—+€BV,-.
j—+i'

After applying Fz-", K gets replaced by

K'= EBV. /<Im¢>.

j—+i'

But
Im2,b=K
and therefore

1-H" 1-” J'—*%'

K'= / ker(<p:®Vj—>V,-) =Im(<p:®Vj—>V;)
by the homomorphism theorem. Since cp was assumed to be surjective,
we get
K’ = V,-.
E}

Proposition 6.6.7. Let Q be a quiver, and let V be an indecompos­

able representation of Q. Then F2-+V and F,-"V (whenever defined)
are either indecomposable or 0.

Proof. We prove the proposition for F1-+V; the case Fz.'V follows
similarly. By Proposition 6.6.5 it follows that either

so = EB V: n V2‘
j—>i'

is surjective or dim V,- = 1, dimV,- = 0,3’ 7E i. In the last case

Fz.+V = 0.

So we can assume that go is surjective. In this case, assume that F1-+V

is decomposable as
F,.+V=XeBY
168 6. Quiver representations
with X, Y 7E 0. But F2-+V is injective at 2, since the maps are canonical
projections, whose direct sum is the tautological embedding. There­
fore X and Y also have to be injective at 2 and hence (by Proposition
6.6.6)
F2—+F[X = X, Fi+FfY = Y.
In particular
Fz—'X 7E 0, F2.'Y 7E 0.
Therefore
V = F2-“F2.+V = Fi'X EB F2.'Y,
which is a contradiction, since V was assumed to be indecomposable.
So we can infer that

is indecomposable. Cl F.+
2
V

Proposition 6.6.8. Let Q be a qu2'ver and let V be a representation

of Q.
(1) Let 2 E Q be a s2'nk and let V be surjectrve at 2. Then
d(F2-+V) = S2(d(V))­
(2) Let 2 E Q be a source and let V be 2'nject2've at 2. Then
d<F.~*V> = s.<d<v>>.

Proof. We only prove the first statement; the second one follows
similarly. Let 2 E Q be a sink and let

90:563.‘/3-—>l/2

be surjective. Let K = ker (,0. Then

dimK = Z:dimVj — dimV,-.
j—>2'

Therefore we get

(d(F,.+V) — d(V)) = Z dim Vj — 2 dim V,- = —B (d(V), a,-)

1 j—>2'
and
j_
(d(F2-+V) — d(V)) 0, 3' 7E 2.
6. 7. Coxeter elements 169
This implies

d(F,.+V) — d(V) = -3 (d(V), a.) a.

<=» d(F,.+V) = d(V)—B(d(V),a,-)a,- = s,-(d(V)).
1:

6.7. Coxeter elements

Definition 6.7.1. Let Q be a quiver and let I‘ be the underlying
graph. Fix any labeling 1, . . . ,n of the vertices of I‘. Then the Cox­
eter element c of Q corresponding to this labeling is defined as
C=S1S2...Sn.

Lemma 6.7.2. Let

3 = Z kiai
with k,- _>_ O for all i but not all k,- = 0. Then there is N E N, such
that
CNB
has at least one strictly negative coefiicient.

Proof. The Coxeter element c belongs to a finite group W. So there

is M E N, such that
M = 1.
Weclaimthat
1+c+c2+---+cM’1=O
as operators on IR". This implies what we need, since B has at least
one strictly positive coefficient, so one of the elements
cB.c2B.---.cM’1fi
must have at least one strictly negative coefficient. Furthermore, it is
enough to show that 1 is not an eigenvalue for c, since
(1+c+c2+...+cM_1)'v=w;-E0
=> cw=c(1+c+c2+---+cM"1)'v
=(c+c2+c3+---+cM’1+1)’U=w.
170 6. Quiver representations
Assume the contrary, i.e., 1 is an eigenvalue of c and let 2) be a corre­
sponding eigenvector:
cv=v —=> s1...sn'v='v
4-—-> s2...snv = 3122.
But since s,- only changes the 2th coordinate of v, we get
3122 :22 and s2...sn'v =21.
Repeating the same procedure, we get
s,-v = 22
for all 2. But this means
B (22, a,-) = 0
for all 2', and since B is nondegenerate, we get 22 = 0. But this is a
contradiction, since 22 is an eigenvector. D
6.8. Proof of Gabriel’s theorem
Let V be an indecomposable representation of Q. We introduce a
fixed labeling 1, . . . , n on Q, such that 2 < j if one can reach j from
2. This is possible, since we can assign the highest label to any sink,
remove this sink from the quiver, assign the next highest label to a
sink of the remaining quiver, and so on. This way we create a labeling
of the desired kind.
We now consider the sequence
W’) = V, V“) = F,',*V, V”) = F,;*_,F,;*V, . . ..
This sequence is well defined because of the selected labeling: n has to
be a sink of Q, n—1 has to be a sink of E (where Q” is obtained from
Q by reversing all the arrows at the vertex 2") and so on. Furthermore,
we note that V(") is a representation of Q again, since every arrow
has been reversed twice (since we applied a reflection functor to every
vertex). This implies that we can define
V<"+1> = F,-,*V<">, . ..
and continue the sequence to infinity.
6.8. Proof of Gabriel’s theorem 171
Theorem 6.8.1. There exists m E N, such that

d (VW) = ap
for some p.

Proof. If V“) is surjective at the appropriate vertex k, then

d (V(‘+1)) = d (F,jV<">) = skd (V09) .

This implies that if V(°), . . . , Va”) are surjective at the appropriate
vertices, then

d (Vm) = . . . sn_1snd(V).
By Lemma 6.7.2 this cannot continue indefinitely, since d (Vm) may
not have any negative entries. Let i be the smallest number such that
V“) is not surjective at the appropriate vertex. By Proposition 6.6.7
it is indecomposable. So, by Proposition 6.6.5, we get

for some p. D d(V('7) = ap

We are now able to prove Gabriel’s theorem. Namely, we get the

following corollaries.

Corollary 6.8.2. Let Q be a quiver, and let V be any indecomposable

representation. Then d(V) is a positive root.

Proof. By Theorem 6.8.1

s,-1...s,-m(d(V))= ap.
Since the s,- preserve B, we get

B<d<v>,d<v>> = B<a.. a.) = 2.

D

Corollary 6.8.3. Let V, V’ be indecomposable representations of Q

such that d(V) = d(V’ Then V and V’ are isomorphic.
172 6. Quiver representations
Proof. Let i be such that
d (W?) = ap­
Then we also get d (V’(i)) = ap. So
V’('7 = V“) =; Vi.
Furthermore we have
V“) = F; . . . F,;*_1F,;*V<°>,
V’<"> = F; . . . F;f_1F,jV'<°>.
But both V(“1),.. .,V(°) and V’(i"1),. ..,V’(°) have to be surjective
at the appropriate vertices. This implies
F,;F,;_1...F,;V"

Fn‘Fn‘_1...F,:F,;*...F,;*_1F,j“V’(°) == V'<°>
{Fn‘Fn‘_1...Fk‘F,;*...F,;*_1F,;*V(°) W“) ==V,V’.
B

These two corollaries show that there are only finitely many inde­
composable representations (since there are only finitely many roots)
and that the dimension vector of each of them is a positive root. The
last statement of Gabriel’s theorem follows from
Corollary 6.8.4. For every positive root a, there is an indecompos­
able representation V with
d(V) = a.
Proof. Consider the sequence
sna, sn_1sna, . . ..
Consider the first element of this sequence which is a negative root
(this has to happen by Lemma 6.7.2) and look at one step before that,
calling this element B. So B is a positive root and s,-B is a negative
root for some i. But since the s,- only change one coordinate, we get
3 = 02'
and
(sq . ..sn_1sn)a = a,-.
6.9. Problems 173
We let (C0) be the representation having dimension vector a,-. Then
we define
V = Fn—Fn__1...Fq—<C(,-).
This is an indecomposable representation and
d(V) = a.
D

Example 6.8.5. Let us demonstrate by example how reflection func­

tors work. Consider the quiver D4 with the orientation of all arrows
towards the node (which is labeled by 4). Start with the 1-dimensional
representation VQ4 sitting at the fourth vertex. Apply to VQ4 the func­
tor F3’ F2’ F1’. This yields
F1_FEF3—Va4 = 0r1+a2+0r3+0r4'
Now applying F4", we get
F4=_Fl_F2_F3_V04 = 0r1+a2+0r3+20l4'
Note that this is exactly the inclusion of three lines into the plane,
which is the most complicated indecomposable representation of the
D4 quiver.

6.9. Problems
Problem 6.9.1. Let Qn be the cyclic quiver of length n, i.e., n
vertices connected by n oriented edges forming a cycle. Obviously,
the classification of indecomposable representations of Q1 is given by
the Jordan normal form theorem. Obtain a similar classification of
indecomposable representations of Q2. In other words, classify pairs
of linear operators A : V ——> W and B : W ——> V up to isomorphism.
Namely:
(a) Consider the following pairs (for n 2 1):
(1) En)‘: V = W = C", A is the Jordan block of size n with
eigenvalue x\, B = 1 (x\ E (C).
(2) Enm: is obtained from En,0 by exchanging V with W and A
with B.
(3) Hn: V = C" with basis 12,-, W = (C""1 with basis 11)., Avi =
112., Bw,- = vz-+1 for 2' < n, and Ann = O.
174 6. Quiver representations
(4) K n is obtained from H” by exchanging V with W and A with
B.
Show that these are indecomposable and pairwise nonisomorphic.
(b) Show that if E is a representation of Q2 such that AB is not
nilpotent, then E = E’ 69 E”, where E” = En)‘ for some /\ 75 0.
(c) Consider the case when AB is nilpotent, and consider the
operator X on V 69 W given by X(v,w) = (Bw,Av). Show that X
is nilpotent and admits a basis consisting of chains (i.e., sequences
11,, Xu, X2u, . . .X“1'u, where Xlu = O) which are compatible with the
direct sum decomposition (i.e., for every chain 21, E V or u E
Deduce that (1)—(4) are the only indecomposable representations of
Q2.
(d) (Harder!) Generalize this classification to the Kronecker quiver,
which has two vertices 1 and 2 and two edges both going from 1 to 2.
(e) (Still harder!) Can you generalize this classification to Qn,
n > 2 with any orientation? (Easier version: consider only the cyclic
orientation).

Problem 6.9.2. Let L C %Z8 be the lattice of vectors where the

coordinates are either all integers or all half-integers (but not integers)
and the sum of all coordinates is an even integer.
(3.) Let C13‘: €z'—6z'+1,’i‘—‘ 1,. . . , 6, C17 = 65-I-67, G3 = 2&1 62'.
Show that a,- are a basis of L (over Z).
(b) Show that roots in L (under the usual inner product) form a
root system of type E3 (compute the inner products of a,-).
(c) Show that the E7 and E5 lattices can be obtained as the sets
of vectors in the E3 lattice L where the first two, respectively three,
coordinates (in the basis e,-) are equal.
(d) Show that E5, E7, E3 have 72, 126, and 240 roots, respectively
(enumerate types of roots in terms of the presentations in the basis
e,-, and count the roots of each type).
Problem 6.9.3. Let Va be the indecomposable representation of a
Dynkin quiver Q which corresponds to a positive root a. For instance,
if a,- is a simple root, then Va, has a 1-dimensional space at 2' and is
O everywhere else.
6.9. Problems 175
(a) Show that if 2' is a source, then Ext1(V,Va1.) = O for any
representation V of Q, and if 2' is a sink, then Ext1(Va1., V) = O.
(b) Given an orientation of the quiver, find a Jordan-Hélder series
of V0, for that orientation.
Chapter 7

Introduction to
categories

7.1. The definition of a category

We have now seen many examples of representation theories and of
operations with representations (direct sum, tensor product, induc­
tion, restriction, reflection functors, etc.). A context in which one can
systematically talk about this is provided by category theory.
Category theory was founded by Saunders Mac Lane and Samuel
Eilenberg in 1942-1945. It is a fairly abstract theory which seem­
ingly has no content, for which reason it was christened “abstract
nonsense” (see Section 7.10). Nevertheless, it is a very flexible and
powerful language, which has become totally indispensable in many
areas of mathematics, such as algebraic geometry, topology, represen­
tation theory, and many others.
We will now give a very short introduction to category theory,
highlighting its relevance to the topics in representation theory we
have discussed. For a serious acquaintance with category theory, the
reader may use, for instance, the classical book [McL].

Definition 7.1.1. A category C is the following data:

(i) A class of objects Ob(C).

177
178 7. Introduction to categories
(ii) For every objects X,Y E Ob(C), the class Homc(X, Y) =
Hom(X, Y) of morphisms (or arrows) from X, Y (for f E Hom(X, Y ,
one may write f : X —> Y).
(iii) For any objects X, Y, Z 6 Ob(C), a composition map
Hom(Y, Z) X Hom(X, Y) —> Hom(X, Z), (f, g) »—> fog.
This data is required to satisfy the following axioms:
1. The composition is associative, i.e., (f o g) o h. = f o (g o h).
2. For each X E Ob(C), there is a morphism 1X 6 Hom(X,X),
called the unit morphism, such that 1X 0 f = f and g o 1X = g for
any f, g for which compositions make sense.

Remark 7.1.2. We will write X E C instead of X E Ob(C).

Example 7.1.3. 1. The category Sets of sets (morphisms are arbi­
trary maps).
2. The categories Groups, Rings (morphisms are homomor­
phisms).
3. The category Vectk of vector spaces over a field k (morphisms
are linear maps).
4. The category Rep(A) of representations of an algebra A (mor­
phisms are homomorphisms of representations).
5. The category of topological spaces (morphisms are continuous
maps).
6. The homotopy category of topological spaces (morphisms are
homotopy classes of continuous maps).

Important remark. Unfortunately, one cannot simplify this

definition by replacing the word “class” by the much more famil­
iar word “set”. Indeed, this would rule out the important Example
7.1.3(1), as it is well known that there is no set of all sets and that
working with such a set leads to contradictions. The precise defini­
tion of a class and the precise distinction between a class and a set is
the subject of set theory and cannot be discussed here. Luckily, for
most practical purposes (in particular, in these notes) this distinction
is not essential.
7.2. Functors 179
We also mention that in many examples, including Examples
7.1.3(1)—(6), the word “class” in part (ii) of Definition 7.1.1 can be
replaced by “set”. Categories with this property (that Hom(X, Y) is
a set for any X, Y) are called locally small; many categories that
we encounter are of this kind.

Definition 7.1.4. A full subcategory of a category C is a category

C’ whose objects are a subclass of objects of C, and Homcr(X, Y) =
Homc(X, Y).

Example 7.1.5. The category AbelianGroups is a full subcategory

of the category Groups.

Sometimes the collection Hom(X , Y) of morphisms from X to

Y in a given locally small category C is not just a set but has some
additional structure (say, the structure of an abelian group, or a vector
space over some field). In this case one says that C is enriched over
another category D (which is a monoidal category, i.e., has a product
operation and a unit object under this product, e.g., the category of
abelian groups or vector spaces with the tensor product operation).
This means that for each X,Y E C, Hom(X, Y) is an object of D,
and the composition Hom(Y, Z) X Hom(X, Y) —> Hom(X, Z) is a
morphism in D. E.g., if D is the category of vector spaces, this
means that the composition is bilinear, i.e., gives rise to a linear
map Hom(Y, Z) <8) Hom(X, Y) —> Hom(X, Z). For a more detailed
discussion of this, we refer the reader to [MCL].
Example 7.1.6. The category Rep(A) of representations of a k­
algebra A is enriched over the category of k-vector spaces.

7 .2. Functors
We would like to define arrows between categories. Such arrows are
called functors.

Definition 7.2.1. A functor F : C —> D between categories C and

D is
(i) a map F : Ob(C) —> Ob(D);
180 7. Introduction to categories
(ii) for each X,Y E C, a map F = FX,y : Hom(X,Y) —>
Hom(F(X), F which preserves compositions and identity mor­
phisms.

Note that functors can be composed in an obvious way. Also, any

category has the identity functor.
Example 7.2.2. 1. A (locally small) category C with one object X
is the same thing as a monoid. A functor between such categories is
a homomorphism of monoids.
2. Forgetful functors
Groups —-> Sets, Rings —> AbelianGroups.
3. The opposite category of a given category is the same category
with the order of arrows and compositions reversed. Then V I—> V*
is a functor Vectk —> Vectzp.
4. The Hom functors: if C is a locally small category, then we
have the functor C —> Sets given by Y I—> Hom(X, Y) and C°p —> Sets
given by Y I—> Hom(Y, X).
5. The assignment X I—> F‘un(X, Z) is a functor Sets —> Rings°p.
6. Let Q be a quiver. Consider the category C (Q) whose objects
are the vertices and morphisms are oriented paths between them.
Then functors from C (Q) to Vectk are representations of Q over is.
7. Let K C G be groups. Then we have the induction func­
tor Indg : Rep(K) —> Rep(G) and the restriction functor Resg :
Rep(G) —> Rep(K).
8. We have an obvious notion of the Cartesian product of cat­
egories (obtained by taking the Cartesian products of the classes of
objects and morphisms of the factors). The functors of direct sum
and tensor product are then functors Vectk >< Vectk ——> Vectk. Also
the operations V I—> V®”, V I—> S'”V, V I—> /\"V are functors on
Vectk. More generally, if 7r is a representation of Sn, we have functors
V I—> Homgn (7r, V8’ ”). Such functors are called the Schur functors.
Thus, the irreducible Schur functors are labeled by Young diagrams.
9. The reflection functors Ff : Rep(Q) —> Rep(E) are functors
between representation categories of quivers.
7.3. Morphisms of functors 181
7 .3. Morphisms of functors
One of the important features of functors between categories which
distinguishes them from usual maps or functions is that the functors
between two given categories themselves form a category; i.e., one can
define a nontrivial notion of a morphism between two functors.
Definition 7.3.1. Let C,D be categories and let F,G : C —> D be
functors between them. A morphism a : F —> G (also called a natu­
ral transformation or a functorial morphism) is a collection of
morphisms a X : F (X) —> G (X ) labeled by the objects X of C, which
is functorial in X; i.e., for any morphism f : X —> Y (for X, Y E C)
one has ay oF(f) = G(f) oax.
A morphism a : F —> G is an isomorphism if there is another
morphism a“1 : G —> F such that aoa‘1 and a‘1 oa are the identities.
The set of morphisms from F to G is denoted by Hom(F, G).
Example 7.3.2. 1. Let FVectk be the category of finite dimensional
vector spaces over k. Then the functors id and ** on this category
are isomorphic. The isomorphism is defined by the standard maps
av 2 V —> V** given by av(u)(f) = f(u), u 6 V, f 6 V*. But
these two functors are not isomorphic on the category of all vector
spaces Vectk, since for an infinite dimensional vector space V, V is
not isomorphic to V**.
2. Let FVect;c be the category of finite dimensional k-vector
spaces, where the morphisms are the isomorphisms. We have a func­
tor F from this category to itself sending any space V to V* and any
morphism a to (a*)‘1. This functor satisfies the property that V is
isomorphic to F (V) for any V, but it is not isomorphic to the identity
functor. This is because the isomorphism V I—> F(V) = V* cannot
be chosen to be compatible with the action of GL(V), as V is not
isomorphic to V* as a representation of GL(V).
3. Let A be an algebra over a field k, and let F : A — mod —>
Vectk be the forgetful functor. Then as follows from Problem 2.3.17,
EndF = Hom(F, F) = A.
4. The set of endomorphisms of the identity functor on the cate­
gory A — mod is the center of A (check it!).
182 7. Introduction to categories
7.4. Equivalence of categories
When two algebraic or geometric objects are isomorphic, it is usually
not a good idea to say that they are equal (i.e., literally the same).
The reason is that such objects are usually equal in many different
ways, i.e., there are many ways to pick an isomorphism, but by saying
that the objects are equal, we are misleading the reader or listener into
thinking that we are providing a certain choice of the identification,
which we actually do not do. A vivid example of this is a finite
dimensional vector space V and its dual space V*.
For this reason, in category theory, most of the time one tries to
avoid saying that two objects or two functors are equal. In particular,
this applies to the definition of isomorphism of categories.
Namely, the naive notion of isomorphism of categories is defined
in the obvious way: a functor F : C ——> D is an isomorphism if there
exists F‘1 : D ——> C such that F o F‘1 and F‘1 o F are equal to
the identity functors. But this definition is not very useful. We
might suspect so since we have used the word “equal” for objects of
a category (namely, functors) which we are not supposed to do. In
fact, here is an example of two categories which are “the same for
all practical purposes” but are not isomorphic; it demonstrates the
deficiency of our definition.
Namely, let C1 be the simplest possible category: Ob(C1) consists
of one object X, with Hom(X,X) = {lx}. Also, let C2 have two
objects X, Y and four morphisms: 1X, 1y, a : X ——> Y, and b : Y —> X.
So we must have aob= 1y, boa = 1X.
It is easy to check that for any category D, there is a natural
bijection between the collections of isomorphism classes of functors
C1 ——> D and C2 ——> D (both are identified with the collection of iso­
morphism classes of objects of D). This is what we mean by saying
that C1 and C2 are “the same for all practical purposes”. Neverthe­
less they are not isomorphic, since C1 has one object and C2 has two
objects (even though these two objects are isomorphic to each other).
This shows that we should adopt a more flexible and less restric­
tive notion of isomorphism of categories. This is accomplished by the
definition of an equivalence of categories.
7.5. Representable functors 183
Definition 7.4.1. A functor F : C —> D is an equivalence of cat­
egories if there exists F’ : D —> C such that F o F’ and F’ o F are
isomorphic to the identity functors.
In this situation, F’ is said to be a quasi-inverse to F.
In particular, the above categories C1 and C2 are equivalent (check
it!).
Also, the category FSet of finite sets is equivalent to the category
whose objects are nonnegative integers, and morphisms are given by
Hom(m,n) = Maps({1, . . . ,m}, {1, . . .
Are these categories isomorphic? The answer to this question depends
on whether you believe that there is only one finite set with a given
number of elements, or that there are many of those. It seems better
to think that there are many (without asking “how many”), so that
isomorphic sets need not be literally equal, but this is really a matter
of choice. In any case, this is not really a reasonable question; the
answer to this question is irrelevant for any practical purpose, and
thinking about it will give you nothing but a headache.

7 .5. Representable functors

A fundamental notion in category theory is that of a representable
functor. Namely, let C be a (locally small) category, and let F :
C ——> Sets be a functor. We say that F is representable if there
exists an object X E C such that F is isomorphic to the functor
Hom(X, '2). More precisely, if we are given such an object X, together
with an isomorphism 5 : F E’ Hom(X, 7), we say that the functor F
is represented by X (using 5
In a similar way, one can talk about representable functors from
C°P to Sets. Namely, one calls such a functor representable if it is of
the form Hom(?, X) for some object X E C, up to an isomorphism.
Not every functor is representable, but if a representing object X
exists, then it is unique. Namely, we have the following lemma.
Lemma 7.5.1 (The Yoneda Lemma). If a functor F is represented
by an object X, then X is unique up to a unique isomorphism. I.e.,
if X ,Y are two objects in C, then for any isomorphism of functors
184 7. Introduction to categories
qb : Hom(X, 7) —> Hom(Y, 7) there is a unique isomorphism a¢ : X ——>
Y inducing qb.

Proof (Sketch). One sets a¢ = q5}",1(1y) and shows that it is invert­

ible by constructing the inverse, which is aqjl = q5X(1X). It remains
to show that the composition both ways is the identity, which we
will omit here. This establishes the existence of a¢. Its uniqueness is
verified in a straightforward manner. Cl
Remark 7.5.2. In a similar way, if a category C is enriched over
another category D (say, the category of abelian groups or vector
spaces), one can define the notion of a representable functor from C
to D.

Example 7.5.3. Let A be an algebra. Then the forgetful functor

from the category of left A—modules to the category of vector spaces
is representable, and the representing object is the free rank 1 module
(= the regular representation) M = A. But if A is infinite dimen­
sional and we restrict attention to the category of finite dimensional
modules, then the forgetful functor, in general, is not representable
(this is so, for example, if A is the algebra of complex functions on Z
which are zero at all but finitely many points).

7.6. Adjoint functors

Another fundamental notion in category theory is the notion of ad­
joint functors.
Definition 7.6.1. Functors F : C —> D and G : D —> C are said to
be a pair of adjoint functors if for any X E C, Y E D we are given
an isomorphism fxy : Homc(F(X), Y) —> HOH11)(X, G(Y)) which is
functorial in X and Y, i.e., if we are given an isomorphism of functors
Hom(F(?), ?) —> Hom(?, G(?)) (C x D ——> Sets). In this situation, we
say that F is left adjoint to G and G is right adjoint to F.
Not every functor has a left or right adjoint, but if it does, it
is unique and can be constructed canonically (i.e., if we somehow
found two such functors, then there is a canonical isomorphism be­
tween them). This follows easily from the Yoneda lemma, since if
7.6. Adjoint functors 185

Table 1. Dictionary between category theory and linear algebra.

Category C Vector space V with a nondegen—

erate inner product
The set of morphisms Hom(X, Y) Inner product (:z:,y) on V (maybe
nonsymmetric)
Opposite category C°P Same space V with reversed inner
product
The category Sets The ground field k
F‘ull subcategory in C Nondegenerate subspace in V
Functor F : C —> D Linear operator f : V —> W
Functor F : C —> Sets Linear functional f E V* =
Hom(V, k)
Representable functor Linear functional f E V* given by
f(v) = (uxv), u 6 V
Yoneda lemma Nondegeneracy of the inner prod­
uct (on both sides)
Not all functors are representable If dimV = oo, not Vf E V*,
f('v) = (uwv)
Left and right adjoint functors Left and right adjoint operators
Adjoint functors don’t always exist Adjoint operators may not exist if
dim V = oo
If they do, they are unique If they do, they are unique
Left and right adjoints may not co­ The inner product may be non­
incide symmetric

F, G are a pair of adjoint functors, then F (X) represents the func­

tor Y »—> Hom(X,G(Y)) and G(Y) represents the functor X »—>
Hom(F(X), Y).
Remark 7.6.2. The terminology “left and right adjoint functors”
is motivated by the analogy between categories and inner product
spaces. More specifically, in Table 1 we have a useful dictionary
between category theory and linear algebra, which helps to better
understand many notions of category theory.
186 7. Introduction to categories
Example 7.6.3. 1. Let V be a finite dimensional representation of
a group G or a Lie algebra g. Then the left and right adjoint to the
functor V® on the category of representations of G is the functor
V*®.
2. The functor Resfi is left adjoint to Indg. This is nothing but
the statement of the Frobenius reciprocity.
3. Let Assock be the category of associative unital algebras, and
let Liek be the category of Lie algebras over some field k. We have a
functor L : Assock —> Liek which attaches to an associative algebra
the same space regarded as a Lie algebra, with bracket [a, b] = ab—ba.
Then the functor L has a left adjoint, which is the functor U of taking
the universal enveloping algebra of a Lie algebra.
4. We have the functor GL1 : Assock —> Groups, given by
A »—> GL1(A) = Ax. This functor has a left adjoint, which is the
functor G v—> k:[G], the group algebra of G.
5. The left adjoint to the forgetful functor Assock ——> Vectk
is the functor of tensor algebra: V »—> TV. Also, if We denote by
Commk the category of commutative algebras, then the left adjoint
to the forgetful functor Commk —> Vectk is the functor of the sym­
metric algebra: V »—> S’ V.
One can give many more examples, spanning many fields. These
examples show that adjoint functors are ubiquitous in mathematics.

7.7. Abelian categories

The type of categories that most often appears in representation the­
ory is abelian categories. The standard definition of an abelian cate­
gory is rather long, so We will not give it here, referring the reader to
the textbook rather, We will use as the definition what is really
the statement of the Freyd-Mitchell theorem:
Definition 7.7.1. An abelian category is a category (enriched over
the category of abelian groups) which is equivalent to a full subcat­
egory C of the category A-mod of left modules over a ring A, closed
under taking finite direct sums, as Well as kernels, cokernels, and
images of morphisms.
7.8. Complexes and cohomology 187
Example 7.7.2. The category of modules over an algebra A and the
category of finite dimensional modules over A are abelian categories.
Problem 7.7.3. Let A be a finitely generated commutative ring.
Show that the category of finitely generated A—modules is an abelian
category.
Hint: Use the Hilbert basis theorem.
We see from this definition that in an abelian category, Hom(X, Y)
is an abelian group for each X ,Y, compositions are group homo­
morphisms with respect to each argument, there is the zero object,
the notions of an injective morphism (monomorphism) and surjec­
tive morphism (epimorphism), and every morphism has a kernel, a
cokernel, and an image.
Remark 7.7.4. The good thing about Definition 7.7.1 is that it
allows us to visualize objects, morphisms, kernels, and cokernels in
terms of classical algebra. But the definition also has a big drawback,
which is that even if C is the whole category A-mod, the ring A is not
determined by C. In particular, two different rings can have equivalent
categories of modules (such rings are called Morita equivalent).
Actually, it is worse than that: for many important abelian categories
there is no natural (or even manageable) ring A at all. This is why
people prefer to use the standard definition, which is free from this
drawback, even though it is more abstract.
Let k: be a field. We say that an abelian category C is kt-linear
if the groups Homc(X, Y) are equipped with a structure of a vector
space over k: and composition maps are k:—linear in each argument. In
particular, the categories in Example 7.7.2 are k—linear.

7.8. Complexes and cohomology

Definition 7.8.1. A sequence of objects C,-,2’ E Z, of an abelian
category C and morphisms di : Ci —> C,-+1 is said to be a complex if
the composition of any two consecutive arrows is zero. The morphisms
d,- are called the differentials.1 The cohomology of this complex is
1A famous example of a complex is the de Rham complex, in which Cm is the
space of differential m-forms on a Euclidean space or, more generally, a manifold, and
dm is the exterior differential of differential forms. This explains the term “differential”.
188 7. Introduction to categories
H i = Ker(d,-)/Im(d,--1). The complex is said to be exact in the ith
term if H i = 0 and is said to be an exact sequence if it is exact in
all terms.

There is an obvious notion of a morphism of complexes. Name­

ly, a morphism f : C —> D is a collection of morphisms f,- : C, —> D,­
such that d,- o f,. = ft.“ 0 dz-. Clearly, such morphisms can be com­
posed, which makes the class of all complexes over C into a category,
denoted Compl(C).
In particular, one can consider complexes of abelian groups, vec­
tor spaces, modules over a ring, etc. In this case, elements of Ker(d,-)
are called zlcocycles, the elements of Im(d,--1) are called z'—cobound­
aries, and the elements of H‘ (C) are called z'th cohomology classes.
This terminology is adopted from topology. For this reason, exact
sequences are also called acyclic complexes (as they are complexes
which have no nontrivial cocycles, i.e., cocycles that are not cobound—
aries).
Often one considers complexes that are bounded in one or both
directions; i.e., the objects Ci are zero for 2' >> 0, 2' << 0, or both.
In this case one writes one zero on each side where the complex is
bounded. For example, a complex bounded on both sides with n + 1
terms will look like

0—>C0—>C1—>---——>Cn—>0.

Definition 7.8.2. A short exact sequence is an exact sequence of

the form
0—>X—>Y—>Z——>0.

Clearly, 0 —> X —> Y —> Z —> 0 is a short exact sequence if and

only if X —> Y is injective, Y —> Z is surjective, and the induced map
Y/ X —> Z is an isomorphism. In other words, short exact sequences
correspond to extensions of Z by X.
Example 7.8.3. The sequence 0 —> X —> X 619 Z —> Z —> 0 with
the obvious morphisms is a short exact sequence. Such a sequence is
called split. It corresponds to the trivial extension of Z by X.
7.8. Complexes and cohomology 189
Exercise 7.8.4. Show that any exact sequence of vector spaces is
isomorphic to a direct sum of complexes of the form
0—>V—>V—>0,
where V stands at the places 2' and 2' + 1 and the map V —> V is the
identity (in particular, any short exact sequence of vector spaces is
split). Is this true in the category of abelian groups?
Problem 7.8.5. Let D. be a complex of abelian groups with differ­
entials d.-, 2' E Z, let C. be a subcomplex of D. (i.e. a collection of
subgroups C. C Di such that d.-(Ci) C CH1), and let E. = D. / C.
be the quotient complex (i.e., E. = D,-/Ci with differentials induced
by dz’)­
Define a homomorphism c,; : H —> H f+1(C) as follows. Given
:1: E H i(E), pick a representative :13’ of :1: in E.-. Let :13” be a lift of :13’
to D,-. Let y’ = dx" E D,-+1 (We abbreviate the notation, denoting d.,;
just by d). Since dx’ = 0, y’ = dx" E C.-+1. Also, dy’ = d2:I:” = 0. So
y’ represents an element y E H l+1(C). We will set C,-(:13) = y.
(i) Show that c,- is Well defined, i.e., C.-(:13) does not depend on the
choice of :13’ and :13”.
(ii) Show that the sequence
(7.8.1) ...H"’1(E) —> H‘(C) —> H‘(D) —> H*'(E) —> H‘+1(C) . . .,
where the first map is c,--1, the middle two maps are induced by the
maps C,; —> D,- —> E.-, and the last map is c,-, is exact.
Definition 7.8.6. The map c,- is called the connecting homomor­
phism, and the sequence (7.8.1) is called the long exact sequence
of cohomology.

Problem 7.8.7. Let C. and D. be complexes of modules over a

commutative ring A. Define the tensor product complex (C ® D). by
the formula
(C® D). = Q3 C, 69,; Dm,
j+m=1l
with differentials

d.C®D|Cj®Dm = dgc 59 1 + (—1)j ° 1 ® da­

190 7. Introduction to categories
(i) Show that this is a complex.
Now assume that A = Is: is a field.
(ii) Show that if C or D is an exact sequence, then so is C (8 D.
Hint: Use the decomposition of Exercise 7.8.4.
(iii) Show that any complex C can be identified with a direct
sum of an exact sequence and the complex consisting of H with
the zero differentials, in such a way that the induced isomorphism
H‘(C) —> H‘(C) is the identity.
(iv) Show that there is a natural isomorphism of vector spaces
H‘(C® D) 2 Q9 Hj(C) ca H”"(D).
j+m=z'
This is the Kiinneth formula.

7.9. Exact functors

Definition 7.9.1. A functor F between two abelian categories is
additive if it induces homomorphisms on Hom groups. Also, for
k—linear categories one says that F is ls:-linear if it induces k—linear
maps between Hom spaces.
It is easy to show that if F is an additive functor, then F (X EBY)
is canonically isomorphic to F (X ) 69 F
Example 7.9.2. The functors Indfz, Resg, HomG(V, ?) in the theory
of group representations over a field is: are additive and k—linear.
Definition 7.9.3. An additive functor F : C —> D between abelian
categories is left exact if for any exact sequence
0—+X—>Y—>Z,
the sequence
0 —+ F(X) —> F(Y) —> F(Z)
is exact. F is right exact if for any exact sequence
X—>Y—>Z—>0,
the sequence
F(X) —>F(Y) —> F(Z) —>0
is exact. F is exact if it is both left and right exact.
7.9. Exact functors 191
Definition 7.9.4. An abelian category C is semisimple if any short
exact sequence in this category splits, i.e., is isomorphic to a sequence

0—>X—>Xe9Y—>Y—>0

(where the maps are obvious).

Example 7.9.5. The category of representations of a finite group G

over a field of characteristic not dividing |G'| (or 0) is semisimple.

Note that in a semisimple category, any additive functor is auto­

matically exact on both sides.

Example 7.9.6. The functors Indfz, Res?» are exact.

(ii) The functor Hom(X, ?) is left exact, but not necessarily right
exact. To see that it need not be right exact, it suffices to consider
the exact sequence

0—>Z—>Z—>Z/2Z—>0

and to apply the functor Hom(Z/2Z, ?).

(iii) The functor X (8 A for a right A—module X (on the category
of left A—modules) is right exact but not necessarily left exact. To see
this, it suflices to tensor multiply the above exact sequence by Z/2Z.

Exercise 7.9.7. Show that if (F, G) is a pair of adjoint additive

functors between abelian categories, then F is right exact and G is
left exact.

Exercise 7.9.8. (a) Let Q be a quiver and let 2' E Q be a source.

Let V be a representation of Q and let W be a representation of
(the quiver obtained from Q by reversing arrows at the vertex
Prove that there is a natural isomorphism between Hom (F,."V, W)
and Hom (V, F,-+W). In other words, the functor F,.+ is right adjoint
to Ff.
(b) Deduce that the functor F,+ is left exact and F,’ is right
exact.
192 7. Introduction to categories
7.10. Historical interlude: Eilenberg, Mac Lane,
and “general abstract nonsense”
Saunders Mac Lane’s (1909-2005) father and grandfather were both
Congregational ministers, so young Saunders seemed destined for an
ecclesiastical career. Early on, however, he started having “various
theological doubts and questions”. “[A] ministerial career would have
required my belief of things of which I was uncertain”, Mac Lane
later recalled. “Mathematics, however, provided a different sort of
certainty”. One thing he was certain about was that the minister’s
income was woefully inadequate to provide for comfortable retire­
ment. As a result, Mac Lane decided to pursue a career that would
be “scientific rather than ministerial” [33, pp. 349-350]. He earned
his bachelor’s degree at Yale and his master’s at Chicago.
Inspired by the “elegance of German abstract algebra”, in which
the notions of rings and fields axiomatized addition and multiplica­
tion, Mac Lane devoted his master’s thesis to a similar axiomatization
of exponentials, as well as plus and times. At Chicago, Mac Lane was
put off by inadequate teaching and by examples of what he saw as
a “display of pedantry”, and he concluded that the mathematics de­
partment there “had ceased to be really first class” [36, p. 152]. He
recalled that he had been taught that a vector was an n-tuple, “an
idea that I soon had to unlearn” [33, p. 39]. Mac Lane wanted to
write a doctoral dissertation on logic but could not find an advisor
and decided to leave for Gottingen, the preeminent center of mathe­
matical logic research.
At Gottingen, the home of the famous Hilbert school, Mac Lane
learned group representations from Hermann Weyl and linear asso­
ciative algebras from Emmy Noether. “To live in Gottingen was to
be immersed in mathematical culture”, Mac Lane wrote, recalling
Gottingen as “an amalgam of research, teaching, and inspiration”.
Despite this immersion, Mac Lane could hardly ignore the tense po­
litical climate, the vociferous clashes among 27 political parties, and
the students dressed in the Nazi brownshirt uniform. Mac Lane did
write a thesis on logic but managed to overlook G6del’s famous incom­
pleteness theorem, which appeared at that time. His thesis elicited
7.10. Historical interlude: Eilenberg, Mac Lane 193
common schemes of proof in Russell and Whitehead’s P7‘z'ncz'pz'a Math­
ematica with the aim of abbreviating and perhaps mechanizing new
proofs. The thesis was “naive in supposing that all mathematics
would be actually written in Whitehead and Russell style”, Mac Lane
admitted. Mac Lane was in G6ttingen in 1933, when Hitler came
to power, and he witnessed the ruthless destruction of the Hilbert
school, as many leading Jewish mathematicians, including Courant
and Noether, were dismissed from their jobs. The once famous math­
ematics department became “but a skeleton”. Mac Lane rushed up
to finish his thesis. Unimpressive by G6ttingen standards, the thesis
was accepted with the grade Geniigend, the lowest grade. Mac Lane
became the last American to earn a doctorate in the famous Hilbert
school. One more thing Mac Lane managed to do before departure
was to get married to Dorothy Jones. Being at Géttingen as he was,
Mac Lane went to hear a mathematical lecture right after his wedding
ceremony [33, pp. 48, 54, 51-52, 58, 52, 56].
In the 1930s mathematical logic was looked down upon by most
mathematicians, who considered it part of philosophy rather than of
mathematics proper. When he started teaching as an instructor at
Harvard, Mac Lane intended to offer a logic course, but departmen­
tal needs forced him to lecture on algebra. Eventually he realized
that it was easier to get a job as an algebraist than as a logician. In
1941, Mac Lane and Garrett Birkhoff co—authored Survey of Modern
Algebra, the first American undergraduate textbook that espoused
the abstract approach of Emmy Noether and B. L. van der Waer—
den. Initially judged as a text that “won’t fly beyond the Hudson”,
the book eventually became a standard textbook for undergraduate
algebra courses [33, pp. 62, 68, 82].
In the spring of 1941, Mac Lane received a semester leave from
Harvard and applied to visit the Institute for Advanced Study at
Princeton. His application was turned down, and Mac Lane ended
up going to the University of Michigan to give a series of six lectures.
One of the Michigan mathematicians in the audience was Samuel
(“Sammy”) Eilenberg (1913-1988). He sat through the first five lec­
tures but had to miss the last one, and he asked Mac Lane to give
him the gist of it privately. Mac Lane did, and Eilenberg noticed a
194 7. Introduction to categories
similarity between Mac Lane’s results on group extensions and his col­
league Norman Steenrod’s homology theory of solenoid—type spaces.
“It was startling” to find that connection, Mac Lane later recalled.
They stayed up all night and by morning they found not only “a new
intrusion of algebra into topology” but also formed a lifelong collabo­
rative duet, which eventually churned out 15 joint papers. Mac Lane
later remarked that the rejection of his application to Princeton was
fortuitous, for “had it been accepted, I might have missed working
with Sammy” [33, pp. 101-104].
A Polish Jew, Eilenberg left Poland shortly before the start of
World War II, following the truly wise advice of his father, once the
best student in his town’s yeshiva. Among other European refugee
mathematicians, Eilenberg was welcomed by Oswald Veblen and Solo­
mon Lefschetz of Princeton University, who helped him find a posi­
tion at Michigan. For Eilenberg, mathematics was a social activity.
“Though his mathematical ideas may seem to have a kind of crys­
talline austerity, Sammy was a warm, robust, and very animated hu­
man being”, whose charm and humor were “hard to resist”, recalled
his former colleague [6, pp. 1351-1352]. Described as “energetic ...,
expressive, charismatic, quick witted, often confrontational, and a
brilliant thinker and good—humored conversationalist”, Eilenberg eas­
ily established contacts [7, p. 360]. His entire career was punctuated
by various productive collaborations. Together with Steenrod, he
“drained the Pontine Marshes of homology theory, turning an ugly
morass of variously motivated constructions into a simple and elegant
system of axioms applied, for the first time, to functors”. In collabo­
ration with Henri Cartan, Eilenberg systematized homology theories
in their Homological Algebra (1956). This field developed so rapidly
that within a few years Alexander Grothendieck reportedly branded
this foundational text “le diplodocus”, relegating it to the subject of
paleontology [6, p. 1348].
Eilenberg believed that “the highest value in mathematics was to
be found, not in specious depth nor in the overcoming of overwhelm­
ing difficulty, but rather in providing the definitive clarity that would
illuminate its underlying order”. He insisted on “lucidity, order, and
understanding as opposed to trophy hunting”. Eilenberg was called
7.10. Historical interlude: Eilenberg, Mac Lane 195
the “greatest mathematical stylist”, perhaps to the point of champi­
oning a “triumph of style over substance” [6, pp. 1349, 1348, 1344,
1350}
Eilenberg’s style called for a minimalist approach, which his col­
league described as “always to find the absolutely essential ingredients
in any problem and work only with those ingredients and nothing else
— in other words, get rid of all the superfluous information”. Once
asked if he could eat Chinese food with three chopsticks, he immedi­
ately answered in the affirmative and explained, “I’ll take the three
chopsticks, I’ll put one of them aside on the table, and I’ll use the
other two” [7, p. 361].
Eilenberg and Mac Lane pushed the homological envelope into
more complicated spaces in their joint 1942 paper “Natural Isomor—
phisms in Group Theory”, which introduced the notion of a functor.
André Weil, who reviewed their article in Mathematical Reviews, did
not merely summarize it but added a remarkable note of praise. He
wrote that the authors succeeded in finding “a precise definition” for
the “vague idea of covariance and contravariance”, “which is likely to
be helpful in classifying and systematizing known results and also in
looking for new relations between groups” (Weil in [4, p. 743]).
In the war years both Mac Lane and Eilenberg came to New York
to work for the Applied Mathematics Group at Columbia University
under the National Defense Research Council. They toiled on prob­
lems of airborne fire control during the day and on pure mathematics
at night. While dealing with the concrete problem of computing the
cohomology of the solenoid, they arrived at their universal coefficient
theorem and the very general and abstract notions of category and
functor. “These notions were so general”, Mac Lane recalled, “that
they hardly seemed to be real mathematics — would our mathemati­
cal colleagues accept them?” Mac Lane and Eilenberg hid their anx­
iety behind the jocular label “general abstract nonsense”, by which
they occasionally referred to their subject. “We didn’t really mean
the nonsense part, and we were proud of its generality”, admitted
Mac Lane [33, pp. 125-126].
In 1945 Eilenberg and Mac Lane summed up their findings in
the article “General Theory of Natural Equivalences”, which for the
196 7. Introduction to categories
first time defined categories. The article was reportedly “rejected
by the editor of an inauspicious journal as ‘more devoid of content
than any I have read,’ to which Mac Lane is said to have replied,
‘That’s the point’ ” [7, p. 362]. Eventually they managed to publish
it in the Transactions of the American Mathematical Society. This in­
volved a fair amount of wheeling and dealing. “Eilenberg, who knew
the editor well, persuaded him to choose as referee a young mathe­
matician”, Mac Lane recalled, “one whom we could influence because
he was then a junior member of the Applied Mathematics Group at
Columbia University (war research), where Eilenberg and I were then
also members, and I was Director” [37, p. 5983]. The article openly
declared Eilenberg and Mac Lane’s ambitious goals. “In a metamath­
ematical sense our theory provides general concepts applicable to all
branches of mathematics”, they wrote, “and so contributes to the
current trend towards uniform treatment of different mathematical
disciplines” [12, p. 236].
The wide recognition of the power of Eilenberg and Mac Lane’s
ideas was a bit slow to come. Years passed before the words category
and functor “could be pronounced unapologetically in diverse math­
ematical company” (Freyd in [6, p. 1351]). Luckily, the Air Force
Office of Scientific Research became an important sponsor of research
and conferences on category theory — an unexpected benefit of Mac
Lane and Eilenberg’s wartime work. Only by the mid-1960s did the
Air Force eventually figure out that category theory was too abstract
to bring any tangible improvements to air combat and discontinued
its support [33, p. 239].
Category theory itself underwent quite a transformation after the
initial insight. As Mac Lane recalled, when the 1945 paper came out,
Eilenberg declared that “this would be the only paper necessary on
the subject. It turned out that he was wrong” [33, p. 209]. Eilenberg
and Mac Lane originally thought that the very concept of a category
was “essentially an auxiliary one; our basic concepts are essentially
those of a functor and of a natural transformation”. They even sug­
gested dropping the category concept altogether and adopting “an
even more intuitive standpoint”, according to which a functor “is not
defined over the category of ‘all’ groups, but for each particular pair
7.10. Historical interlude: Eilenberg, Mac Lane 197
of groups which may be given” [12, p. 247]. Moreover, although Mac
Lane and Eilenberg believed that category theory “offered a concep­
tual view of parts of mathematics”, they regarded it merely as “a
handy language”, “a language of orientation”, not as a “field for fur­
ther research effort” [35, pp. 334-335].
Although Eilenberg and Mac Lane invented category theory, for
them “it was always an applied subject, not an end in itself. Cat­
egories were defined in order to define functors, which in turn were
defined in order to define natural transformations, which were de­
fined finally in order to prove theorems that could not be proved
before”. They viewed categories as instruments of mainstream math­
ematics. The idea that categories could be used “to state theorems
that could not be stated before, that they were not tools but ob­
jects of nature worthy of study in their own right”, sounded alien to
them, for it seemed to place category theory on the fringe, outside
the mainstream (Freyd in [6, p. 1351]). In the hands of Alexander
Grothendieck, Daniel Kan, and Eilenberg’s students, however, cat­
egory theory became a dynamic field in its own right, profoundly
transformed algebraic geometry, topology, and representation theory,
and resonated across a wide range of disciplines, from mathematical
logic and theoretical computer science to linguistics and philosophy.
In the work of William Lawvere, category theory came to be regarded
as a foundation for all of mathematics (Heller in [6, p. 1349]; [7, p.
362]).
Category theory caused a controversy within the group of French
mathematicians who worked under the pseudonym Nicolas Bourbaki
with the aim of producing a multivolume rigorous treatment of all
contemporary mathematics from a unified conceptual point of view.
On Weil’s invitation, Eilenberg joined the group in 1950 and re­
mained an active member for 15 years. “He knew admirably how
to present his point of view, and he often made us agree to it”,
recalled Henri Cartan, another Bourbaki member (Cartan in [6, p.
1345]). Mac Lane was invited to attend a private Bourbaki meeting
in 1954, perhaps with the idea that he would persuade the group to
adopt the category-theoretical approach. Cartier, Chevalley, Eilen­
berg, and Grothendieck supported Mac Lane, but André Weil took
198 7. Introduction to categories
a cautious approach, which outweighed them [39, p. 5]. While gen­
uinely interested in incorporating category theory, Bourbaki found it
very difficult to abandon its prewar foundational concept of struc­
ture. Switching to categories would mean revising the whole body
of previous work, while combining the two approaches seemed to un­
dermine the Bourbakist ideal of unity of mathematics [5, p. 236].
“My facility in the French language was not sufficient to categorize
Bourbaki”, remarked Mac Lane. As a result, important parts of re­
cent mathematical research fell out of Bourbaki’s purview, perhaps
contributing to the decline of their overall project. Mac Lane barely
hid his disappointment in the outcome, calling the Bourbaki oeuvre
a “magnificent multi-volume monster” and Bourbaki’s definition of
structure “a cumbersome piece of pedantry”, which even Bourbaki
never used (quoted in [39, p.
Ironically, the Bourbakist vision of unified mathematics found
a powerful vehicle in category theory. General category-theoretical
concepts have enjoyed wide use across many different mathematical
fields; in fact, the very rigid division of mathematics into fields has
been called into question. “Mathematics as it is today ...can no
longer be presented by piecemeal courses”, argued Mac Lane in a
1954 talk, “for it is simply no longer true that advanced mathematics
can be split neatly into compartments labelled ‘algebra’, ‘analysis’,
‘geometry’, and ‘applied mathematics”’. He championed the “infu­
sion of a geometrical point of view” and reminded his audience that
“a vector is geometrical”. “A vector is not an n-tuple of numbers until
a coordinate system has been chosen. Any teacher and any text book
which starts with the idea that vectors are n-tuples is committing a
crime for which the proper punishment is ridicule” (quoted in [40, p.
245]).
Mac Lane’s favorite pastime was sailing, which he considered
“a much more serious sport” than skiing, which he had learned at
Gottingen. On some sailing expeditions he spent time with prominent
philosophers but showed little interest in their subject. “Thankfully,
we paid more attention to sailing than to philosophical doctrine”, he
recalled. He found that most studies in philosophy of mathematics
“paid little attention to the actual substance of mathematics beyond
7.10. Historical interlude: Eilenberg, Mac Lane 199
the most elementary concerns” and set out to repair the situation. In
his 1986 book Mathematics, Form and Function, Mac Lane quickly
dispatched with an array of mathematical philosophies, from set the­
ory (“often artificial”) to formalism (“can’t explain which of many
formulas matter”) to intuitionism (“can be dogmatic”) to empiricism
(“mathematics originates not just in facts”) to Platonism (“a useful
mythology and a speculative ontology”). He even dismissed the R0­
mantic ideal of mathematics as “a search for austere forms of beauty”
and offered instead “formal functionalism” — his vision of mathemat­
ics as “an extensive network of formal rules, definitions and systems,
tightly tied here and there to activities and to science”. Since mathe­
matical entities do not necessarily correspond to any physical objects,
Mac Lane argued, it would be meaningless to ask if mathematics is
true. The “appropriate” questions about a particular piece of math­
ematics are different: is it correct (i.e., proved), responsive to some
problem, illuminating, promising, relevant? According to Mac Lane,
mathematics is correct but not true [33, pp. 143, 145, 455-456, 440,
441,443]
Like Mac Lane, Eilenberg “had no patience for metaphysical ar­
gument” (Heller in [6, p. 1349]). He had his own passion outside
mathematics, however. “A sophisticated and wise man who took a
refined pleasure in life” (Bass in [6, p. 1352]), Eilenberg, in fact, had
a parallel life. Known as “Sammy” to mathematicians, Eilenberg was
equally famous as an art dealer and collector under the nickname
“Professor” in the art world. Over the years, he amassed one of the
most significant collections of Southeast Asian sculpture in the world
and became a leading expert in the field. His biographer surmised
that Eilenberg “found in ancient Hindu sculpture a formal elegance
and imagination that resonated well with the same aesthetic sensibil­
ity — ‘classical rather than romantic,’ in the words of Alex Heller ­
that animated his mathematical work” [7, p. 362]. Eilenberg’s two
lives rarely intersected. His colleagues recalled only one occasion on
which he “moved from a conversation about sculpture to one about
mathematics. Sculptors, he said, learn early to create from the in­
side out: what finally is to be seen on the surface is the result of a
lot of work in conceptualizing the interior. But there are others for
whom the interior is the result of a lot of work on getting the surface
200 7. Introduction to categories
right. ‘And,’ Sammy asked, ‘isn’t that the case for my mathemat­
ics?’” (Freyd in [6, p. 1350]).
Mac Lane called his collaboration with Eilenberg “one small, but
typical, example of East meets West. The great influx of refugee
mathematicians from Europe presented a decisive stimulus for Amer­
ican mathematics in the 1940s” [33, p. 347]. The Eilenberg-Mac
Lane collaboration not only brought together algebra and topology;
it epitomized the cross-fertilization of different mathematical schools
and the breaking down of internal disciplinary barriers Within math­
ematics in the second half of the twentieth century.
Chapter 8

Homological algebra

In this chapter We develop some basic homological algebra tools,

which are necessary to study the fine structure of representation cat­
egories.

8.1. Projective and injective modules

Let A be an algebra, and let P be a left A—module.
Theorem 8.1.1. The following properties of P are equivalent:
Ifa : M —> N is a surjectiue morphism and 1/ : P —> N is
any morphism, then there exists a morphism ,u : P —> M such that
Q 0 /J, Z I/.
Any surjectiue morphism Oz : M —> P splits; i.e., there exists
,u:P—>M such thatozopzid.
There exists another A-module Q such that P 619 Q is a free
A—module, i.e., a direct sum of copies of A.
(iv) The functor HomA (P, ?) on the category of A-modules is ex­
act.

Proof. To prove that implies (ii), take N = P. To prove that

(ii) implies (iii), take M to be free (this can always be done since
any module is a quotient of a free module). To prove that (iii) im­
plies (iv), note that the functor HomA (P, ?) is exact if P is free (as

201
202 8. Homological algebra
HomA (A, N) = N), so the statement follows, since if the direct sum
of two complexes is exact, then each of them is exact. To prove that
(iv) implies (i), let K be the kernel of the map Oz, and apply the exact
functor HomA (P, ?) to the exact sequence
0—>K—>M—>N—>0.
Cl

Definition 8.1.2. A module satisfying any of the conditions (i)—(iv)

of Theorem 8.1.1 is said to be projective.

Problem 8.1.3. A right A-module M is said to be flat if the functor

M 8) A on the category of left A-modules is exact.
(i) Show that any projective module is flat.
(ii) Let A be a commutative ring and let S be any multiplicatively
closed subset of A. Then, the localization S ‘IA is a flat A-module.
(iii) Let A = (C[x], M = (C[x,a:‘1]. Show that M is flat but not
projective.

Exercise 8.1.4. Let A be a ring, let M1,M2 be left A-modules,

let P1,P2 be projective left A-modules, and let f,- : P, —> M, be
homomorphisms. Let M be a left A-module containing M1 such that
M /M1 = M2. Show that there exists a homomorphism f : P1 619 P2 —>
M such that f | P1 = fl and the induced homomorphism P2 —> M2 is
f2.

There is also a notion of an injective module, which is dual to

the notion of a projective module. Namely, we have the following
theorem.
Let A be an algebra and let I be a left A-module.
Theorem 8.1.5. The following properties of I are equivalent:
Ifa : N —> M is an injective morphism and 1/ : N —> I is
any morphism, then there exists a morphism ,u : M —> I such that
,u 0 Oz = 1/.
Any injective morphism oz : I —> M splits; i.e., there exists
,u:M—>I such that,uooz=id.
8.2. Tor and Ext functors 203
The functor HomA(?, I) on the category of A-modules is ex­
act.

Proof. The proof of the implications “(i) implies (ii)” and “(iii) im­
plies (i)” is similar to the proof of Theorem 8.1.1. Let us prove that
(ii) implies (iii). Let
0 —> N —> M —> K —> 0

be an exact sequence. Denote the embedding N —> M by j. Consider

the corresponding sequence
0 —> Hom(K, I) —> Hom(M, I) —> Hom(N, I) —> 0.
Let f E Hom(N, I), and define the module E := (MEBI)/N, Where N
is embedded into M 619 I via :17 r—> (j(a:), —f Clearly, we have an
inclusion I —> E, since the image of N 619 I in E is naturally identified
with I. So there is a splitting E —> I of this inclusion, i.e., a map
M 619 I —> I, (m,i) »—> g(m) +71 such that — f(x) = 0. This
means that the map j* : Hom(M, I) —> Hom(N, I) is surjective, i.e.,
the functor Hom(?, I) is exact, as desired. El
Definition 8.1.6. A module satisfying any of the conditions (i)—(iii)
of Theorem 8.1.5 is said to be injective.

Example 8.1.7. Let A be an algebra and P be a left A—module.

Then P is projective if and only if P* is an injective right A-module.

Theorem 8.1.1(iv) and Theorem 8.1.5(iii) can be used to define a

projective and an injective object in any abelian category. Namely,
We make the following definition.
Definition 8.1.8. A projective object in an abelian category C is
an object P such that the functor Home (P, ?) is exact.
An injective object in an abelian category C is an object I such
that the functor HomC(?, I) is exact.

8.2. Tor and Ext functors

Let A be a unital ring. As We have mentioned in Example 7.9.6, the
functors M 8) A? and HomA (M, ?) (where M is a right, respectively
204 8. Homological algebra
left, A-module) on the category of left A-modules are, in general,
not exact (they are only exact on one side). The job of the functors
Tor;'l(M, ?) and Extf'4 (M, ?) is to quantify the extent to which the
functors M 8) A? and HomA(M,?) fail to be exact. Namely, these
functors are defined as follows.
Definition 8.2.1. A projective resolution of M is an exact se­
quence
---—>P2—>P1—>P0—>M—>0
such that all modules P,-, 2 Z 0, are projective.

Exercise 8.2.2. Show that any module has a projective resolution

(for example, one consisting of free modules).
Definition 8.2.3. Let M be a right A-module, P. a projective resolu­
tion of M, and N a left A—module. For 2 2 0 we define Tor;"(M, N) =
Tor,-(M, N) to be the 2th cohomology of the complex
---—>P2®AN—>P1®AN—)P0®AN—>0
induced by the resolution P..
Definition 8.2.4. Let M be a left A-module, P. a projective resolu­
tion of M, and N a left A-module. For 2 Z 0 we define Ext; (M, N) =
Ext‘(M, N) to be the 2th cohomology of the complex
0 —> HomA(P0,N) —> HomA(P1,N) —> HomA(P2,N) —>
induced by the resolution P..
Problem 8.2.5. In this problem we will show that the cohomology
groups Tor, and Exti don’t really depend on the projective resolution
P., in a fairly strong sense, which justifies the fact that we don’t
mention P. in the notation for them.
Let P., Q. be two projective resolutions of M. Let df : P,- —>
P,--1, d? : Q, —> Q,--1 be the corresponding differentials; in particular,
d(I)D2P0—)M,d§ZQ0—)M.
(i) Show that there exists a homomorphism fo : P0 —> Q0 such
that 213 0 f0 = d{,’.
(ii) Proceed to show by induction in j that there exists a homo­
morphism fj : P, —> Q, such that d? o fj = f,-_1 0 df.
8.2. Tor and Ext functors 205
The collection of homomorphisms satisfying the conditions of
and (ii) is called a morphism of resolutions, f : P. —> Q..
(iii) Clearly, such a morphism f defines a linear map w,-(P, Q, f) :
Torf (M, N) —> Tor,.Q(M, N), where the superscripts P and Q mean
that the Tor groups are defined using the resolutions P. and Q.. Show
that the maps lb,-(P, Q, f) don’t really depend on f (so they can be
denoted by lb,-(P,
(iv) Deduce that lb,-(P, Q) are isomorphisms (use that lb,-(P, P) =

(v) Define similar maps {Z-(Q, P, f) : Ext’);(M, N) —> ExtiQ(M, N)

and show that they are independent on f and are isomorphisms.
Problem 8.2.6. Show that Tor0(M, N) = M ®A N and that
Ext°(M, N) = HomA(M, N).
(ii) Show that the group Extl (M, N) as defined here is canonically
isomorphic to the one defined in Problem 3.9.1.
(iii) Let
0—>N1—>N2—>N3—>0
be a short exact sequence of left A-modules.
Show that there are long exact sequences

0 —> HomA(M,N1) —> HomA(M,N2) —> HomA(M, N3) —>

Ext1(M, N1) —> Ext1(M, N2) —> Ext1(M, N3) —> Ext2(M, N1) —> ...
and

—> Tor2(M, N3) —> Tor1(M,N1) —> Tor1(M,N2) —> Tor1(M, N3)
—)M®AN1—>M®AN2—)M®AN3—)0.
This shows that, as we mentioned above, the Tor and Ext groups
“quantify” the extent to which the functors M 8) A and HomA (M, ?)
are not exact.
Hint: Use Problem 7.8.5.
(iv) Show that Torfl (M, N) can be computed by taking a projec­
tive resolution of N, tensoring it with M, and computing cohomology.
206 8. Homological algebra
Hint: Show first that Tor;4(M, N) can be computed by tensoring
two projective resolutions, for M and for N, and computing cohomol­
ogy.
(v) Let 0 —> M1 —> M2 —> M3 —> 0 be a short exact sequence
of left A-modules. Let P} and P,3 be projective resolutions of M1
and M3. Construct a projective resolution of M2 with terms P,-2 :=
P,-1 EDP,-3 (use Exercise 8.1.4). Use it to show that for any left A-module
N there are long exact sequences
0 —> HomA(M3,N) —> HomA(M2,N) —> HomA(M1,N) —>
Ext1(M3, N) —> Ext1(M2, N) —> Ext1(M1, N) —> Ext2(M3, N) —> ...
and
--- —> Tor2(M3,N) —> Tor1(M1, N) —> Tor1(M2,N) —> Tor1(M3, N)
—>M1®AN—>M2®AN—>M3®AN—>0.
Problem 8.2.7. Let A = Z and let ,N be finitely generated
abelian groups. Compute Tor,-(M, N), Ext’(M, N) (Hint: Reduce to
the case of cyclic groups using the classification theorem for finite
abelian groups.)
(ii) Do the same for A = k[:n] and M, N being any finitely gener­
ated A-modules.
Problem 8.2.8. Show that if A1, A2 are algebras over a field k and
M2-, N2 are left A2--modules, then
Tor;41®A2(M1 ® M2, N1 ® N2) =

Q3 Torfl (M1, N1) ® Tor;‘,2 (M2, N2).

j+m=’i
Similarly,
EXl3:41®A2 (M1 ® M2, N1 ® N2) =
Q3 Extih (M1, N1) ® Extfffz (M2, N2),
j+m='i
if N2 are finite dimensional.

In a similar Way one can define Tor and Ext groups in any abelian
category which has enough pro jectives; i.e., any object is a quotient
of a projective object (this condition insures that every object has a
projective resolution).
8.2. Tor and Ext functors 207
Exercise 8.2.9. Show that the category of finite abelian groups or
finite dimensional k[:n]-modules does not contain nonzero projective
objects (so it does not have enough pro jectives).
(ii) Let A be a finitely generated commutative ring. Show that
the category of finitely generated A-modules has enough projectives.
Problem 8.2.10. Let V be a finite dimensional vector space over
a field k. Let Ci = SV 8) /\"V. We can view Ci as the space of
polynomial functions on V* with values in /\"V. Define the differential
di : Ci —> Ci_1 by the formula
: "uf(u)> U’ E V*>
Where Lu : /\"V —> /\"‘1V is the contraction defined by the formula

Lu(’U1 /\ ° ° ° /\ vi‘) = Z(—1)ju('vj)'v1 /\ . . . '03‘ ° ° ° /\ ’Ui'.

J'=1

(i) Show that C. is a projective (in fact, free) resolution of k

(with trivial action of V) as an SV-module. (It is called the Koszul
resolution.)
(ii) Let V = U 69 W. Then We can view SW as an SV-module
(U acts by zero). Construct a resolution of SW by free SV-modules
whose terms are Ci = SV 8) /\"U.
Hint: Tensor the Koszul resolution of k as an S U -module by SW.
(iii) Regard SV as an S (V 619 V) = SV 8) SV-module using left
and right multiplication. Construct a free resolution of S’ V as an
S’ V ®S'V-module with terms Ci = S'V® /\"V®S' V (called the Koszul
bimodule resolution).
(iv) By tensoring the resolution of (iii) over S’ V with any SV­
module M, construct a projective (in fact, free) resolution P. of M
with Pi = 0 for 2' > dim V. Deduce that for any SV-module N and
any 2' > dim V, one has
Tor?‘/(M, N) = ExtgV(M, N) = 0
(the Hilbert syzygies theorem).
(v) Compute Extfgv (k, k) and Torggv (k, k).
Chapter 9

Structure of finite
dimensional algebras

In this chapter we return to studying the structure of finite dimen­

sional algebras. Throughout the chapter, we work over an alge­
braically closed field k (of any characteristic).

9.1. Lifting of idempotents

Let A be a ring, and let I C A be a nilpotent ideal.
Proposition 9.1.1. Let e0 6 A/I be an idempotent, i.e., eg = e0.
There exists an idempotent e E A which is a lift of eg (i.e., it projects
to e0 under the reduction modulo I This idempotent is unique up to
conjugation by an element of 1 + I.

Proof. Let us first establish the statement in the case when I 2 = 0.

Note that in this case I is a left and right module over A / I . Let e... be
any lift of eg to A. Then ef — e... = a E I, and ega = (L60. We look for
e in the form e = e... +b, b E I. The equation for b is e0b+be0 — b = a.
Set b = (2e0 —— 1)a. Then
eob + beg — b = 2e0a — (2e0 — 1)a = a,
so e is an idempotent. To classify other solutions, set e’ = e + c.
For e’ to be an idempotent, we must have ec + ce — c = O. This

209
210 9. Structure of finite dimensional algebras
is equivalent to saying that ece = O and (1 —— e)c(1 —— e) = 0, so
c = ec(1—e)+(1——e)ce = [e, [e, Hence e’ = (1+[c, e])e(1+[c, e])“1.
Now, in the general case, we prove by induction in k that there
exists a lift ek of eg to A/ I k+1, and it is unique up to conjugation
by an element of 1 + I k (this is sufficient as I is nilpotent). Assume
it is true for k = m — 1, and let us prove it for k = m. So we have
an idempotent em_1 E A/Im, and we have to lift it to A/Im+1. But
(I"‘)2 = O in A/I"‘+1, so we are done. [3
Definition 9.1.2. A complete system of orthogonal idempo­
tents in a unital algebra B is a collection of elements e1, . . . , en E B
such that e,ej = 5,]-e, and 22:1 e, = 1.
Corollary 9.1.3. Let em, . . . , 60m be a complete system of orthogonal

61). D
idempotents in A/I. Then there exists a complete system of orthog­
onal idempotents e1, . . . , em (e,ej = 5,]-e,, Ze, = 1) in A which lifts
601, . . .,60m.

Proof. The proof is by induction in m. For m = 2 this follows from

Proposition 9.1.1. For m > 2, we lift em to e1 using Proposition 9.1.1
and then apply the induction assumption to the algebra (1 — e1)A(1 —

=
9.2. Pro jective covers
Obviously, every finitely generated projective module over a finite
dimensional algebra A is a direct sum of indecomposable projective
modules, so to understand finitely generated projective modules over
A, it suffices to classify indecomposable projective modules.
Let A be a finite dimensional algebra with simple modules
M1, . . . , Mn.
Theorem 9.2.1. For each i = 1,. ..,n there exists a unique in­
decomposable finitely generated projective module P, such that

() Any indecomposable finitely generated projective module over

A is isomorphic to P, for some i.
9.3. The Cartan matrix of a finite dimensional algebra 211
Proof. Recall that A/ Rad(A) = $221 End(M,) and that Rad(A)
is a nilpotent ideal. Pick a basis of M,, and let egj = E3"-J-, the
rank 1 projectors projecting to the basis vectors of this basis (j =
1, . . . ,dim Then egj are orthogonal idempotents in A/ Rad(A .
So by Corollary 9.1.3 we can lift them to orthogonal idempotents
6,3‘ in NOW define P,'j = A6,)". Then A = P,'j,
so P,j are projective. Also, we have Hom(P,j,M;,) = e,]-Mk, so
dim Hom(P,j,M;,) = 5,;,. Finally, P,j is independent of j up to an
isomorphism, as e,, for fixed i are conjugate under AX by Proposition
9.1.1; thus we will denote P,j by P,.
We claim that P, is indecomposable. Indeed, if P, = Q1 69 Q2,
then Hom(Q,, M3) = O for all j either for l = 1 or for l = 2, so either
Q1 =- 0 01‘ Q2 = 0.
Also, there can be no other indecomposable finitely generated
projective modules, since any indecomposable projective module has
to occur in the decomposition of A. The theorem is proved. D
Definition 9.2.2. The projective module P, is called the projective
cover of M,.
Proposition 9.2.3. Let N be any finite dimensional A-module. Then
one has dim HomA(P,, N) = [N : M,], the multiplicity of occurrence
on M, in the Jordan-Holder series of N.

Proof. If N = M,, the statement is clear. Also, if

O——>N1——>N2——>N3——>O

is an exact sequence of A-modules, then the corresponding sequence

0 -5 HOH1A(P,, N1) -9 HomA(P,, N2) -9 HomA(P,, N3) -9 O
is exact, as P, is projective. This implies the statement. D

9.3. The Cartan matrix of a finite dimensional

algebra
Let A be a finite dimensional algebra with simple modules M,, i =
1,...,n, and projective covers P,. Let c,j := dimHomA(P,,Pj) =
[P7 2
212 9. Structure of finite dimensional algebras
Definition 9.3.1. The matrix C = (c,-j) is called the Cartan ma­
trix of A.

Obviously, the Cartan matrix of A is a matrix with nonnegative

entries, whose diagonal entries are positive.
Problem 9.3.2. Let A be the algebra over complex numbers gener­
ated by elements g, :1: with defining relations ga: = —:1:g, £132 = 0, g2 = 1.
Find the simple modules, the indecomposable projective modules, and
the Cartan matrix of A.

9.4. Homological dimension

Let A be a ring, and let M be a left A-module.
Definition 9.4.1. The projective dimension pd(M) of M is the
length of the shortest finite projective resolution of M (where we agree
that the length of O ——> Pd ——> ——> P0 is d). If a finite projective
resolution of M does not exist, then the projective dimension of M
is infinite.

For instance, the projective modules are the modules of projective

dimension 0.

Problem 9.4.2. Show. that pd(M) 3 d if and only if for any left
A-module N, one has Ext’(M, N) = O for i > d.
Hint: To prove the “if” part, use induction in d and the long
exact sequence of Ext groups in Problem 8.2.6(v).
(ii) Let 0 ——> M ——> P ——> N ——> 0 be a nonsplit short exact
sequence, and assume that P is projective. Show that pd(N) =
pd(M) + 1.
(iii) Show that if pd(M) = d > O and P. is any projective res­
olution of M, then the kernel Kd of the map Pd_1 ——> Pd_2 in this
resolution is projective (where we agree that P_1 = M Thus, by
replacing Pd with Kd and all terms to the left of Pd by zero, we get
a projective resolution of M of length d. Deduce that if A and M
are finite dimensional, then there is a finite resolution P. of M with
finite dimensional P,-.
9.5. Blocks 213
Definition 9.4.3. One says that a ring A has left (respectively, right)
homological dimension 3 d if every left (respectively, right) A­
module M has projective dimension 3 d. The homological dimension
of A is exactly d if it is 3 d but not 3 d — 1. If such a d does not
exist, one says that A has infinite homological dimension.

Example 9.4.4. By the Hilbert syzygies theorem (see Problem

8.2.10(iv)), the homological dimension of the polynomial algebra
k[:1:1,...,:1:,,] is n.
Problem 9.4.5. Show that if a finite dimensional algebra A has
finite homological dimension d and if C is the Cartan matrix of A,
then det(C) = :l:1.
(ii) What is the homological dimension of t", n > 1? Of the
algebra of Problem 9.3.2?
Problem 9.4.6. Show that the path algebra PQ of any quiver Q
with at least one edge has homological dimension 1. In particular,
the homological dimension of the free algebra k(:1:1, . . . ,:1:,,) is 1 (for
n 2 1).
(ii) Let Q be a finite oriented graph without oriented cycles. Find
the Cartan matrix of its path algebra PQ.
9.5. Blocks
Let A be a finite dimensional algebra over an algebraically closed field
k, and let C denote the category of finite dimensional A-modules.
Definition 9.5.1. Two simple finite dimensional A-modules X ,Y
are said to be linked if there is a chain X = M0, M1, . . . , Mn = Y
such that for each 13 = O,...,n — 1 either Ext1(M,-,M,~+1) 7E O or
Ext1(M,-+1, M,-) 75 0 (or both).

Here we agree that X is linked to itself (by a chain of length 0).

This linking relation is clearly an equivalence relation, so it defines
a splitting of the set S of isomorphism classes of simple A-modules
into equivalence classes Sk, k E B. The kth block Ck of C is, by
definition, the category of all objects M of C such that all simple
modules occurring in the Jordan-Holder series of M are in Sk.
214 9. Structure of finite dimensional algebras
Example 9.5.2. If A is semisimple, each block is equivalent to the
category of vector spaces (and thus has only one simple object). If A
is a commutative local artinian algebra, then there is only one block.
Also there is just one block for the algebra of Problem 9.3.2.
Problem 9.5.3. Show that there is a natural bijection between
blocks of C and indecomposable central idempotents ek of A (i.e., ones
that cannot be nontrivially split in a sum of two central idempotents),
such that Ck is the category of finite dimensional ekA-modules.
(ii) Show that any indecomposable object of C lies in some Ck and
that Hom(M, N) = 0 if M E Ck, N E C1, is 7é l. Thus, C = ®kEB Ck.
(iii) Determine the blocks in the category of left A-modules for
A = k[S3], where k is of characteristic 2.

9.6. Finite abelian categories

Let C be a k-linear abelian category in which any object has finite
length (i.e., a finite filtration whose successive quotients are simple).
Assume that for every simple object X of C, one has End(X) = k.
Note that in such a category, Hom spaces are finite dimensional (check
it!).
Definition 9.6.1. Let us say that C is finite if it has enough pro­
jectives and finitely many simple objects (up to an isomorphism).
For example, as we have shown, the category of finite dimensional
modules over a finite dimensional algebra is finite. Below we will see
that the converse is also true.
Definition 9.6.2. An object P of an abelian category C is said to be
a projective generator (or progenerator) if it is projective and
every object is a quotient of a multiple of P.
For example, in the category of modules over a ring A, the free
module A is a projective generator.
Exercise 9.6.3. Show that in a finite abelian category, P is a pro­
jective generator if and only if for every simple object L, one has
Hom(P, L) 75 O. Deduce that any finite abelian category has a pro­
jective generator.
9.7. Morita equivalence 215
Now let C be a finite abelian category with a projective generator
P. Let B = End(P)°p (a finite dimensional algebra acting on the
right on P). Let B — fmod denote the category of finite dimensional
left B-modules. Consider the functor F : C ——> B — fmod given by the
formula F(M) = Hom(P,
Theorem 9.6.4. The functor F is an equivalence of categories. Thus,
any finite abelian category over a field k is equivalent to the category
of modules over a finite dimensional k-algebra.

The proof of Theorem 9.6.4 is contained in the following problem.

Problem 9.6.5. Let G : B — fmod -> C be the functor defined by
G(X) := P (83 X, where P <83 X is the cokernel of the morphism
zp : P®B®X -> P®X given by w = ap®Id-—Id®aX (where
a p : P <8) B ——> P, a X : B <8) X ——> X are the morphisms representing
the actions of B on P and X
(i) Show that F o G § Id. That is, for every X E B — fmod, show
that X is naturally isomorphic to Hom(P,P <83 X (For this you
should only need that P is a nonzero projective object.)
(ii) For any X E C, construct a natural morphism
§:P®B Hom(P,X) ——> X,
and show that it is surjective.
(iii) Show that G 0 F E Id. To this end, consider the short exact
sequence
O—>K—>P®BHom(P,X)—->X—>0,
where the third map is 5. Apply the functor F to this sequence
and use to conclude that K = O and hence 5 is an isomorphism.
Conclude that the functors G and F are quasi-inverse to each other
and hence are equivalences of categories.

9.7 . Morita equivalence

In this section we will discuss the theory of Morita equivalence of
finite dimensional algebras. We note that this theory extends with
appropiate changes to the infinite dimensional case.
216 9. Structure of finite dimensional algebras
Theorem 9.6.4 shows that a finite abelian category C can be iden­
tified with the category of finite dimensional modules over a finite
dimensional algebra B, once we choose a projective generator P of
C. However, the projective generator is not unique. Namely, let
P1, . . . ,Pm be the indecomposable projective objects of C (they make
sense by Theorem 9.6.4). Then all the projective generators of C are
the objects of the form P := n,;P,;, where n, 2 1 (check it!).
Defining Bn = Bn(C) := End(Pn)°p, we see that the algebras Bn are
all the finite dimensional algebras whose category of finite dimensional
modules is equivalent to C.
Definition 9.7.1. Finite dimensional algebras A and B are said to be
Morita equivalent if the abelian categories A — fmod and B — fmod
are equivalent.

Thus, we obtain that Morita equivalence classes of finite dimen­

sional algebras are the collections of the form {Bn (C), n E Nm}.
Note that the dimension of Bn is Zcijninj, where (c,-9-) is the
Cartan matrix of C. In particular, the algebra B = Bn with the
smallest possible dimension, 2 c,-3-, corresponds to the case when all
ni = 1, i.e., to the projective generator P = P,;. It is easy to see
that this algebra is the only one of the Bn for which Bn/Rad(Bn) =
@z- Matm is commutative.
Definition 9.7.2. A finite dimensional algebra B with commutative
B / Rad(B) is called basic.
So we have the following corollary of Theorem 9.6.4.
Corollary 9.7.3. Any finite abelian category C is equivalent to
the category of finite dimensional modules over a unique basic algebra
B = B(C).
Any finite dimensional algebra A is Morita equivalent to a
unique basic algebra B = B A, such that dim B A 3 dim A.
References for historical
interludes

[1] Adelmann, Clemens, and Eberhard H.—A. Gerbracht. “Letters from

William Burnside to Robert Fricke: Automorphic Functions, and the
Emergence of the Burnside Problem”. Archive for History of Exact
Sciences 63:1 (2009): 33-50.
[2] Arnold, Vladimir. “Matematika s chelovecheskim litsom” (Mathe­
matics with a Human Face). Priroda, no. 3 (1988): 117-119.
[3] Atiyah, Michael. “Hermann Weyl, 1885-1955: A Biographical Mem­
oir”. In Biographical Memoirs, vol. 82 (Washington, DC: National
Academy of Sciences, 2003), 321-332.
[4] Baumslag, Gilbert, comp. Reviews on Infinite Groups: As Printed
in Mathematical Reviews 1940 through 1970, vol. 2 (Providence, RI:
American Mathematical Society, 1974).
[5] Beaulieu, Liliane. “Bourbaki’s Art of Memory”. Osiris, 2nd series, 14
(1999): 219-251.
[6] Bass, Hyman, et al. “Samuel Eilenberg (1913-1998)”. Notices of the
American Mathematical Society 45, no. 10 (1998): 1344-1352.
[7] Bass, Hyman. “Eilenberg, Samuel”. New Dictionary of Scientific Bi­
ography, vol. 20 (Detroit, MI: Charles Scribner’s Sons, 2008), pp.
360-363.
[8] Burnside, William. “On the Theory of Groups of Finite Order”. Proc.
London Math. Soc. (1909): 1-7.
[9] Corry, Leo. Modern Algebra and the Rise of Mathematical Structures,
2nd ed. (Basel: Birkhauser, 2004).

217
218 References for historical interludes

[10] Crowe, Michael J. A History of Vector Analysis: The Evolution of

the Idea of a Vectorial System (New York: Dover, 1985).
[11] Curtis, Charles W. Pioneers of Representation Theory: Frobenius,
Burnside, Schur, and Braver, History of Mathematics, vol. 15 (Prov­
idence, RI: American Mathematical Society, 1999).
[12] Eilenberg, Samuel, and Saunders Mac Lane. “General Theory of Nat­
ural Equivalences”. Transactions of the American Mathematical So­
ciety 58 (1945): 231-294.
[13] Everett, Martin G., A. J. S. Mann, and Kathy Young. “Notes of
Burnside’s life”. In The Collected Papers of William Burnside, eds.
Everett, Mann, and Young, vol. 1 (New York: Oxford University
Press, 2004), pp. 89-110.
[14] Feit, Walter. “Foreword”. In The Collected Papers of William Burn­
side, vol. 1, pp. 3-6.
[15] Forsyth, A. R. “William Burnside”. J. London Math. Soc. s1—3:1 (Jan­
uary 1928): 64-80.
[16] Heudenthal, Hans. “Lie, Marius Sophus”. In Dictionary of Scientific
Biography, vol. 8 (New York: Scribner, 1973), pp. 323-327.
[17] Graves, Robert P. Life of Sir William Rowan Hamilton, 3 vols.
(Dublin: Dublin University Press, 1882-1889).
[18] Hankins, Thomas L. “Hamilton, William Rowan”. Complete Dictio­
nary of Scientific Biography, ed. Charles C. Gillespie, vol. 6 (Detroit:
Charles Scribner’s Sons, 2008), pp. 85-93 (Gale Virtual Reference
Library).
[19] Hankins, Thomas L. Sir William Rowan Hamilton (Baltimore, MD:
Johns Hopkins University Press, 1980).
[20] Hankins, Thomas L. “Triplets and Triads: Sir William Rowan Hamil­
ton on the Metaphysics of Mathematics”. Isis 68 (1977): 175-193.
[21] Haubrich, Ralf. “Pkobenius, Schur, and the Berlin algebraic tradi­
tion”. In Mathematics in Berlin, eds. Heinrich Begehr et al. (Berlin:
Birkhauser, 1998), pp. 83-96.
[22] Hawkins, Thomas. Emergence of the Theory of Lie Groups: An Essay
in the History of Mathematics, 1869-1926 (Berlin: Springer, 2000).
[23] Hawkins, Thomas. “Hypercomplex Numbers, Lie Groups, and the
Creation of Group Representation Theory”. Archive for History of
Exact Sciences 8 (1971): 243-287.
[24] Hawkins, Thomas. “New light on Pkobenius’ creation of the theory of
group characters”. Archive for History of Exact Sciences 12:3 (1974):
217-243.
[25] Hawkins, Thomas. “The origins of the theory of group characters”.
Archive for History of Exact Sciences 7:2 (1971): 142-170.
References for historical interludes 219
[26] Hermann, Robert. “Preface”. In Lie Groups: History, Frontiers and
Applications, vol. 3 (Brookline, MA: Math Sci Press, 1976), pp. iii—iv.
[27] Kimberling, Clark. “Dedekind letters” ;
http : / /faculty.evansville .edu/ck6/bstud/dedek. html.
[28] Kromer, Ralf. Tool and Object: A History and Philosophy of Category
Theory (Basel: Birkhauser, 2007).
[29] Lam, T. Y. “Representations of finite groups: A hundred years: Part
I”. Notices of the AMS 45:3 (1998): 361-372.
[30] Lam, T. Y. “Representations of Finite Groups: A Hundred Years:
Part II”. Notices of the AMS 45:4 (1998): 465-474.
[31] Landry, Elaine, and Jean—Pierre Marquis. “Categories in Context:
Historical, Foundational, and Philosophical”. Philosophia Mathemat­
ica 13 (2005): 1-43.
[32] Lawvere, F. W., and Colin McLarty. “Mac Lane, Saunders”. New Dic­
tionary of Scientific Biography, vol. 23 (Detroit, MI: Charles Scrib­
ner’s Sons, 2008), pp. 1-5.
[33] Mac Lane, Saunders. A Mathematical Autobiography (Wellesley, MA:
A.K. Peters, 2005).
[34] Mac Lane, Saunders. Mathematics, Form and Function (New York:
Springer, 1986).
[35] Mac Lane, Saunders. “Concepts and Categories in Perspective”. In A
Century of Mathematics in America, ed. Peter Duren, Part I (Provi­
dence, RI: American Mathematical Society, 1988), pp. 323-365.
[36] Mac Lane, Saunders. “Mathematics at the University of Chicago: A
Brief History”. In A Century of Mathematics in America, ed. Pe­
ter Duren, Part II (Providence, RI: American Mathematical Society,
1989), pp. 127-154.
[37] Mac Lane, Saunders. “The PNAS way back then”. Proc. Natl. Acad.
Sci. USA 94 (June 1997): 5983-5985.
[38] Marquis, Jean-Pierre. From a Geometrical Point of View: A Study of
the History and Philosophy of Category Theory (New York: Springer,
2009).
[39] McLarty, Colin. “The Last Mathematician from Hilbert’s Gottingenz
Saunders Mac Lane as Philosopher of Mathematics”. British Journal
for the Philosophy of Science 58:1 (2007): 1-36.
[40] McLarty, Colin. “Saunders Mac Lane (1909-2005): His Mathematical
Life and Philosophical Works”. Philosophia Mathematica 13 (2005):
237-251.
[41] Neumann, Peter M. “The Context of Burnside’s Contributions to
Group Theory”. In The Collected Papers of William Burnside, vol.
1, pp. 15-54.
220 References for historical interludes

[42] Pesic, Peter. “Introduction”. In Hermann Weyl, Mind and Nature:

Selected Writings on Philosophy, Mathematics, and Physics (Prince­
ton, NJ: Princeton University Press, 2009), pp. 1-19.
[43] Pickering, Andrew. “Concepts and the Mangle of Practice: Con­
structing Quaternions”. In 18 Unconventional Essays on the Nature
of Mathematics, ed. Reuben Hersh (New York: Springer, 2006), pp.
250-288.
[44] Pycior, Helena M. “The Role of Sir William Rowan Hamilton in the
Development of British Modern Algebra”. Ph. D. dissertation, Cor­
nell University, 1976.
[45] Reid, Constance. Hilbert—C'ourant (New York: Springer, 1986).
[46] Rowe, David E. “Hermann Weyl, the Reluctant Revolutionary”.
Mathematical Intelligencer 25:1 (2003): 61-70.
[47] Rowe, David E. “Lie, Marius Sophus”. In New Dictionary of Scientific
Biography, vol. 22 (Detroit, MI: Charles Scribner’s Sons, 2008), pp.
307-310.
[48] Rowe, David E. “Three letters from Sophus Lie to Felix Klein on
Parisian mathematics during the early 1880s”. Mathematical Intelli­
gencer 7:3 (1985): 74-77.
[49] Sigurdsson, Skuli. “Physics, Life, and Contingency: Born, Schr6­
dinger and Weyl in Exile”. In Forced Migration and Scientific Change:
Emigre German—Speaking Scientists and Scholars after 1933, eds.
Mitchell G. Ash and Alfons Sollner (Cambridge, UK: Cambridge Uni­
versity Press, 1996), pp. 48-70.
[50] Sigurdsson, Skuli. “On the Road”. In Positioning the History of Sci­
ence, eds. Kostas Gavroglu and Jiirgen Renn (Dordrecht: Springer,
2007), pp. 149-157.
[51] Sigurdsson, Skuli. “Journeys in Spacetime”. In Hermann Weyl’s
Raum—Zeit-Materie and a General Introduction to His Scientific
Work, ed. Erhard Scholz (Basel: Birkhauser Verlag, 2001), pp. 15-47.
[52] Sigurdsson, Skuli. “Hermann Weyl, Mathematics and Physics, 1900­
1927”. PhD dissertation, Harvard University, 1991.
[53] Sigurdsson, Skuli. “Unification, Geometry and Ambivalence: Hilbert,
Weyl, and the Gottingen Community”. In Trends in the Histori­
ography of Science, eds. Kostas Gavroglu, Jean Christianidis, and
Efthymios Nicolaidis (Dordrecht: Kluwer, 1994), pp. 355-367.
[54] Solomon, Ronald M. “Burnside and Finite Simple Groups”. In The
Collected Papers of William Burnside, vol. 1, pp. 45-54.
[55] Sternberg, Shlomo. Group Theory and Physics (Cambridge, UK:
Cambridge University Press, 1995).
References for historical interludes 221
[56] Stubhaug, Arild. The Mathematician Sophus Lie: It Was the Au­
dacity of My Thinking, transl. Richard H. Daly (Berlin: Springer,
2002)
[57] Weyl, Hermann. The Classical Groups, Their Invariants and Repre­
sentations (Princeton, NJ: Princeton University Press, 1946).
[58] Weyl, Hermann. “David Hilbert. 1862-1943”. Obituary Notices of
Fellows of the Royal Society 4:13 (1944): 547-553.
[59] Weyl, Hermann. “The Open World: Three Lectures on the Metaphys­
ical Implications of Science”. (1932). In Mind and Nature: Selected
Writings on Philosophy, Mathematics, and Physics (Princeton, NJ:
Princeton University Press, 2009), pp. 34-82.
[60] Weyl, Hermann. The Theory of Groups and Quantum Mechanics, 2nd
ed. (New York: Dover, 1950).
[61] Winterbourne, Anthony T. “Algebra and Pure Time: Hami1ton’s
Affinity with Kant”. Historia Mathematica 9 (1982): 195-200.
Mathematical references

[BGP] J. Bernstein, I. Gelfand, V. Ponomarev, Coxeter functors and

Gabriel’s theorem, Russian Math. Surveys 28 (1973), no. 2, 17-32.
[Cu] C. Curtis, Pioneers of Representation Theory: Frobenius, Burnside,
Schur, and Brauer, Amer. Math. Soc., 1999.
[CR] C. Curtis and I. Reiner, Representation Theory of Finite Groups and
Associative Algebras, Amer. Math. Soc., 2006.
[FH] W. Fulton and J. Harris, Representation Theory, A First Course,
Springer, New York, 1991.
[Fr] P. J. Freyd, Abelian Categories, an Introduction to the Theory of
Functors, Harper and Row, 1964.
[Ki] A. A. Kirillov Jr., An Introduction to Lie Groups and Lie Algebras,
Cambridge University Press, 2008.
[MCL] S. Mac Lane, Categories for a Working Mathematician: 2nd ed.,
Graduate Texts in Mathematics 5, Springer, 1998.
[3] J .-P. Serre, Linear Representations of Finite Groups, Springer-Verlag,
1997.
[W] C. Weibel, An Introduction to Homological Algebra, Cambridge
Univ. Press, 1994.

223
Index

Abelian category, 186 Central function, 61

Acyclic complex, 188 Character group, 62
Additive functor, 190 Character of a group
Adjacency matrix, 146 representation, 61
Adjoint functors, 184 Character of a representation, 48
Ado’s theorem, 22 Character table, 71
Affine Dynkin diagram, 148 Class function, 61
Algebra of dual numbers, 55 Clifford algebra, 55
Algebra of endomorphisms of a Coboundary, 188
vector space, 9 Cocycle, 54, 188
Algebra of polynomials, 9 Cohomology, 187
Algebraic conjugate, 95 Cohomology class, 188
Algebraic integer, 93 Commutant of a Lie algebra, 39
Algebraic number, 93 Commutative algebra, 9
Algebraic representation, 130 Companion matrix, 94
Algebras defined by generators and Complete system of orthogonal
relations, 16 idempotents, 210
Artin’s theorem, 141 Completely reducible
Associative algebra, 5, 9 representation, 41
Complex, 187
Basic algebra, 216 Complex conjugate representation,
Block, 213 69
Burnside’s theorem, 98 Compression modulus, 86
Connecting homomorphism, 189
Cartan matrix, 160 Content of a Young diagram, 118
Cartan matrix of a finite Coxeter element, 169
dimensional algebra, 212 Coxeter functor, 164
Casimir operator, 38 Cyclic quiver, 173
Category, 178 Cyclic representation, 16
Central character, 14 Cyclic vector, 16

225
226 Index

De Rham complex, 187 Highest weight, 130

Defining relations, 16 Hilbert syzygies theorem, 207
Deformation tensor, 85 Hilbert’s third problem, 36
Dehn invariant, 36 Homological dimension, 213
Density theorem, 43 Homomorphism of algebras, 9
Differential, 187 Homomorphism of Lie algebras, 23
Dimension vector, 163 Homomorphism of representations,
Direct sum of representations, 5, 11 11
Double centralizer theorem, 119 Hook, 119
Dual group, 62 Hook length formula, 119
Dual representation of a Lie
algebra, 37 Ideal, 15
Dual representation of an Ideal generated by a subset, 15
associative algebra, 44 Idempotent, 67
Dynkin diagram, 8, 146 Indecomposable representation, 6,
11
Einstein summation, 32 Induced representation, 105
Enriched category, 179 Intertwining operator, 11
Equivalence of categories, 182 Irreducible representation
Exact functor, 191 (module), 6, 10
Exact sequence, 188 Isomorphism of representations, 11
Ext functor, 203
Extensions of representations, 53 Jordan—H6lder series, 51
Exterior algebra, 35 Jordan—H6lder theorem, 50
Exterior power, 33
Kostka numbers, 114
Faithful representation, 18 Koszul resolution, 207
Filtration, 45 Krull—Schmidt theorem, 51
Finite abelian category, 214 Kiinneth formula, 190
Flat module, 202
Formal deformation, 55 Left adjoint functor, 184
Left exact functor, 191
Free algebra, 9
Frobenius character formula, 115 Left ideal, 15
Frobenius determinant, 75 Left module, 5, 10
Frobenius divisibility theorem, 96 Length of a representation, 51
Frobenius formula, 106 Lexicographic ordering on
Frobenius reciprocity, 107
partitions, 112
Frobenius—Schur indicator, 92 Lie algebra, 22
Full subcategory, 179 Lie group, 24
Functor, 179 Lie subalgebra, 22
Lie’s theorem, 39
F‘unctorial morphism, 181
Linear abelian category (over a
Gabriel’s theorem, 163 field), 187
General linear group, 13 Linear algebraic group, 25
General linear Lie algebra, 22 Linear functor (over a field), 190
Group algebra, 9 Linked modules, 213
Group determinant, 75 Locally small category, 17 9
Long exact sequence of
Heisenberg group, 81 cohomology, 189
Index 227

Manifold, 25 Representation of a quiver, 7, 19

Maschke’s theorem, 59 Representation of complex type, 91
Maximal ideal, 15 Representation of quaternionic
McKay graph, 149 type, 91
Minimal polynomial, 95 Representation of real type, 91
Module of finite length, 53 Representations of the symmetric
Monoidal category, 179 group, 110
Morita equivalence, 215 Restriction of a representation, 105
Morita equivalent rings, 187 Right adjoint functor, 184
Morphism of complexes, 188 Right exact functor, 191
Right ideal, 15
Natural transformation, 181 Right module, 10
Root, 161
Orthogonality of characters, 65
Orthogonality of matrix elements, Schur functor, 180
70
Schur polynomials, 128
Schur’s lemma, 11
Partition, 110
Path algebra of a quiver, 20 Schur-Weyl duality, 119, 122
Polynomial representation, 130 Semidirect product, 142
Positive root, 162 Semisimple abelian category, 191
Progenerator, 214 Semisimple algebra, 48
Projective cover, 210 Semisimple representation, 41
Projective dimension of a module, Shearing modulus, 86
212 Short exact sequence, 188
Projective generator, 214 Simple algebra, 15
Projective module, 202, 203 Simple reflection, 163
Projective object, 203 Simple representation (module), 10
Projective resolution, 204 Simple root, 161
Pure tensor, 31 Sink, 164
Solvable group, 98
q—Weyl algebra, 17, 18 Solvable Lie algebra, 39
Quasi—inverse functor, 183 Source, 164
Quaternion algebra, 83 Specht module, 111
Quaternion group, 63 Split short exact sequence, 188
Quaternions, 83 Stress tensor, 85
Quiver, 7, 19 Subrepresentation, 5, 10
Quiver of finite type, 7, 148 Sum of squares formula, 60
Quotient algebra, 15 Sylvester criterion, 147
Quotient representation, 15 Symmetric algebra, 35
Symmetric power, 33
Radical, 46
Rational representation, 130 Tensor algebra, 35
Reflection functor, 164 Tensor product of linear maps, 32
Regular representation, 10 Tensor product of representations
Representable functor, 183 of a Lie algebra, 37
Representation, 5, 10 Tensor product of vector spaces, 31
Representation of a group, 59 Tensors of type (m, n), 31
Representation of a Lie algebra, 24 Tor functor, 203
228 Index

Two-sided ideal, 15
Unit, 9
Unit morphism, 178
Unitary representation of a group,
69
Universal enveloping algebra of a
Lie algebra, 24, 35
Virtual representation, 105
Weyl algebra, 17
Weyl group, 163
Young diagram, 110
Young projector, 111
Young tableau, 110
Zariski dense subset, 145
Titles in This Series
59 Pavel Etingof, Oleg Golberg, Sebastian Hensel, Tiankai Liu, Alex
Schwendner, Dmitry Vaintrob, and Elena Yudovina, with
historical interludes by Slava Gerovitch, Introduction to
representation theory, 2011
58 Alvaro Lozano-Robledo, Elliptic curves, modular forms, and their
L-functions, 2011
57 Charles M. Grinstead, William P. Peterson, and J. Laurie Snell,
Probability Tales, 2011
56 Julia Garibaldi, Alex Iosevich, and Steven Senger, The Erd6s
distance problem, 2011
55 Gregory F. Lawler, Random walk and the heat equation, 2010
54 Alex Kasman, Glimpses of soliton theory: The algebra and geometry of
nonlinear PDES, 2010
53 J ii"1' Matousek, Thirty-three miniatures: Mathematical and algorithmic
applications of linear algebra, 2010
52 Yakov Pesin and Vaughn Climenhaga, Lectures on fractal geometry
and dynamical systems, 2009
51 Richard S. Palais and Robert A. Palais, Differential equations,
mechanics, and computation, 2009
50 Mike Mesterton-Gibbons, A primer on the calculus of variations and
optimal control theory, 2009
49 Francis Bonahon, Low-dimensional geometry: From euclidean surfaces
to hyperbolic knots, 2009
48 John Franks, A (terse) introduction to Lebesgue integration, 2009
47 L. D. Faddeev and O. A. Yakubovskii’, Lectures on quantum
mechanics for mathematics students, 2009
46 Anatole Katok and Vaughn Climenhaga, Lectures on surfaces:
(Almost) everything you wanted to know about them, 2008
45 Harold M. Edwards, Higher arithmetic: An algorithmic introduction to
number theory, 2008
44 Yitzhak Katznelson and Yonatan R. Katznelson, A (terse)
introduction to linear algebra, 2008
43 Ilka Agricola and Thomas Friedrich, Elementary geometry, 2008
42 C. E. Silva, Invitation to ergodic theory, 2007
41 Gary L. Mullen and Carl Mummert, Finite fields and applications,
2007
40 Deguang Han, Keri Kornelson, David Larson, and Eric Weber,
Frames for undergraduates, 2007
39 Alex Iosevich, A view from the top: Analysis, combinatorics and number
theory, 2007
38 B. Fristedt, N. Jain, and N. Krylov, Filtering and prediction: A
primer, 2007
 TITLES IN THIS SERIES

37 Svetlana Katok, p-adic analysis compared with real, 2007

36 Mara D. Neusel, Invariant theory, 2007
35 J6rg Bewersdorff, Galois theory for beginners: A historical perspective,
2006
34 Bruce C. Berndt, Number theory in the spirit of Ramanujan, 2006
33 Rekha R. Thomas, Lectures in geometric combinatorics, 2006
32 Sheldon Katz, Enumerative geometry and string theory, 2006
31 John McCleary, A first course in topology: Continuity and dimension,
2006
30 Serge 'I‘abachnikov, Geometry and billiards, 2005
29 Kristopher Tapp, Matrix groups for undergraduates, 2005
28 Emmanuel Lesigne, Heads or tails: An introduction to limit theorems in
probability, 2005
27 Reinhard Illner, C. Sean Bohun, Samantha McCollum, and Thea
van Roode, Mathematical modelling: A case studies approach, 2005
26 Robert Hardt, Editor, Six themes on variation, 2004
25 S. V. Duzhin and B. D. Chebotarevsky, Transformation groups for
beginners, 2004
24 Bruce M. Landman and Aaron Robertson, Ramsey theory on the
integers, 2004
23 S. K. Lando, Lectures on generating functions, 2003
22 Andreas Arvanitoyeorgos, An introduction to Lie groups and the
geometry of homogeneous spaces, 2003
21 W. J. Kaczor and M. T. Nowak, Problems in mathematical analysis
III: Integration, 2003
20 Klaus Hulek, Elementary algebraic geometry, 2003
19 A. Shen and N. K. Vereshchagin, Computable functions, 2003
18 V. V. Yaschenko, Editor, Cryptography: An introduction, 2002
17 A. Shen and N. K. Vereshchagin, Basic set theory, 2002
16 Wolfgang Kiihnel, Differential geometry: curves — surfaces — manifolds,
second edition, 2006
15 Gerd Fischer, Plane algebraic curves, 2001
14 V. A. Vassiliev, Introduction to topology, 2001
13 Frederick J. Almgren, J r., Plateau’s problem: An invitation to varifold
geometry, 2001
12 W. J . Kaczor and M. T. Nowak, Problems in mathematical analysis
II: Continuity and differentiation, 2001

For a complete list of titles in this series, visit the

AMS Bookstore at www.ams.org/bookstore/.