Algebra 4 Lie Algebras Chevalley Groups And

Their Representations 1st Edition Ramji Lal

download

https://ebookbell.com/product/algebra-4-lie-algebras-chevalley-
groups-and-their-representations-1st-edition-ramji-lal-51159100

Explore and download more ebooks at ebookbell.com

Here are some recommended products that we believe you will be
interested in. You can click the link to download.

Lie Groups And Lie Algebras Part Ii Chapters 46 Nicolas Bourbaki

https://ebookbell.com/product/lie-groups-and-lie-algebras-part-ii-
chapters-46-nicolas-bourbaki-5432730

Complex Semisimple Lie Algebras Jeanpierre Serre

https://ebookbell.com/product/complex-semisimple-lie-algebras-
jeanpierre-serre-47365128

Complex Semisimple Lie Algebras Jeanpierre Serre

https://ebookbell.com/product/complex-semisimple-lie-algebras-
jeanpierre-serre-2148846

Introduction To Lie Algebras Karin Erdmann Mark J Wildon Erdmann

https://ebookbell.com/product/introduction-to-lie-algebras-karin-
erdmann-mark-j-wildon-erdmann-22382416
Representations Of Lie Algebras An Introduction Through Gln Anthony
Henderson

https://ebookbell.com/product/representations-of-lie-algebras-an-
introduction-through-gln-anthony-henderson-34257482

Introduction To Lie Algebras K Erdmann Mark J Wildon

https://ebookbell.com/product/introduction-to-lie-algebras-k-erdmann-
mark-j-wildon-42998074

Representations Of Lie Algebras And Partial Differential Equations 1st

Xiaoping Xu

https://ebookbell.com/product/representations-of-lie-algebras-and-
partial-differential-equations-1st-xiaoping-xu-6761094

Representations Of Lie Algebras Quantum Groups And Related Topics

Naihuan Jing

https://ebookbell.com/product/representations-of-lie-algebras-quantum-
groups-and-related-topics-naihuan-jing-7219144

Infinitedimensional Lie Algebras Minoru Wakimoto

https://ebookbell.com/product/infinitedimensional-lie-algebras-minoru-
wakimoto-1526684
Infosys Science Foundation Series in Mathematical Sciences

Ramji Lal

Algebra 4
Lie Algebras, Chevalley Groups, and
their Representations
Infosys Science Foundation Series

Infosys Science Foundation Series in Mathematical

Sciences

Series Editors
Irene Fonseca, Carnegie Mellon University, Pittsburgh, PA, USA
Gopal Prasad, University of Michigan, Ann Arbor, USA

Editorial Board
Manindra Agrawal, Indian Institute of Technology Kanpur, Kanpur, India
Weinan E, Princeton University, Princeton, USA
Chandrashekhar Khare, University of California, Los Angeles, USA
Mahan Mj, Tata Institute of Fundamental Research, Mumbai, India
Ritabrata Munshi, Tata Institute of Fundamental Research, Mumbai, India
S. R. S Varadhan, New York University, New York, USA
The Infosys Science Foundation Series in Mathematical Sciences, a Scopus-indexed
book series, is a sub-series of the Infosys Science Foundation Series. This sub-series
focuses on high-quality content in the domain of mathematical sciences and various
disciplines of mathematics, statistics, bio-mathematics, financial mathematics,
applied mathematics, operations research, applied statistics and computer science.
All content published in the sub-series are written, edited, or vetted by the laureates
or jury members of the Infosys Prize. With this series, Springer and the Infosys
Science Foundation hope to provide readers with monographs, handbooks,
professional books and textbooks of the highest academic quality on current topics
in relevant disciplines. Literature in this sub-series will appeal to a wide audience of
researchers, students, educators, and professionals across mathematics, applied
mathematics, statistics and computer science disciplines.

More information about this subseries at http://www.springer.com/series/13817

Ramji Lal

Algebra 4
Lie Algebras, Chevalley Groups, and their
Representations

123
Ramji Lal
University of Allahabad
Prayagraj, Uttar Pradesh, India

ISSN 2363-6149 ISSN 2363-6157 (electronic)

Infosys Science Foundation Series
ISSN 2364-4036 ISSN 2364-4044 (electronic)
Infosys Science Foundation Series in Mathematical Sciences
ISBN 978-981-16-0474-4 ISBN 978-981-16-0475-1 (eBook)
https://doi.org/10.1007/978-981-16-0475-1

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature
Singapore Pte Ltd. 2021
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with regard
to jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Dedicated to the memory of
my younger sister
(Late) Smt Pushpa (Malti),
who left us at the age of 30.
Preface

The present volume, Algebra 4, in this series of books on Algebra, centers around
the study of Lie algebras, Chevalley groups, and their representation theory. Lie
groups and Lie algebras are very intrinsically related. The origin of Lie groups and
Lie algebras lies in the study of geometric spaces with the very crucial observation
that a geometric space is determined by the group of its continuous symmetries. Lie
groups and Lie algebras play a very fundamental role in Physics also.
The main concerns in the book are the following:
1. The structure theory and the classification of semi-simple Lie algebras over C
through root space decomposition, root systems, and Dynkin diagrams.
2. The representation theory of semi-simple Lie algebras including the theorem of
Harish-Chandra and the theorems of Ado and Iwasava.
3. Chevalley groups including the twisted finite simple groups of Lie types.
4. The representation theory of Chevalley groups including the Steinberg charac-
ters, Principal and Discrete series representations, and an introduction to the
Deligne–Lusztig characters.
The book can act as a text for graduate and advanced graduate students
specializing in the field.
There is no prerequisite essential for the book except for some basics in algebra
(as in Algebra 1 and Algebra 2) together with some amount of calculus and
topology. An attempt to follow the logical ordering has been made throughout the
book.
My teacher: (Late) Prof. B. L. Sharma; my colleagues at the University of
Allahabad; my friends: Prof. Satyadeo, Prof. S. S. Khare, Prof. H. K. Mukherji, and
Dr. H. S. Tripathi; my students: Prof. R. P. Shukla, Prof. Shivdatt, Dr. Brajesh
Kumar Sharma, Mr. Swapnil Srivastava, Dr. Akhilesh Yadav, Dr. Vivek Jain,
Dr. Vipul Kakkar, and Dr. Laxmikant; and above all the mathematics students of
Allahabad University had always been a motivating force for me to write a series of
books on various topics in Algebra. Without their continuous insistence, it would
have not come in the present form. I wish to express my warmest thanks to all
of them.

vii
viii Preface

Harish-Chandra Research Institute Allahabad has always been a great source

for me to learn more and more mathematics. I wish to express my deep sense
of appreciation and thanks to HRI for providing me with all the infrastructural
facilities to write these volumes.
Last but not least, I wish to express my thanks to my wife Veena Srivastava who
had always been helpful in this endeavor.
In spite of all the care, some mistakes and misprints might have crept and
escaped my attention. I shall be grateful to any such attention. Criticisms and
suggestions for the improvement of the book will be appreciated and gratefully
acknowledged.

Prayagraj, India Ramji Lal

September 2020
Contents

1 Lie Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Definitions and Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Universal Enveloping Algebras: PBW Theorem . . . . . . . . . . . . . . 22
1.3 Solvable and Nilpotent Lie Algebras . . . . . . . . . . . . . . . . . . . . . . 34
1.4 Semi-Simple Lie Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
1.5 Extensions of Lie Algebras and Co-homology . . . . . . . . . . . . . . . 59
2 Semi-Simple Lie Algebras and Root Systems . . . . . . . . . . . . . . . . . . 77
2.1 Root Space Decomposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
2.2 Root Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
2.3 Dynkin Diagram and the Classification of Root Systems . . . . . . . . 97
2.4 Conjugacy Theorem, Existence and Uniqueness Theorems . . . . . . 105
3 Representation Theory of Lie Algebras . . . . . . . . . . . . . . . . . . . . . . 127
3.1 Theorems of Ado and Iwasawa . . . . . . . . . . . . . . . . . . . . . . . . . . 127
3.2 Cyclic Modules and Weights . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
3.3 Characters and Harish-Chandra’s Theorem . . . . . . . . . . . . . . . . . . 150
3.4 Multiplicity Formulas of Weyl, Kostant, and Steinberg . . . . . . . . . 165
4 Chevalley Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
4.1 Classical Linear Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
4.2 Chevalley Basis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
4.3 Chevalley Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
4.4 Twisted Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
5 Representation Theory of Chevalley Groups . . . . . . . . . . . . . . . . . . 247
5.1 Language of Representation Theory . . . . . . . . . . . . . . . . . . . . . . . 247
5.2 Representations of Sn , and of GLð2:qÞ . . . . . . . . . . . . . . . . . . . . . 276
5.3 Steinberg Characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
5.4 Principal and Discrete Series Representations . . . . . . . . . . . . . . . . 300
5.5 Deligne–Lusztig Generalized Characters . . . . . . . . . . . . . . . . . . . . 304

ix
x Contents

Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
About the Author

Ramji Lal is Adjunct Professor at the Harish-Chandra Research Institute (HRI),

Allahabad, India. He started his research career at the Tata Institute of Fundamental
Research (TIFR), Mumbai, and served the University of Allahabad in different
capacities for over 43 years, as a professor, head of the department and the
coordinator of the DSA program. He was associated with HRI, where he initiated a
postgraduate program in mathematics and coordinated the Nurture Program of
National Board for Higher Mathematics (NBHM) from 1996 to 2000. After his
retirement from the University of Allahabad, he was an advisor cum adjunct pro-
fessor at the Indian Institute of Information Technology (IIIT), Allahabad, for over
three years. His areas of interest include group theory, algebraic K-theory and
representation theory

xi
Notations

½;  Lie product or bracket product

L a Lie algebra ðusuallyÞ
C L ðSÞ centralizer of S
N L ðVÞ normalizer of V
bn a bracket arrangement of weight n
ZðLÞ center of L
glðVÞðglðn; FÞÞ general linear Lie algebra
slðVÞðslðn; FÞÞ special linear Lie algebra
oðf Þðoð2m þ 1; FÞ orthogonal Lie algebra
spð2m; FÞ symplectic Lie algebra
dðn; FÞ Lie algebra of diagonal matrices
tðn; FÞ Lie algebra of upper triangular matrices
mðn; FÞ Lie algebra of strictly upper triangular matrices
DerðLÞ Lie algebra of derivations
IderðLÞ Lie algebra of inner derivations
ad adjoint representation
AB semi  direct product of A and B
T e ðGÞ Tangent space of G at e
LðGÞ Lie algebra of an algebraic ðLieÞ group G
Ad adjoint representation of an algebraic group on its Lie algebra
TðVÞ tensor algebra on V
SðVÞ symmetric algebra on V
ðUðLÞ; jL Þ universal enveloping algebra of L
RðLÞ radical of L
xs ðxn Þ semi  simpleðnilpotentÞ part of x
j Killing form
cq Casimir element associated with q
CðqÞðCðVÞÞ set of weights of a repnðmoduleÞ qðVÞ
U root system
Pa hyperplane determined by a root a

xiii
xiv Notations

ra reflection determined by a
WðUÞ the Weyl group of U
D a basis of a root system U
BðDÞ standard Borel subalgebra associated with a basis d
Kuniv ðUÞðKr Þ universal weight (root) lattice associated with U
ZðkÞ universal standard cyclic module of highest weight k
VðkÞ standard cyclic simple module with highest weight k
Ck set of weights of VðkÞ
Chk formal character of k
ChV formal character of the module V
cuniv universal Casimir element
vk the character afforded by ZðkÞ
LðKÞ Chevalley algebra of complex semi  simple Lie algebra over K
GðV; KÞ Chevalley group associated with a L-module V and a field K
Guniv universal Chevalley group
Gadj adjoint Chevalley group
Gp principal parabolic subgroup determined by p  D
StGðV;F q Þ Steinberg character of G
RT;H Deligne  Lusztig generalized character
Chapter 1
Lie Algebras

The concept and the theory of Lie algebras originated and took momentum from the
Lie theory of continuous groups. Locally, a Lie group is essentially a Lie algebra. To
every Lie group (complex or real), there is an associated Lie algebra. Structurally,
Lie subgroups and normal Lie subgroups of a Lie group associate faithfully with the
Lie subalgebras, and Lie ideals of the Lie algebra associated with the Lie group. The
isomorphism between Lie algebras corresponds to the local isomorphism between
the corresponding Lie groups. Indeed, the category of simply connected Lie groups is
equivalent to the category of Lie algebras. The theory of Lie algebras is indispensable
in the theory of Lie groups.
In another development, Magnus (refer to the excellent book entitled “Combinato-
rial Group Theory” by Magnus, Karrass, and Solitar) initiated the use of Lie algebras
in the study of discrete groups given in terms of presentations. The Lie algebras over
fields of positive characteristics have been very effectively and successfully used in
dealing with the restricted Burnside problem: “Is there a free object in the category
B(n, r ) of finite groups generated by n elements having the exponent dividing r ?”.
The problem was solved by E. Zelmonov in 1991 for which he got the Fields Medal
in 1994.
In the present chapter, we develop the basic language of Lie algebras including
universal enveloping algebras (PBW theorem), free Lie algebras, solvable, nilpotent,
and semi-simple Lie algebras. We also establish the theorem of Weyl about the com-
plete reducibility of representations of semi-simple Lie algebras. A field is usually
be denoted by F or also sometimes by K .

1.1 Definitions and Examples

Definition 1.1.1 A Lie algebra over a field F is a vector space L over F together
with a binary operation [, ] on L such that the following conditions hold:
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 1
R. Lal, Algebra 4, Infosys Science Foundation Series,
https://doi.org/10.1007/978-981-16-0475-1_1
2 1 Lie Algebras

1. [, ] is bi-linear in the sense that

[αx + β y, z] = α[x, z] + β[y, z],

and
[x, αy + βz] = α[x, y] + β[x, z]

for all x, y, z ∈ L and α, β ∈ F, where [x, y] denotes the image of (x, y) under [, ].
2. [, ] is an alternating map in the sense that [x, x] = 0 for all x ∈ L.
3. [x, [y, z]] + [y, [z, x]] + [z, [x, y]] = 0 for all x, y, z ∈ L.
[x, y] is called the Lie product of x and y. The third identity is termed as the
Jacobi identity.

We also say that (L , [, ]) is a Lie algebra over F, or [, ] is a Lie algebra structure

on the vector space L over F.
Proposition 1.1.2 Let (L , [, ]) be a Lie algebra over F. Then [x, y] = −[y, x]
for all x, y ∈ L. Conversely, let L be a vector space over a field F of characteristic
different from 2 and [, ] be a bilinear product on L such that conditions 1 and 3 of
the Definition 1.1 together with the condition
(2’) [x, y] = −[y, x] hold for all x, y ∈ L.
Then (L , [, ]) is a Lie algebra.

Proof Suppose that (L , [, ]) is a Lie algebra. Then by condition 2, [x + y, x +

y] = 0 for all x, y ∈ L. Using condition 1 and condition 2 again, we see that
[x, y] + [y, x] = 0. Conversely, suppose that [, ] is bi-linear and [x, y] = −[y, x]
for all x, y ∈ L. In particular, [x, x] = −[x, x] for all x ∈ V . Since the characteristic
of F is different from 2, [x, x] = 0 for all x ∈ L. 

Proposition 1.1.3 A Lie algebra structure determines,

 and is determined uniquely
by, a vector space homomorphism φ from L L to L satisfying the conditions

φ(x ∧ φ(y ∧ z)) + φ(y ∧ φ(z ∧ x)) + φ(z ∧ φ(x ∧ y)) = 0


for all x, y, z ∈ L, where L L denotes the exterior square of L.

Proof Let (L , [, ]) be a Lie algebra. Since [, ] is an alternating map, from the universal
property
 of the exterior square, there is a unique vector space homomorphism φ from
L L to L given by φ(x ∧ y) = [x, y]. The identity

φ(x ∧ φ(y ∧ z)) + φ(y ∧ φ(z ∧ x)) + φ(z ∧ φ(x ∧ y)) = 0

is a consequence of the Jacobi identity. Conversely, if such a homomorphism φ exists,

then we have the Lie product [, ] on L given by [x, y] = φ(x ∧ y). 

Let (L , [, ]) be a Lie algebra over a field F. A subspace V of L is said to be a Lie

subalgebra of L if [x, y] ∈ V for all x, y ∈ V . Evidently, V is also a Lie algebra
1.1 Definitions and Examples 3

with respect to the induced Lie product at its own right. A subalgebra V of L is said
be a Lie ideal if [x, v] ∈ V for all x ∈ L and v ∈ V . Evidently, a subspace V of L
is an ideal of L if and only if [v, x] ∈ V for all v ∈ V and x ∈ L. If V is an ideal
of L, then it can be easily observed that the quotient space L/V is a Lie algebra
with respect to the Lie product [, ] given by [x + V, y + V ] = [x, y] + V . This Lie
algebra is called the quotient Lie algebra of L modulo V . A Lie algebra L having
no nonzero proper ideal is called a simple Lie algebra.
Let (L , [, ]) and (L  , [, ] ) be Lie algebras over a field F. A linear transformation
f from L to L  is called Lie algebra homomorphism if f ([x, y]) = [ f (x), f (y)]
for all x, y ∈ L. This gives us a category L A F of Lie algebras over F. We have two
forgetful functors from L A F : One from L A F to the category of vector spaces, and the
second from L A F to the category S E T of sets. The question of the existence and the
construction of adjoints to these functors will be discussed in detail in the following
section. As usual, the image of a Lie subalgebra under a Lie homomorphism is a Lie
subalgebra, whereas the image of a Lie ideal need not be a Lie ideal. The inverse
image of a Lie subalgebra under a Lie homomorphism is a Lie subalgebra and also the
inverse image of a Lie ideal under a Lie homomorphism is a Lie ideal. In particular,
the kernel K er f = f −1 ({0}) of a Lie homomorphism is a Lie ideal. Further, as
usual the correspondence theorem, isomorphism theorems, and the Jordan-Holder
theorem follow for Lie algebras.
Let L be a Lie algebra over a field F. Let A and B be ideals of L. Let [A, B]
denote the subspace generated by the set {[a, b] | a ∈ A and b ∈ B}. By the Jacobi
identity
[x, [a, b]] = − [a, [b, x]] − [b, [x, a]]

for all x ∈ L , a ∈ A and b ∈ B. It follows that [x, [a, b]] ∈ [A, B] for all x ∈ L , a ∈
A and b ∈ B. This shows that [A, B] is also an ideal of L. In particular, [L , L] is also
an ideal of L. The ideal [L , L] is called the commutator or the derived subalgebra
of L. A Lie algebra L is said to be abelian if [L , L] = {0}. Evidently, [L , L] is the
smallest ideal of L by which if we factor, we get an abelian Lie algebra. L/[L , L] is
the largest quotient of L which is abelian. As in case of groups, we term L/[L , L]
as an abelianizer of L, and it is denoted by L ab . Any homomorphism f from L to
an abelian Lie algebra factors through L ab . A Lie algebra L is said to be perfect if
[L , L] = L. Evidently, every simple Lie algebra is perfect.
Let L be a Lie algebra over a field, and S be a subset of L. Let C L (S) denote the
subset {x ∈ L | [x, s] = 0 ∀s ∈ S}. The Jacobi identity ensures that C L (S) is a Lie
subalgebra of L. This Lie subalgebra is called the centralizer of S in L. If S = {a},
then we denote the centralizer of S by C L (a). If A is an ideal of L, then [z, [x, a]] =
−[x, [a, z]] − [a, [z, x]] = 0 for all z, x ∈ L and a ∈ C L (A). This shows that the
centralizer of an ideal is an ideal. In particular C L ([L , L]) is an ideal. The centralizer
C L (L) = {x ∈ L | [x, y] = 0 ∀y ∈ L} is called the center of L, and it is denoted
by Z (L). Let V be a Lie subalgebra of L. The set N L (V ) = {x ∈ L | [x, V ] ⊆ V }
is a subalgebra of L which is called the normalizer of V in L. Evidently, N L (V ) is
characterized by the fact that it is the largest subalgebra of L in which V is an ideal.
Note that a simple Lie algebra L is centerless in the sense that Z (L) = {0}.
4 1 Lie Algebras

Let L be a Lie algebra over a field F. Evidently, the intersection of a family of Lie
subalgebras/ideals of L is a Lie subalgebra/ideal of L. Let X be a subset of L. The
smallest Lie subalgebra of L containing X (the intersection of all Lie subalgebras
of L containing X ) is called the Lie subalgebra generated by X and it is denoted by
< X >. Similarly, we can talk of the ideal generated by a subset of L.
In non-associative structures, the bracket arrangements are important tools. To
describe the form of the elements of the subalgebra < X > generated by X in terms
of elements
 of X , we introduce the concept of bracket arrangements. For each
n ∈ N {0}, we introduce the set Bn of elements called the bracket arrangements
of weight n. This we do by the induction on n. B0 = {∅}, B1 = {()}, B2 =
{(()())}, B3 = {((()())()), (()(()()))}. Assume that the set Br has already
been defined for all r < n. For each r, s ≤ n − 1, r + s = n, let Br,s denote the set
{(βγ) | β ∈ Br and γ ∈ Bs }. Define Bn = r +s=n Br,s . Each  is called a place
holder. This defines bracket arrangements of different weights. Given a sequence
x1 , x2 , · · · xm , xm+1 , · · · of elements of a Lie algebra L, for each bracket arrangement
βn of weight n, we define the element βn (x1 , x2 , · · · , xn ) of L. This, again, we do
by induction on n as follows. Define β1 (x1 ) = x1 . Assume that βr (x1 , x2 , · · · , xr )
has been defined for all r < n. Suppose that βn = (βr , βs ), where βr is a bracket
arrangement of weight r and βs is that of weight s. Then we put βn (x1 , x2 , · · · , xn ) =
[βr (x1 , x2 , · · · , xr ), βs (xr +1 , xr +2 , · · · , xr +s )]. The elements of the type
βn (x1 , x2 , · · · , xn ) are called the bracket arrangements of the sequence
x1 , x2 , · · · xm , xm+1 , · · · of elements of L. Let X be a subset of L. Let β(X ) denote the
set of all bracket arrangements corresponding to all sequences in X . Clearly, the Lie
subalgebra < X > of L generated by X is precisely the set of all linear combinations
of members of β(X ). The reader is asked to describe the form of the elements of the
ideal generated by X .
Lie Algebras of Low Dimensions

Example 1.1.4 In this example, we describe the isomorphism classes of Lie algebras
of dimension at the most 2. Evidently, there is a unique one-dimensional Lie algebra
which is abelian. Suppose that L is a non-abelian two-dimensional Lie algebra. Let
{x, y} be a basis of L. Then [L , L] is the subspace generated by [x, y] = z = 0.
Thus, [L , L] is a one-dimensional subspace generated by {z}. Let {z, u} be a basis of
L. Then [z, u] = αz for some α = 0. Taking v = α−1 u, we see that L is the Lie
algebra generated by {z, v} subject to the relation [z, v] = z. An arbitrary element
of L is of the form αz + βv and [αz + βv, γz + δv] = (αδ − βγ)z. Thus, up to
isomorphism, there is only one non-abelian Lie algebra of dimension 2. Further, if
αz + βv ∈ Z (L), then −βz = [αz + βv, z] = 0 and αz = [αz + βv, v] = 0.
This implies that α = 0 = β. It follows that Z (L) = {0}. Evidently, no Lie algebra
of dimension less than 3 is perfect.

Example 1.1.5 In this example, we classify all Lie algebras of dimension 3. Let L
be a Lie algebra of dimension 3.
Case (1). [L , L] = {0}. In this case, L is abelian and there is only one such Lie
algebra up to isomorphism.
1.1 Definitions and Examples 5

Case (2). Dim[L , L] = 1 and [L , L] ⊆ Z (L). Suppose that [L , L] = F x1 ,

where x1 = 0. Embed {x1 } into a basis {x1 , x2 , x3 } of L. Evidently, [x2 , x3 ] = αx1
for some α = 0. We may choose x2 so that α = 1. Since x1 ∈ Z (L), the Lie product
in L is uniquely given by

[α1 x1 + α2 x2 + α3 x3 , β1 x1 + β2 x2 + β3 x3 ] = (α2 β3 − α3 β2 )x1 .

It can be easily seen that the Jacobi identity is satisfied. Thus, in this case also we
have a unique Lie algebra up to isomorphism given as above.
Case(3). Dim [L , L] = 1 and [L , L]  Z (L). Suppose that [L , L] = F x1 ,
where x1 = 0. Since [L , L]  Z (L), there is a nonzero element x2 of L such that
[x1 , x2 ] = 0. We may take x2 such that [x1 , x2 ] = x1 . Evidently, {x1 , x2 } is linearly
independent. Embed it into a basis {x1 , x2 , x3 } of L. Suppose that [x1 , x3 ] = βx1 . If
β = 0, we may choose x1 so that [x1 , x3 ] = x1 . Note that the relation [x1 , x2 ] = x1
will not change. We can replace x3 by x3 − x2 and then [x1 , x3 ] = 0. Suppose that
[x2 , x3 ] = γx1 . If γ = 0, we may further modify x3 by taking it to be γ −1 x3 so that
[x2 , x3 ] = x1 . Observe that the remaining relations remain the same. Again replace
x3 by x1 + x3 . Then [x2 , x3 ] becomes 0. Still note that the remaining relations remain
the same. Also note that all the time {x1 , x2 , x3 } remains a basis. We get a Lie algebra
structure on L given by

[α1 x1 + α2 x2 + α3 x3 , β1 x1 + β2 x2 + β3 x3 ] = (α1 β2 − α2 β1 )x1 .

It can be easily seen that the Jacobi identity is satisfied. Thus, in this case also we
have a unique Lie algebra up to isomorphism described as above.
Case (4). Dim [L , L] = 2. We first show that [L , L] is abelian. Suppose the
contrary. Then as in Example 1.1.4, there is a basis {x1 , x2 } of [L , L] with [x1 , x2 ] =
x2 . In particular, Z ([L , L]) = {0}. Consider
 the centralizer C L ([L , L]) of [L , L]
which is an ideal of L. Evidently, [L , L] C L ([L , L]) = {0}. Let x be an element
of L. Suppose that [x, x1 ] = αx1 + βx2 and [x, x2 ] = γx1 + δx2 . Then using the
Jacobi identity,

γx1 + δx2 = [x, x2 ] = [x, [x1 , x2 ]] = −[x1 , [x2 , x]] − [x2 , [x, x1 ] = (α + δ)x2 .

This means that α = 0 = γ. Thus, [x, x1 ] = βx2 and [x, x2 ] = δx2 . Hence

[x − δx1 + βx2 , x1 ] = 0 = [x − δx1 + βx2 , x2 ].

This shows that x − δx1 + βx2 belongs to C L ([L , L]). Consequently, L =

[L , L] ⊕ C L ([L , L]). Since C L ([L , L]) is a one-dimensional abelian Lie algebra,
[L , L] = [[L , L], [L , L]] = 0. This is a contradiction. Thus, [L , L] is abelian.
Let {x1 , x2 , x3 } be a basis of L, where {x1 , x2 } is a basis of [L , L]. Suppose that
[x3 , x1 ] = αx1 + βx2 and [x3 , x2 ] = γx1 + δx2 . Since [L , L] is abelian and of
dimension 2, {[x3 , x1 ], [x3 , x2 ]} is also a basis of [L , L]. This gives us a non-singular
2 × 2 matrix
6 1 Lie Algebras
 
αβ
A = .
γ δ

Thus, for a Lie algebra L of dimension 3 for which the dimension of [L , L] is

2, a choice of basis {x1 , x2 , x3 } of L with {x1 , x2 } as a basis of [L , L] determines a
non-singular 2 × 2 matrix A described as above. Conversely, given a non-singular
2 × 2 matrix A, take a vector space L with a set {x1 , x2 , x3 } as a basis and define the
product [, ] by

[λ1 x1 +λ2 x2 + λ3 x3 , μ1 x1 + μ2 x2 + μ3 x3 ] =
(−λ1 μ3 + λ3 μ1 )(αx1 + βx2 ) + (λ3 μ2 − λ2 μ3 )(γx1 + δx2 ).

It can be easily observed that (L , [, ]) is a Lie algebra with [L , L] being of dimen-

sion 2, and also the matrix associated with L with respect to the basis {x1 , x2 , x3 } is
the given matrix A. However, different matrices may determine isomorphic Lie alge-
bras. We try to classify it faithfully. If we fix an element x3 = 0 outside [L , L] and
change the basis {x1 , x2 } of [L , L] to a basis {x1 , x2 }, then the matrix A changes to a
matrix A which is similar to A. However, if we fix a basis {x1 , x2 } of L and change x3
to x3 so that {x1 , x2 , x3 } becomes a basis of L, then x3 = λx3 + x, for some nonzero
element λ in F and x ∈ [L , L]. Since [L , L] is abelian, [x3 , x1 ] = λ(αx1 + βx2 )
and [x3 , x2 ] = λ(γx1 + δx2 ). The matrix A changes to the matrix λA. It follows that
every Lie algebra L of dimension 3 in which [L , L] is of dimension 2 determines,
and is determined uniquely up to isomorphism by, a conjugacy class of Collineation
group of 2 × 2 matrices. If F is an algebraically closed field, then using the Jor-
dan theorem, we see that a conjugacy class of Collineation group of 2 × 2 matrices
contains one and only one member from the class
   1 β 
10
| α = 0} { |β=0
0α 01

of matrices. Thus, in this case, we have infinitely many Lie algebras L having a basis
{x1 , x2 , x3 } satisfying one and only one of the following two types of relations:

1. [x1 , x2 ] = 0, [x3 , x1 ] = x1 , [x3 , x2 ] = αx2 or

2. [x1 , x2 ] = 0, [x3 , x1 ] = x1 + βx2 , and [x3 , x2 ] = x2 .

Case (5). Dim [L , L] = 3 or equivalently [L , L] = L. We describe and classify
such Lie algebras up to isomorphisms. Let P L(3, F) denote the set of isomorphism
classes of perfect Lie algebras of dimension 3 over a field F. The isomorphism class
determined by L will be denoted by [L]. A non-singular symmetric matrix A is said
to be multiplicatively cogredient to a non-singular symmetric matrix B if there is a
nonzero scalar λ and a non-singular matrix P such that A = λP B P t . Thus, if every
element of the field F has a square root (for example C), then the multiplicative
cogredience is the same as the usual congruence. Let Ŝ(3, F) denote the set of
1.1 Definitions and Examples 7

multiplicatively cogredient classes of 3 × 3 non-singular symmetric matrices with

entries in F. The cogredient class determined by the matrix A will be denoted by [A].
We exhibit a natural bijective correspondence between P L(3, F) to Ŝ(3, F). Let L
be a three-dimensional perfect Lie algebra, and let {x1 , x2 , x3 } be a basis of L. Put
y1 = [x2 , x3 ], y2 = [x3 , x1 ], and y3 = [x1 , x2 ]. Evidently, {y1 , y2 , y3 } generates
3
[L , L] = L, and as such it is also a basis of L. Suppose that y j = i=1 ai j x i . Then
A = [ai j ] is a non-singular matrix. We term A as the matrix of the basis {y1 , y2 , y3 }
with respect to the basis {x1 , x2 , x3 }. The Jacobi identity in L is equivalent to the
identity
[x1 , y1 ] + [x2 , y2 ] + [x3 , y3 ] = 0.

This is further equivalent to

3 3 3
[x1 , ai1 xi ] + [x2 , ai2 xi ] + [x3 , ai3 xi ] = 0.
i=1 i=1 i=1

Using the bi-linear and alternating property of the product [, ], we obtain the identity

(a21 − a12 )y3 + (a13 − a31 )y2 + (a32 − a23 )y1 = 0.

This shows that A = [ai j ] is a symmetric 3 × 3 matrix. Thus, every perfect Lie
algebra of dimension 3 together with a choice {x1 , x2 , x3 } of a basis of L determines
a unique non-singular symmetric matrix A described as above. Conversely, suppose
that we are given a 3 × 3 symmetric matrix A = [ai j ]. Let L be a vector space with
a basis {x1 , x2 , x3 } consisting of 3 elements. Define a product [, ] on L by
3 3 3
αi xi , βi xi = (α1 β2 − α2 β1 ) ai3 xi + (−α1 β3
i=1 i=1 i=1
3 3
− α3 β1 ) ai2 xi + (α2 β3 − α3 β2 ) ai1 xi .
i=1 i=1

The fact that A is a non-singular symmetric matrix implies that L is a perfect Lie
algebra of dimension 3 and the basis {x1 , x2 , x3 }, in turn, determines back the given
matrix A. To classify these Lie algebras, we analyze the effect of the change of
basis of L on the matrix A. Let {u 1 , u 2 , u 3 } be another basis with the associated basis
{v1 , v2 , v3 }, where v1 = [u 2 , u 3 ], v2 = [u 3 , u 1 ], and v3 = [u 1 , u 2 ]. Let B = [bi j ]
3
be the associated non-singular symmetric matrix. Then v j = i=1 bi j u i . We relate
3
A and B. Suppose that u j = i=1 ci j x i . Clearly, C = [ci j ] is non-singular matrix.
Now,
v1 = [u 2 , u 3 ]
3 3
= [ i=1 ci2 xi , i=1 ci3 xi ]
= (c22 c33 − c32 c23 )y1 + (−c12 c33 + c32 c13 )y2 + (c12 c23 − c22 c12 )y3
3
= i=1 di1 yi ,
where [di1 ] represent the the first column of (C t )ad j = (C ad j )t = (Det C)(C t )−1 ,
3
where C ad j denotes the adjoint of C. Similarly, for each j, v j = i=1 di j yi , where
8 1 Lie Algebras

[di j ] represent the jth column of adC t . We have the following transformations
among the bases of L:

B −1 (C t )ad j A C
U → V → Y → X = U → X,

where U denotes the basis {u 1 , u 2 , u 3 }, V denotes the basis {v1 , v2 , v3 }, Y repre-

sents the basis {y1 , y2 , y3 }, and X represents the basis {x1 , x2 , x3 }. This shows that
A (C t )ad j B −1 = C. Since (C t )ad j = (Det C)(C t )−1 ,

A = (Det C)−1 C BC t ,

where C is a non-singular matrix. This shows that different choice of bases for L give
rise to cogredient matrices. More generally, if f is an isomorphism from a perfect
Lie algebra L of dimension 3 with a basis {x1 , x2 , x3 } to another perfect Lie algebra
L  of dimension 3 with a basis {x1 , x2 , x3 }, then the associated matrices A and A are
cogredient to each other. Consequently, we get a map η from P L(3, F) to Ŝ(3, F)
which is given by η([L]) = [A], where A is the non-singular symmetric matrix
associated with L with respect to a basis of L. Further, suppose that we are given a
3 × 3 symmetric matrix A = [ai j ]. Let L be a vector space with a basis {x1 , x2 , x3 }
consisting of 3 elements. Define a product [, ] on L by

3 3 3
αi xi , βi xi = (α1 β2 − α2 β1 ) ai3 xi + (−α1 β3 +
i=1 i=1 i=1
3 3
α3 β1 ) ai2 xi + (α2 β3 − α3 β2 ) ai1 xi .
i=1 i=1

The fact that A is a non-singular symmetric matrix implies that L is a perfect

Lie algebra of dimension 3 and basis {x1 , x2 , x3 }, in turn, determines back the given
matrix A. Thus, the map η is surjective. Next, let A and B be non-singular symmetric
matrices which are cogredient. Let L be a Lie algebra together with a basis {x1 , x2 , x3 }
which is associated with the matrix A and let L  be a Lie algebra with a basis
{x1 , x2 , x3 } which is associated with the matrix B. Then there is a nonzero scalar λ
and a non-singular matrix P such that A = λP B P t . Put P = μQ. Then A =
λμ2 Q B Q t . We wish to choose μ so that λμ2 = (Det Q)−1 . This is possible,
since (Det Q)−1 = μ3 (Det P)−1 . Thus, we have a non-singular matrix Q such
that A = (Det Q)−1 Q B Q t . It follows from the above discussion that the linear
transformation from L to L  having Q as the matrix representation with respect to
the basis {x1 , x2 , x3 } of L and the basis {x1 , x2 , x3 } of L  is an isomorphism from L
to L  . This shows that the map η is a bijective correspondence.
Now suppose that the field F is of characteristic different from 2. Then every
symmetric matrix with entries in F is congruent to a diagonal matrix (see Theorem
5.6.15, Algebra 2). Consequently, every equivalence class in Ŝ(3, K ) contains a
matrix of the form Diag(α, β, 1), α = 0 = β. Thus, every perfect three-dimensional
Lie algebra over a field of characteristic different from 2 is isomorphic to a Lie algebra
1.1 Definitions and Examples 9

L α,β having a basis {x1 , x2 , x3 } such that [x1 , x2 ] = x3 , [x2 , x3 ] = αx1 , and
[x3 , x1 ] = βx2 . Note that L α,β may be isomorphic to L γ,δ even if {α, β} = {γ, δ}.
If the field F is the field R of real numbers, then it follows (see Theorem 5.6.22,
Algebra 2 ) that every equivalence class contains one and only one of the following
two matrices Diag(1, 1, 1) or Diag(−1, 1, 1). Thus, there are only two different
perfect Lie algebras of dimension 3 over the field R of real numbers, and they are
given by the set {[x2 , x3 ] = x1 , [x3 , x1 ] = x2 , [x1 , x2 ] = x3 } of relations or else
it is given by the set {[x2 , x3 ] = −x1 , [x3 , x1 ] = x2 , [x1 , x2 ] = x3 } of relations.
Identify them.
Suppose that the field F is the field C of complex numbers. Then every non-
singular symmetric 3 × 3 matrix is congruent to the identity matrix. Hence there is
only one perfect Lie algebra L of dimension 3 over the field C of complex numbers. L
has a basis {x1 , x2 , x3 } with relations [x2 , x3 ] = x1 , [x3 , x1 ] = x2 , [x1 , x2 ] = x3 .
Indeed, this Lie algebra has a nice representation as the Lie algebra sl(2, C) consisting
of 2 × 2 matrices with entries in C and with trace 0. The Lie product is given by
[x, y] = x y − yx (see Example 1.1.7).
A cyclic Lie algebra is a Lie algebra which is generated by a single element. Evidently,
a nontrivial cyclic Lie algebra is an abelian algebra on a one-dimensional space. A
Lie algebra L has no proper subalgebras if and only if L is a cyclic Lie algebra. A
Lie algebra L is said to be a simple Lie algebra if it has no proper ideals. Evidently,
an abelian Lie algebra is simple if and only if it is cyclic. As in the case of finite
groups, the problem of classification of finite-dimensional Lie algebras reduces to
the following two problems:
1. Classify all finite-dimensional simple Lie algebras up to isomorphisms.
2. Given a pair of (A, B) of Lie algebras, to classify Lie algebras L having an ideal
A isomorphic to A as Lie algebra such that L/A isomorphic to B.
These problems will be discussed in due course.
Proposition 1.1.6 L/Z (L) cannot be a nontrivial cyclic Lie algebra.
Proof Suppose that L/Z (L) = < x + Z (L) > is a cyclic Lie algebra. Then any
element of L is of the form αx + u, where u ∈ Z (L). Evidently, [αx + u, βx +
v] = 0 for all α, β ∈ K and u, v ∈ Z (L). This implies that L = Z (L). 
Example 1.1.7 Let L be a non-abelian Lie algebra of dimension 3. It follows from
the above proposition that Z (L) = {0} or else Z (L) is of dimension 1. Suppose that
L is perfect. Then every quotient of L is perfect. Since no Lie algebra of dimension
less than 3 is perfect, it follows that L has no proper ideals. Thus, every perfect
Lie algebra of dimension 3 is simple. Let sl(2, F) denote the vector space of 2 × 2
matrices with entries in the field F having 0 trace. Then sl(2, F) is a Lie algebra
with respect to the Lie product given by [A, B] = AB − B A. Evidently, sl(2, F)
is three dimensional with {e12 , e21 , h} as a basis, where
     
01 00 1 0
e12 = , e21 = , and h = .
00 10 0 −1
10 1 Lie Algebras

It can be easily observed that [e12 , e21 ] = h, [h, e12 ] = 2e12 , and [h, e21 ] =
−2e21 . Thus, if the characteristic of F is different from 2, then sl(2, F) is perfect.
Consequently, sl(2, F) is simple provided that the characteristic of F is different
from 2.
Next, suppose that Z (L) is of dimension 1 generated by {z}. Let {x, y, z} be a basis
of L. L/Z (L) is a two-dimensional Lie algebra generated by {x + Z (L), y + Z (L)}.
If L/Z (L) is abelian, then [L , L] ⊆ Z (L), and, there is a unique (up to isomorphism)
non-abelian Lie algebra L of dimension 3 as described in Case 2 of Example 1.1.5.
Now, suppose that L/Z (L) is non-abelian. Then [L , L]  Z (L). If dim [L , L] = 1,
then we have a unique Lie algebra L as described in Case 3 of Example 1.1.5. Finally,
if dim [L , L] = 2, then as described in Case 4 of Example 1.1.5, [L , L] is abelian
and there are two types of Lie algebras.

Now, we list some more important examples.

Example 1.1.8 Consider the usual vector space R3 over R. The vector product ×
in R3 is an alternating product which satisfies the Jacobi identity. As such, (R3 , ×)
is a three-dimensional real Lie algebra. All nontrivial proper subalgebras of (R3 , ×)
are one dimensional. However, it has no nontrivial proper ideals. Thus, (R3 , ×) is a
simple Lie algebra over R.

Example 1.1.9 Let L be a three-dimensional vector space over F with {x, y, z} as a

basis. Then there is a unique Lie product [, ] on L subject to [x, y] = z, [x, z] = y
and [y, z] = 0. Observe that the abelian subalgebra A of L generated by {y, z} is
an ideal such that L/A is also abelian.

Example 1.1.10 Let A be an associative algebra over a field F, for example, a

polynomial algebra in non-commuting variables or a group algebra F(G) over a
group G. Define a product [, ] on A given by [a, b] = ab − ba. It is easily observed
that (A, [, ]) is a Lie algebra. This Lie algebra will be denoted by A L , and it will be
termed as the Lie algebra associated with the associative algebra A. As such, we have
a functor from the category ASS F of associative algebras over F to the category L A F
of Lie algebras over the field F. The adjoint to this functor is an important functor
which is extremely useful in the representation theory and the structure theory of
Lie algebras. This adjoint functor will be discussed in detail in the next section. This
also provides several examples of Lie algebras.

Example 1.1.11 Let (L , [, ]) be a finite-dimensional Lie algebra over F. Let

{x1 , x2 , · · · , xn } be an ordered basis of L. This gives us n skew symmetric matrices
M 1 = [ai1j ], M 2 = [ai2j ], · · · , M n = [ainj ] given by

n
[xi , x j ] = a k xk .
k=1 i j

Further, the Jacobi identity induces the following identity among the entries of the
matrices M k , 1 ≤ k ≤ n:
1.1 Definitions and Examples 11

n p p p
(aikj akl + a kjl aki + alik ak j ) = 0
k=1

for all i, j, k, l, p. Conversely, given the set of skew symmetric matrices M k = [aikj ]
satisfying the above condition, there is a unique Lie algebra structure [, ] on F n given
by
n
[ei , e j ] = aikj ek ,
k=1

where {e1 , e2 , · · · , en } is the standard basis of F n . The entries aikj are called the
structure constants of the Lie algebra L associated with the basis
{x1 , x2 , · · · , xn }. Describe the effect of change of the base on the structure constants.

Example 1.1.12 Let V be a vector space over a field F. Then End F (V ) is an

associative algebra over F. The associated Lie algebra is denoted by gl(V ). Thus,
gl(V ) = End F (V ) and the Lie product [, ] is given by [A, B] = AB − B A for all
A, B ∈ gl(V ). The Lie algebra (gl(V ), [, ]) is called the general linear Lie alge-
bra on V . Any Lie subalgebra of gl(V ) is called a linear Lie algebra. Indeed, it
is a fact (Ado and Iwasava theorems: Theorems 3.1.23 and 3.1.25) that every finite-
dimensional Lie algebra is a linear Lie algebra over a vector space of finite dimension.
If V is of dimension n, then gl(V ) is of dimension n 2 . Fixing a basis of gl(V ), gl(V )
can be identified with the Lie algebra gl(n, F) of n × n matrices with entries in F.
Clearly, {ei j | 1 ≤ i ≤ n, 1 ≤ j ≤ n} is a basis of gl(n, F), where ei j is the matrix
all of whose entries are 0 except the i j entry which is 1. Evidently, ei j ekl = δ jk eil ,
where δ jk is the Kronecker delta. Thus,

[ei j , ekl ] = δ jk eil − δli ek j

for all i, j, k, l. This means that the structure constants of gl(n, F) with respect to
the standard basis {ei j | 1 ≤ i ≤ n, 1 ≤ j ≤ n} of gl(n, F) are members of the prime
field in F.

Next, we introduce some important families of Lie subalgebras of gl(V ) ≈

gl(n, F) termed as classical linear Lie algebras.
Example 1.1.13 The family An . Let V be a vector space of dimension n + 1 over
a field F. Let sl(V ) ≈ sl(n + 1, F) denote the set of linear endomorphisms of V
which are of trace 0. Evidently, sl(V ) is a vector subspace of gl(V ) which is of
dimension (n + 1)2 − 1. Since T r (AB − B A) = 0, it follows that [A, B] ∈ sl(V )
for all A, B ∈ sl(V ). This shows that sl(V ) is a Lie subalgebra of gl(V ). Observe
that sl(V ) is not an associative subalgebra of the associative algebra gl(V ). The Lie
subalgebra sl(V ) ≈ sl(n + 1, F) is called a special linear Lie algebra. Evidently,
{ei j | i = j} {eii − ei+1i+1 | 1 ≤ i ≤ n + 1} is a basis of sl(V ). This basis is called
the standard basis of sl(V ). Determine the structure constants.

Example 1.1.14 The family Bn . Let f be a nondegenerate symmetric bi-linear form

on a vector space V of odd dimension m = 2n + 1 over a field F. Let o( f ) denote
12 1 Lie Algebras

the set of all endomorphisms A of V such that f (A(v), w) = − f (v, A(w)) for all
v, w ∈ V . Evidently, o( f ) is a subspace of gl(V ). If A, B ∈ o( f ), then

f ((AB − B A)(v), w) = f (A(B(v)), w) − f ((B(A(v)), w) =

− f (v, (AB − B A)(w))

for all v, w ∈ V . Hence [A, B] ∈ o( f ) for all A, B ∈ o( f ). Thus, o( f ) is a Lie

subalgebra of gl(V ). This Lie algebra is called the orthogonal Lie algebra associated
with f . We try to represent it in matrix form. Since f is nondegenerate, there is a
basis (see Algebra 2, Sect. 5.6) {v1 , v2 , · · · , v2n+1 } such that the matrix M( f ) =
[ f (vi , v j ] of f with respect to this basis is given by
⎡ ⎤
1 01×n 01×n
M( f ) = ⎣ 0n×1 0n×n In ⎦ ,
0n×1 In 0n×n

where 0r ×s denotes the r × s zero matrix and In is the n × n identity matrix. Evi-
dently, A ∈ o( f ) if and only if M( f )M(A) = −M(A)t M( f ). Suppose that the
matrix M(A) of A with respect to the above basis is expressed as
⎡ ⎤
a α β
M(A) = ⎣ γ t P Q ⎦ ,
t
δ R S

where a ∈ F, α, β, γ, δ are row vectors in F n , whereas P, Q, R, and S are members

of gl(n, F). The condition M( f )M(A) = −M(A)t M( f ) implies that a = 0, α =
−γ, β = −δ, R and Q are skew symmetric matrices, whereas P t = −S. This, in
turn implies that the trace of each member of o( f ) is 0, and o( f ) is a Lie subalgebra
of sl(V ). The Lie algebra o( f ) is called an orthogonal Lie algebra associated with
f . The matrices of the form M(A) described above form a Lie algebra under the
Lie product of matrices. This Lie algebra is isomorphic to o( f ), and it is denoted
by o(2n + 1, F). Evidently, the dimension of o(2n + 1, F) is 2n 2 + n. Determine
a standard basis of o(2n + 1, F) as in the above example. Determine the structure
constants with respect to this basis. Check that they also belong to the prime field
contained in F.

Example 1.1.15 The family Cn . Let f be a nondegenerate alternating bi-linear

form on a vector space V of dimension m = 2n over a field F. Observe that
there is no degenerate alternating form on odd-dimensional spaces unless the field
is of characteristic 2. Let sp( f ) denote the set of all endomorphisms A of V such
that f (A(v), w) = − f (v, A(w)) for all v, w ∈ V . Evidently, sp( f ) is a subspace
of gl(V ). As in the above example, sp( f ) is a Lie subalgebra of gl(V ). This Lie
algebra is called the symplectic Lie algebra associated with f . We try to represent
it in matrix form. Since f is nondegenerate, there is a basis (see Algebra 2, Sect. 5.6)
{v1 , v2 , · · · , v2n } such that the matrix M( f ) = [ f (vi , v j )] of f with respect to this
basis is given by
1.1 Definitions and Examples 13
 
0n In
M( f ) = .
−In 0n

Evidently, A ∈ sp( f ) if and only if M( f )M(A) = −M(A)t M( f ). Suppose that

the matrix M(A) of A with respect to the above basis is expressed as
 
P Q
M(A) = ,
R S

where P, Q, R, and S are members of gl(n, F). The condition M( f )M(A) =

−M(A)t M( f ) implies that R and Q are symmetric matrices, whereas P t = −S.
This, in turn implies that the trace of each member of sp( f ) is 0, and sp( f ) is a
Lie subalgebra of sl(V ). The set of matrices of the form M(A) described above is
denoted by sp(2n, F) and it is a Lie algebra with respect to the Lie product of matri-
ces. This Lie algebra is isomorphic to sp( f ). Evidently, the dimension of sp(2n, F)
is 2n 2 + n. Determine a standard basis of sp(2n, F) and also the structure constants
with respect to this basis. Observe that they belong to the prime field of F.
Example 1.1.16 The family Dn . In Example 1.1.14, we described the orthogonal
Lie algebras on odd-dimensional spaces. In this example, we describe orthogonal Lie
algebras on even-dimensional spaces. Let f be a symmetric nondegenerate bi-linear
form on a vector space V of even dimension 2n. There is a basis {v1 , v2 , · · · , v2n }
of V such that the matrix M( f ) of f with respect to this basis is
 
0n In
M( f ) = .
In 0n

As in Example 1.1.14, we have a Lie subalgebra o( f ) of sl(2n, F) which is

isomorphic to the Lie algebra o(2n, F) = {A | M( f )A = −At M( f )} of matrices.
As in previous examples, it can be seen that the dimension of o(2n, F) is 2n 2 − n.
The reader may determine the standard basis, and also the structure constants.
Apart from these family of linear Lie algebras, there are other important linear Lie
subalgebras of gl(n, F) which are important for the structure theory of Lie algebras.

Example 1.1.17 The set d(n, F) of all diagonal matrices in gl(n, F) form an abelian
Lie subalgebra of gl(n, F) which is called a total subalgebra of gl(n, F). The set
t (n, F) of upper triangular matrices in gl(n, F) is also a Lie subalgebra of gl(n, F)
which is termed as a Borel subalgebra of gl(n, F). The set n(n, F) of strict upper
triangular matrices (diagonal entries 0) also form a Lie subalgebra of gl(n, F). It is
easy to observe that [t (n, F), t (n, F)] = {[A, B] | A, B ∈ t (n, F)} = n(n, F).
Find the dimension of each of these subalgebras.
Derivations and Lie Algebras
Let A be an algebra over a field F which may be non-associative. More explicitly,
A is a vector space over a field F together with a bi-linear product on A denoted by
14 1 Lie Algebras

juxtaposition. A F-linear transformation d from A to A is called a derivation on A

if
d(ab) = d(a)b + ad(b)

for all a, b ∈ A. Let Der (A) denote the set of all derivations on A. Evidently, Der (A)
is a subspace of End F (A). The composite of two derivations need not be a derivation
(give an example). However, [d, d  ]] = dd  − d  d can easily be seen to be a deriva-
tion. Thus, Der (A) is a Lie subalgebra of gl(A). In particular, if L is a Lie algebra
over F, a derivation d on L is, by the definition, a linear transformation from L to L
such that
d([v, w]) = [d(v), w] + [v, d(w)]

for all v, w ∈ L. For each v ∈ L, the map ad(v) from L to L given by ad(v)(w) =
[v, w] is clearly a linear transformation. Further,

ad(v)([w, u]) = [v, [w, u]] = [ad(v)(w), u] + [w, ad(v)(u)]

for all u, v, w ∈ L, thanks to the Jacobi identity. Hence for each v ∈ L, ad(v) is a
derivation on L. Such a derivation is called an inner derivation of L determined by
the element v. Other derivations are called the outer derivations. In turn, we get a
map ad from L to Der (L) which associates with each v ∈ L the inner derivation
ad(v). Evidently, ad is a linear transformation. Further,

[ad(v), ad(w)](u) = ad(v)(ad(w)(u)) − ad(w)(ad(v)(u)) =

[v, [w, u]] − [w, [v, u]] = [[v, w], u] = ad([v, w])(u)

for all u, v, w ∈ L, thanks to the Jacobi identity. Hence ad is a Lie homomorphism

from L to Der (L). This is called the adjoint representation of L. The image of ad
is precisely the Lie algebra of inner derivations of L, and it is denoted by ad(L) or
I der (L). The kernel K er ad = {v ∈ L | [v, w] = 0 ∀w ∈ L} of ad is the center
Z (L) of the Lie algebra L. By the fundamental theorem of homomorphism

L/Z (L) ≈ I der (L) = ad(L).

Let d ∈ Der (L) and v ∈ L. Then

[d, ad(v)](w) = d(ad(v)(w)) − ad(v)(d(w)) = d([v, w]) − [v, d(w)] =

[d(v), w] + [v, d(w)] − [v, d(w)] = [d(v), w] = ad(d(v))(w)

for all v, w ∈ L. This shows that I der (L) is an ideal of Der (L). The quotient Lie
algebra Der (L)/I der (L) is called the Lie algebra of outer derivations of L which
will be denoted by Oder (L). We get a short exact sequence
1.1 Definitions and Examples 15

i ν
0 −→ I der (L) → Der (L) → Oder (L) −→ 0,

and also an exact sequence

i ad ν
0 −→ Z (L) → L → Der (L) → Oder (L) −→ 0

of Lie algebras.
Semi-Direct Product and Split Extension
As in the case of groups, we introduce the notion of semi-direct product and split
extensions in the category of Lie algebras. Here in this case, the derivation algebras
play the role of automorphism groups. Thus, let A and B be Lie algebras over a field
F. Let σ be a Lie algebra homomorphism from B to the derivation algebra Der (A).
Consider the vector space L = A × B. Define the product [, ] on L by

[(a, b), (a  , b )] = ([a, a  ] + σ(b )(a) − σ(b)(a  ), [b, b ])

for all a, a  ∈ A and b, b ∈ B. It is a straightforward verification to show that L

is a Lie algebra. The first inclusion map i 1 from A to L given by i 1 (a) = (a, o)
is a monomorphism whereas the second projection map p2 from L to B given by
p2 ((a, b)) = b is an epimorphism. Further, A × {0} is the kernel of p2 and so it is
an ideal of L. We get a short exact sequence

i1 p2
0 −→ A → L → B −→ 0.

Evidently, the second inclusion i 2 from B to L is a splitting of the short exact

sequence. Note that L = A + B  , where A = A × {0} and B  = {0} × B,
A  is an ideal isomorphic to A, and B  is a subalgebra isomorphic to B. Further

A B  = {0}. We say that L is an external semi-direct product of A with B, and
it is said to be an internal semi-direct product of A with B  . Note that if σ is trivial
homomorphism, then the semi-direct product is the direct product. Suppose that L is
 semi-direct product of its ideal A by its subalgebra B. Then L = A + B
the internal
and A B = {0}. An element x of L is uniquely expressible as x = a + b, where
a ∈ A and b ∈ B. For each element b of B, we have a derivation σ(b) of A given by
σ(a)(b) = [a, b]. It can easily be seen that σ is a homomorphism from B to Der (A).
Further, the map (a, b) → a + b from the external semi-direct product A × B to the
internal semi-direct product L is easily seen to be an isomorphism. Thus, an internal
semi-direct product is isomorphic to an external semi-direct product. The external
semi-direct product of A with B relative to a homomorphism σ is denoted by A σ B.
If L is a Lie algebra having an ideal A and a subalgebra B such that L = A + B
and A B = {0}, then we simply denote it by A  B.
Linear Algebraic Groups and Linear Lie Algebras
The discussions to follow will be useful in the representation theory. Recall that
an algebraic variety G together with a group structure on G is called an algebraic
16 1 Lie Algebras

group or a group variety if the group operations ((a, b) → ab, a → a −1 ) are

morphisms between the corresponding varieties. Let F be an algebraically closed
field. Consider the general linear group G L(n, F) of non-singular n × n matrices
with entries in F. Then G L(n, F) can be identified with the affine algebraic sub-
set {(A, (Det (A))−1 ) | A ∈ G L(n, F)} = V (Det ([X i j ])X − 1) of F n +1 . Thus,
2

G L(n, F) is an algebraic variety. Since the matrix multiplication and inversion on

G L(n, F) can be treated as polynomial maps on the entries of the matrices, the
operations on G L(n, F) are morphisms of varieties. Consequently, G L(n, F) is
an algebraic group. This algebraic group is called a general linear algebraic group
over F. A closed (with respect to the Zariski topology) subgroup G of G L(n, F)
is also an algebraic group which is termed as a linear algebraic group over F. Let
G ⊆ G L(n, F) be a connected linear algebraic group, where F is an algebraically
closed field. Let (G) denote the coordinate ring of G which consists of poly-
nomial maps on G. Then (G) is an algebra over F. For each x ∈ G, the right
multiplication Rx from G to G given by Rx (g) = gx is an isomorphism of G con-
sidered as a variety. Thus, for each x ∈ G, we have a map ρ(x) from (G) to itself
given by ρ(x)( f )(g) = f (gx). Evidently, ρ(x) ∈ G L((G)) = Aut F ((G)) for
each x ∈ G, and also ρ(x y) = ρ(x)ρ(y) for all x, y ∈ G. This gives us homomor-
phism ρ from G to G L((G)) = Aut F ((G)), and correspondingly, (G) is a
right F(G)-module, where F(G) denotes the group algebra over F. Now, consider
the Lie subalgebra Der ((G)) of gl((G)) consisting of derivations on (G). A
member of Der ((G)) need not be a G-homomorphism from (G) to (G). A
derivation d on (G) is a F(G)-homomorphism
 if doρ(x) = ρ(x)od for all x ∈ G.
Let (Der ((G)))G = End F(G) ((G)) Der ((G)) denote the set of all deriva-
tions on (G) which are F(G)-homomorphisms. Evidently, End F(G) ((G)) is a Lie
subalgebra of gl((G)). Hence (Der ((G)))G is a Lie subalgebra of gl((G)).
Treat the field F as a (G)-module by putting f α = f (e)α = α f (e) =
α f, f ∈ (G), and α ∈ F. This (G)-module F is denoted by Fe . A (G)-
derivation from (G) to Fe is called a point derivation of (G) at e. Recall (Definition
4.3.29, Algebra 3) that the tangent space Te (G) of G at e is the space Der ((G), Fe )
of point derivations of (G) at e. More explicitly, Te (G) is the F vector space of all
maps d from (G) to F satisfying the condition

d( f g) = d( f )g(e) + f (e)d(g)

for all f, g ∈ (G), where e denotes the identity of G.

Proposition 1.1.18 We have a natural F-isomorphism χ from (Der ((G)))G to
Der ((G), Fe ) given by χ(d)( f ) = d( f )(e).

Proof Let d be a member of (Der ((G)))G . Then the map χ(d) from (G) to F
given by χ(d)( f ) = d( f )(e) is a point derivation of (G) at e, since χ(d)( f f  ) =
d( f f  )(e) = (d( f ) f  + f d( f  ))(e) = d( f )(e) f  (e) + f (e)d( f  )(e) = χ
(d)( f ) f  (e) + f (e)χ(d)( f  ) for all f, f  ∈ (G). This gives us a natural linear
map χ from (Der ((G)))G to Der ((G), Fe ) defined by χ(d)( f ) = d( f )(e).
Suppose that χ(d) = χ(d  ), where d, d  ∈ (Der ((G)))G . Then d( f )(e) =
1.1 Definitions and Examples 17

d  ( f )(e) for all f ∈ (G). Since d and d  are G-module endomorphisms of (G),
d( f )(x) = xd( f (e)) = xd  ( f )(e) = d  ( f )(x) for all x ∈ G. This shows that
d( f ) = d  ( f ) for all f ∈ (G). It follows that χ is injective. Next, let φ be a member
of Der ((G), Fe ). Then φ( f f  ) = φ( f ) f  + f φ( f  ) for all f, f  ∈ (G). Define
a map d from (G) to itself by putting d( f )(x) = φ(ρ(x)( f )) = φ( f · x). Using
the fact that φ is a point derivation of (G) at e, it is easily observed that d is a
derivation of (G) which is a G-module homomorphism. Evidently, χ(d) = φ. .
The vector space isomorphism χ introduced above induces a Lie algebra structure
on Te (G). Te (G) with this Lie algebra structure is called the Lie algebra of G, and it
is denoted by L(G). Let φ be an algebraic homomorphism from an algebraic group G
to an algebraic group G  . Define a map dφe from L(G) = Te (G) = Der ((G), Fe )
to L(G  ) = Te (G  ) = Der ((G  ), Fe ) by putting dφe (d)( f ) = d( f oφ). It can be
easily seen that dφe is a Lie algebra homomorphism. There is no loss in adopting the
notation dφ for dφe . If φ is a homomorphism from G to G  and ψ is a homomorphism
from G  to G  , then d(ψoφ) = d(ψ)od(φ). This gives us a functor L from the
category ALG of connected algebraic groups over an algebraically closed field F
to the category L A F of Lie algebras over F which associates with each connected
algebraic group G the Lie algebra L(G) of G, and with each homomorphism φ from
G to G  , the Lie algebra homomorphism L(φ) = dφ.
For each g ∈ G, let i g denote the inner automorphism of G determined by g.
Then i g is an algebraic automorphism of G, and di g is a Lie automorphism of L(G).
This automorphism is denoted by Ad(g). We have a map Ad from G to the group
Aut (L(G)) of the automorphisms of the Lie algebra L(G) of G. It can be observed
that Ad(gh) = Ad(g)o Ad(h). Thus, Ad is a representation of G on its Lie algebra
L(G). This representation is called the adjoint representation of the algebraic group
G. We shall have occasions to discuss the adjoint representations.
Example 1.1.19 Consider the general linear algebraic group G L(n, F) =
V ({Det [X i j ]X − 1}) ⊆ AnF +1 . The coordinate ring (G L(n, F)) of G L(n, F) is
2

the F-algebra

F[X 11 , X 12 , · · · , X 1n , X 21 , X 22 , · · · , X 2n , · · · , X nn , X ]
.
I (G L(n, F))

Evidently, the map φ from (G L(n, F)) to the polynomial ring

F[X 11 , X 12 , · · · , X 1n , X 21 , X 22 , · · · , X 2n , · · · , X nn , Det [X i j ]−1 ]

defined by

φ( f [X 11 , X 12 , · · ·, X 1n , X 21 , X 22 , · · · , X 2n , · · · , X nn , X ] + I (G L(n, F))) =
f [X 11 , X 12 , · · · , X 1n , X 21 , X 22 , · · · , X 2n , · · · , X nn , Det [X i j ]−1 ]

is a natural isomorphism. The map η from Der ((G L(n, F)), FI ) to gl(n, F) is
defined by η(d) = [ai j ], where d(X i j ) = ai j can be easily seen to be a Lie algebra
isomorphism. Thus, L(G L(n, F)) = gl(n, F).
18 1 Lie Algebras

Next, consider the special linear group S L(n, F) = V ({Det [X i j ] − 1}). Clearly,

F[X 11 , X 12 , · · · , X 1n , X 21 , X 22 , · · · , X 2n ]
(S L(n, F)) = .
I (S L(n, F))

Let X i j denote the coset X i j + I (S L(n, F)). Let d ∈ Der ((S L(n, F)), FI ). Put
d(X i j ) = αidj ∈ F. Since d is a derivation

0 = d(1) = d(Det [X i j ]) = D(Det [X i j ])(I ) = i D(X ii )(I ) =

i d(X ii ) = i αii ,
d

where D is the derivation of F[X 11 , X 12 , · · · , X 1n , X 21 , X 22 , · · · , X 2n ] associ-

ated with d. This shows that the matrix [αidj ] is of trace 0. The map η from
Der ((S L(n, F)), FI ) to sl(n, F) given by η(d) = [αidj ] is easily seen to be a
Lie algebra isomorphism. Thus, L(S L(n, F)) = sl(n, F).
Lie Groups and Lie Algebras
Recall that a map f from an open subset U of Rn to R is a C r -map if for each k ≤ r
and an n-tuple (t1 , t2 , · · · , tn ) of nonnegative integers with t1 + t2 + · · · + tn = k,
the partial derivative
∂k f
∂ 1 x1 ∂ 2 x2 · · · ∂ tn xn
t t

exists and it is continuous on U . It is said to be a C ∞ -map if it is a C r -map for

each r ≥ 0. A map φ from an open subset U of Rn to Rm can be expressed as
φ = (φ1 , φ2 , · · · , φm ), where φi = pi oφ are maps from U to R. We say that φ is
a C ∞ -map if each φi is a C ∞ -map.
A Hausdorff topological space M (usually second countable) is called a topolog-
ical manifold or a locally Euclidean space if there is a nonnegative integer n such
that every point p ∈ M has an open neighborhood U p which is homeomorphic to
Rn . By the invariance of the domain (see Corollary 3.2.3, Algebra 3), n is uniquely
determined and it is called the dimension of the manifold. Thus, every nonempty
open subset U of Rn is a manifold of dimension n. Since S n − { p} is homeomorphic
to Rn for each p ∈ S n , S n is a manifold of dimension n.
Let M be a manifold of dimension n. A family  = {(Uα , h α ) | α ∈ } is called
a C r (C ∞ )-differential atlas on M if the following hold:
(i) {(Uα | α ∈ } is an open cover of M.
(ii) For each α ∈ , h α is a homeomorphism from Uα to Rn .
(iii)  each pair α,
For β ∈ , the restriction map h α oh −1 
β |h β (Uα Uβ ) from h β (Uα
Uβ ) to h α (Uα Uβ ) is a C r (C ∞ )-map.

A C r (C ∞ )-atlas  is said to compatible with a C r (C ∞ )-atlas   if    is also a
C r (C ∞ )-atlas. We have an obvious partial ordering on the set of all C r (C ∞ )-atlases.
A maximal C r (C ∞ )-atlas  on M is called C r (C ∞ )- differential structure on M.
If M admits a C r (C ∞ )-atlas , then the union of all atlases compatible with 
is a C r (C ∞ )-differential structure containing . Thus, every C r (C ∞ )-atlas  is
1.1 Definitions and Examples 19

contained in a unique C r (C ∞ )-differential structure. A manifold may admit several

differential structures. Indeed, Milnor showed the existence of 28 distinct differential
structures on S 7 . A C r (C ∞ )-manifold is a pair (M, ), where M is a manifold and
 is a C r (C ∞ ) differential structure. We shall be interested in C ∞ -manifolds.
Example 1.1.20 1. If U is an open subset of Rn , then U is union of a countable
family {Bn | n ∈ N} of open balls in Rn . For each n, we have a homeomorphism h n
from Bn to Rn . Clearly, {(Bn , h n ) | n ∈ N} is a C ∞ -atlas on U which determines a
unique C ∞ -differential structure  on U .
2. G L(n, R), being an open subset of Mn (R), is a C ∞ -manifold as described
above. Similarly, G L(n, C) being an open subset of Cn ≈ R(2n) is a C ∞ -manifold
2 2

of dimension 4n 2 . It can be shown (see Chap. 5) that all closed subgroups of G L(n, C)
are C ∞ -manifolds.
3. Consider S m ⊆ Rm+1 . Let U1 denote the open subset S m − { pn } and U2 the
open subset S m − { ps }, where pn = (0, 0, · · · , 0, 1) is the north pole and ps =
(0, 0, · · · , 0, −1) is the south pole of S m . Define a map h 1 from U1 to Rn by putting
h 1 (x1 , x2 , · · · , xm+1 ) = ( 1−xx1m+1 , 1−xx2m+1 , · · · , 1−xxmm+1 ) and a map h 2 from U2 to Rn
by putting h 2 (x1 , x2 , · · · , xm+1 ) = ( 1+xx1m+1 , 1+xx2m+1 , · · · , 1+xxmm+1 ). It can be easily
observed that {(U1 , h 1 ), (U2 , h 2 )} is a C ∞ -atlas which determines a unique C ∞ -
manifold structure on S n .
Let (M1 , 1 ) and M2 , 2 ) be C ∞ -manifolds of dimensions n and m, respectively.
Clearly, M1 × M2 is a manifold of dimension n + m. Further, 1 × 2 is a C ∞ -
atlas which determines a unique C ∞ -differential structure . The C ∞ -manifold
(M1 × M2 , ) is called the product of (M1 , 1 ) and (M2 , 2 ).
Let (M, ) be a C ∞ -manifold and U be an open subset of M. We say that a

map f from U to R is a C ∞ -function if f oh −1 α |h α (Uα U ) is a C
∞
function for each
−1 
(Uα , h α ) ∈ . This is also equivalent to say that f okβ |kβ (Vβ U ) is a C ∞ function
for each (Vβ , kβ ) ∈   , where   is a C ∞ -atlas defining the differential structure
. Let (M, ) and (N ,   ) be two C ∞ -manifolds. A continuous map f from M to
N is called a C ∞ -map if for any open subset V of N and a C ∞ -map h from V to
R, ho f is a C ∞ -map on f −1 (V ). This is equivalent to say that h α o f okβ−1 is a C ∞ -
map for each (Uα , h α ) ∈  and (Vβ , kβ ) ∈   . Clearly, composites of C ∞ -maps are
C ∞ -maps. Thus, we have the category of C ∞ -manifolds.
Let C ∞ (M) denote the set of all C ∞ -functions from M to R. Then C ∞ (M) is
a commutative algebra over R. A derivation on C ∞ (M) is called a vector field
on M. The set Der (C ∞ (M)) of all vector fields is a vector space over R with
respect to the obvious operations. If d, d  ∈ Der (C ∞ (M)), then it is easily seen that
[d, d  ] = dd  − d  d is also a member of Der (C ∞ (M)). In turn, Der (C ∞ (M)) is a
Lie subalgebra of gl(C ∞ (M)) over R.
Let (M, ) be a C ∞ -manifold and p ∈ M. Consider the set  p = {(U, φ) |
U is an open set containing p and φ is C ∞ map on U }. Define a rela-
tion ≈ on   p by putting (U, φ) ≈ (V, ψ) if there is an open subset W , p ∈
W ⊆ U V such that φ = ψ on W . It is clear that ≈ is an equivalence rela-
tion. Let C ∞ ( p) = {(U, φ) | (U, φ) ∈  p } denote the quotient set  p modulo
20 1 Lie Algebras

≈. C ∞ ( p) is called the set of germs of the C ∞ functions defined at p. Sup-

posethat (U, φ) = (U  , φ ) and (V, ψ) = (V  , ψ  ). It is easily observed that

 
(U V, φ + ψ) = (U V  , φ + ψ  ), (U V, φ · ψ) = (U  V  , φ · ψ  ), and
(U, aφ) = (U  , aφ ), a ∈ R. This makes C ∞ ( p) a commutative R-algebra. Further,
R is a C ∞ ( p)-module with the external product given by (U, φ) · a = φ( p)a =
a · φ( p) = a · (U, φ). The tangent space T p (M) of M at p is defined to be the
space Der (C ∞ ( p), R) of all point derivations of C ∞ ( p) at p.
If M is a complex manifold, then we can talk of an analytic atlas and an analytical
structure on M together with other related concepts by replacing C ∞ - maps with
analytic maps.
A group G together with C ∞ -structure on G is called a Real Lie Group if all the
group operations are C ∞ -maps. For example, G L(n, C), G L(n, R), S L(n, C), and
SU (n) are all examples of real Lie groups. Similarly, A group G together with an
analytic-structure on G is called a Complex Lie Group if all the group operations
are analytic-maps. For example, G L(n, C), S L(n, C) are all examples of complex
Lie groups. Note that SU (n) is not a complex Lie group (why?).
Let G be a real Lie group. Then for each g ∈ G, the left multiplication L g and
the right multiplication Rg are bijective C ∞ -maps whose inverses are also C ∞ -
maps on G. For each g ∈ G, we have a map ρ(g) from C ∞ (G) to C ∞ (G) defined by
(ρ(g)( f ))(x) = f (xg). Evidently, ρ(g) ∈ G L(C ∞ (G)) = AutR (C ∞ (G)) for each
g ∈ G, and also ρ(x y) = ρ(x)ρ(y) for all x, y ∈ G. This gives us a homomorphism
ρ from G to G L(C ∞ (G)) = AutR (C ∞ (G)), and correspondingly C ∞ (G) is a right
R(G)-module, where R(G) denotes the group algebra over R. Now, consider the
Lie subalgebra Der (C∞ (G)) of gl(C ∞ (G)) consisting of derivations on C ∞ (G). A
member of Der (C ∞ (G)) need not be a G-homomorphism from C ∞ (G) to C ∞ (G).
A derivation d on C ∞ (G) is a R(G)-homomorphism if and only  if doρ(x) = ρ(x)od
for all x ∈ G. Let (Der (C ∞ (G)))G = EndR(G) (C ∞ (G)) Der (C ∞ (G)) denote
the set of all derivations on C ∞ (G) which are R(G)-homomorphisms. Evidently,
EndR(G) (C ∞ (G)) is a Lie subalgebra of gl(C ∞ (G)). Hence (Der (C ∞ (G)))G is a
Lie subalgebra of gl(C ∞ (G)). The members of (Der (C ∞ (G)))G are called the G-
invariant vector fields. Each f ∈ C ∞ (G) determines an element (M, f ) of Ce∞ (G)
which we denote by f itself. As in Proposition 1.1.18, we obtain a natural R-
isomorphism χ from (Der (C ∞ (G)))G to Te (G) = Der (C ∞ (e), R) by putting
χ(d)( f ) = d( f )(e). It follows that Te (G) is a real Lie algebra. This Lie algebra
is called the Lie algebra of G and it is denoted by L(G). We can do the same for
complex Lie groups. In Chap. 5, we shall describe Lie algebras of linear Lie groups.

Remark 1.1.21 Hilbert’s fifth problem posed by Hilbert in 1900 was to see if a
Locally Euclidean group is a Lie group. Gleason and Montgomery settled it by prov-
ing that every Locally Euclidean Group (a group together with a manifold structure
such that the group operations are continuous) is a Lie group.

Let L be a real Lie algebra. Then L C = L ⊗R C is a Lie algebra over C with

obvious Lie product. This Lie algebra is called the complexification of L. Show that
the complexification of su(n) is G L(n, C) which is also the complexification of
gl(n, R).
1.1 Definitions and Examples 21

Exercises

1.1.1. Assuming that the characteristic of F is 0, show that all classical Lie algebras
in the families An , Bn , Cn , and Dn are perfect and center-less.
1.1.2. Determine the centers of all the Lie algebras discussed so far.
1.1.3. Show that for any Lie algebra L, ad(L) cannot be one dimensional.
1.1.4. Show that sl(3, F) is simple if and only if the characteristic of F is different
from 3.
1.1.5. Let T ∈ gl(n, F) which has all its eigenvalues in F and all are different.
Show that ad(T ) is diagonalizable.
1.1.6. Show that the derived algebra of gl(n, F) is sl(n, F).
1.1.7. Describe the derived algebras of d(n, F), t (n, F), and n(n, F). What are
their normalizers in gl(n, F).
1.1.8. Describe the Lie algebras of the algebraic groups D(n, F), T (n, F), and
U (n, F), where D(n, F) is the group of non-singular diagonal matrices,
T (n, F) is the group of non-singular upper triangular matrices, and U (n, F)
is the group of uni-upper triangular matrices.
1.1.9. Describe all three-dimensional Lie algebras over Z3 .
1.1.10. Show that (R3 , ×) is a Lie algebra over R. Is it simple? Describe its com-
plexification.
1.1.11. Describe all three-dimensional Lie algebras over Q. What are their com-
plexifications?
1.1.12. Establish the Leibnitz rule
n
δ n (x y) = n
Cr δr (x)δ n−r (y)
i=0

for a derivation δ on a Lie algebra over a field of characteristic 0.

1.1.13. Let δ be a nilpotent derivation of an algebra L over a field of characteristic
0. Suppose that δ n = 0. Show that

n−1 δ r
ex p(δ) =
r =0 r!
is an automorphism of L.
1.1.14. Describe the Lie subalgebra of sl(n, F) generated by {e12 , e21 }. Show that
the ideal of sl(n, F) generated by {e12 } is sl(n, F).
1.1.15. Let bn denote the number of bracket arrangements of weight n. Consider
∞
the power series b(t) = r =0 br t in Z[[t]].
r
√
Show that b(t)2 − b(t) + t
with b(0) = 0. Deduce that b(t) = 1 − 2 1−4t . Expanding this in power
series show that bn = n−1 1 2n−2
Cn .
1.1.16. Show that the Lie algebra L over C having a basis {x1 , x2 , x3 } subject to the
relation [x2 , x3 ] = x1 , [x3 , x1 ] = x2 , and [x1 , x2 ] = x3 is isomorphic
to sl(2, C).
1.1.17. Let us call a Lie algebra L to be a complete Lie algebra if Z (L) = {0}
and every derivation is an inner derivation. Let A be an ideal of L which
22 1 Lie Algebras

is complete as an algebra. Show that there is an ideal B of L such that

L = A ⊕ B.
1.1.18. Show that every non-abelian Lie algebra of dimension 2 is complete.

1.2 Universal Enveloping Algebras: PBW Theorem

This section is devoted to introducing and studying some universal objects in the
category of Lie algebras such as universal enveloping algebras, free Lie algebras, and
also to describing Lie algebras through presentations. We also establish the Poincaré–
Birkhoff–Witt (PBW) theorem together with some of its consequences. Every group
G is isomorphic to a subgroup of G L(V ) for some vector space V (consequence
of Cayley’s theorem). We shall show that every Lie algebra is isomorphic to a Lie
subalgebra of gl(V ) for some vector space V (not necessarily finite dimensional). In
case of groups, every finite group is isomorphic to a subgroup of a matrix group over
a field F. The corresponding analogue for the Lie algebras are the theorems of Ado
and of Iwasawa (Theorems 3.1.23 and 3.1.25) which assert that a finite-dimensional
Lie algebra over F is isomorphic to a Lie subalgebra of gl(n, F) for some n.
Let V be a vector space (not necessarily finite dimensional) over an arbitrary
field F. Recall (Sect. 7, Algebra 2) the tensor algebra T (V ) on V . Thus, as
a vector space, T (V ) = ⊕ ∞ n=0 ⊗ V , where ⊗ V = F, ⊗ V = V , and
n 0 1

⊗ V = V ⊗ V ⊗
n
· · · ⊗ V. For each m, n ≥ 0, we have a bi-linear product ηnm
n
from ⊗n V × ⊗m V to ⊗n+m V given by

ηnm (x1 ⊗ x2 ⊗ · · · ⊗ xn , x1 ⊗ x2 ⊗ · · · ⊗ xm ) = x1 ⊗ x2 ⊗ · · · ⊗ xn ⊗ x1 ⊗ x2 ⊗ · · · ⊗ xm .

These partial products can be extended by linearity to a unique associative bi-linear

product on T (V ). This makes T (V ) an associative F-algebra. The algebra T (V )
is called the tensor algebra of V . We have the tautological injective linear map i V
from V to ⊗1 V ⊆ T (V ) given by i V (v) = v. The pair (T (V ), i V ) has the following
universal property: () If φ is a linear map from V to an associative algebra A,
then there is a unique algebra homomorphism φ from T (V ) to A such that φoi V =
φ. Further, if (B, j) is a pair which also satisfies the universal property (), then
there is a unique algebra homomorphism j from T (V ) to B, and a unique algebra
homomorphism i V from B to T (V ) such that joi V = j and i V oj = i V . Thus,
i V o joi V = i V = IT (V ) oi V . From the universal property of (T (V ), i V ), i V o j =
IT (V ) . Similarly, joi V = I B . This shows that (T (V ), i V ) is the unique pair up to
an isomorphism which satisfies the universal property (). Consequently, if f is
a linear map from a vector space V to a vector space W , then there is a unique
algebra homomorphism T ( f ) from T (V ) to T (W ) such that T ( f )oi V = i W o f .
Thus, T defines a functor from the category of vector spaces over F to the category
1.2 Universal Enveloping Algebras: PBW Theorem 23

of associative algebras over F which is adjoint to the forgetful functor from the
category of associative algebras over F to the category of vector spaces over F.
T (V ) has a simpler description as a polynomial algebra: Let B be a basis of V
and let S be a set of symbols in bijective correspondence with B. Let F̂[S] denote
the polynomial algebra over F in non-commuting indeterminates belonging to S.
We have a linear map i from V to F̂[S] induced by the bijective correspondence
from the basis B to the set S of indeterminates. It is an easy observation that the pair
( F̂[S], i) also satisfies the universal property . In fact, F̂[S] is a free associative
algebra on the set S, and V is the free F-module on S.
Let I be the ideal of the tensor algebra T (V ) generated by {x ⊗ y − y ⊗ x |
x, y ∈ V }. The algebra T (V )/I is called the symmetric algebra of V and it is
denoted by S(V ). The natural quotient map from T (V ) to S(V ) will be denoted
by π. Clearly, S(V ) is an  associative and commutative
 algebraover F. Fur-
ther, I (V ) = ⊕ ∞ n=2 (I ⊗ n
V ) (note that I F = {0} = I ⊗1 V ). Thus,
∞
S(V ) =⊕ n=0 S (V ), where S (V ) = F, S (V ) = V , and S m (V ) =
m 0 1

⊗m V /(I ⊗m V ), m ≥ 2. S m (V ) is called the mth symmetric power of V . We

have the tautological linear map i V from V to S(V ). The pair (S(V ), i V ) has the
following universal property: () If (B, j) is a pair, where B is an associative and
commutative algebra and j is a linear map from V to B, then there is a unique homo-
morphism j from S(V ) to B such that joi = j. As in the case of tensor algebra, the
pair (S(V ), i V ) is unique up to isomorphism which satisfies the universal property
(). Evidently, S defines a functor from the category of vector spaces to the cate-
gory of commutative algebras over F which is adjoint to the forgetful functor. The
symmetric algebra of V has a simpler description as polynomial algebra F[S] in the
set S of commuting variables, where S is in bijective correspondence with a basis of
V.
Now, we introduce the adjoint functor to the functor from the category of associa-
tive algebras to the category of Lie algebras which associates with each associative
algebra A the associated Lie algebra A L (see Example 1.1.10).
Definition 1.2.1 A universal enveloping algebra of a Lie algebra L over a field
F is a pair (U (L), jL ), where U (L) is an associative algebra over F and jL is a
Lie algebra homomorphism from L to the associated Lie algebra U (L) L such that
it satisfies the following universal property: (  ) If (B, j) is also a pair, where B
is an associative algebra and j is a Lie algebra homomorphism from L to B L , then
there is a unique algebra homomorphism j from U (L) to B such that jojL = j.
As usual, the pair (U (L), jL ) is unique up to a natural isomorphism.
We show the existence of a universal enveloping algebra by constructing it. Let
T (L) denote the tensor algebra of L considered as a vector space. Let J denote the
ideal of T (L) generated by {x ⊗ y − y ⊗ x − [x, y] | x, y ∈ L} (note that (x ⊗ y −
y ⊗ x) belongs to the component ⊗2 L of T (L) and [x, y] belongs to the component
L of T (L)). Take U (L) to be T (L)/J and jL = νoi L , where ν is the quotient
map from T (L) to T (L)/J . Let B be an associative algebra and j be a linear map
from L to B such that j ([x, y]) = j (x) j (y) − j (y) j (x) for all x, y ∈ L. From the
universal property of (T (L), i L ), there is a unique algebra homomorphism φ from
24 1 Lie Algebras

T (L) to B such that φoi L = j. Since j ([x, y]) = j (x) j (y) − j (y) j (x), it follows
that φ(x ⊗ y − y ⊗ x − [x, y]) = 0. Hence φ induces a unique homomorphism ψ
from U (L) to B such that ψojL = j. This shows that the pair (U (L), jL ) is universal
enveloping algebra of L.
Let η be a homomorphism from a Lie algebra L to a Lie algebra L  . Then jL  oη
is a Lie algebra homomorphism from L to U (L  ) L . From the universal property of
the universal enveloping algebra, we have a unique algebra homomorphism U (η)
from U (L) to U (L  ) such that jL  oη = U (η)ojL . This defines a functor U from
the category of Lie algebras to the category of associative algebras. The functor U
is adjoint to the functor from the category of associative algebras to the category of
Lie algebras which associates with each associative algebra A the corresponding Lie
algebra A L .
Example 1.2.2 If L is an abelian Lie algebra, then [x, y] = 0 for all x, y ∈ L.
Thus, the ideal J is the same as I . Hence in this case, U (L) = S(L). If L is a
non-abelian Lie algebra of dimension 2, then there is a basis {x, y} of L such that
[x, y] = x. In this case, U (L) is isomorphic to F̂[X, Y ]/J , where F̂[X, Y ] is the
polynomial ring in non-commuting variables X and Y , and J is the principal ideal
of K̂ [X, Y ] generated by X Y − Y X − X = X (Y − 1) − Y X . Observe that in both
the cases jL is injective. Also observe that in this case {X r Y s + J | r ≥ 0, s ≥ 0}
forms a basis of U (L).

Theorem 1.2.3 Let L be a Lie algebra. Then the universal enveloping algebra
(U (L), jL ) of L satisfies the following properties:
1. The image jL (L) generates U (L) as an associative algebra.
2. Let A be an ideal of L. Let  denote the ideal of U (L) which is generated by
jL (A). Then the pair (U (L)/ Â, jˆL ) is the universal enveloping algebra of L/A,
where jˆL is given by jˆL (x + A) = jL (x) + Â.
3. There is a unique anti-automorphism ρ of U (L) such that ρ( jL (x)) = − jL (x)
for each x ∈ L. Further ρ2 = IU (L) .
4. We have a unique algebra homomorphism  from U (L) to the algebra U (L) ⊗ F
U (L) (note that U (L) ⊗ F U (L) is an associative algebra) such that

( jL (x)) = jL (x) ⊗ 1 + 1 ⊗ jL (x)

for each x ∈ L. The homomorphism  is called the diagonal homomorphism.

5. For each derivation d on L, we have a unique derivation d̂ on U (L) such that
d̂ojL = jL od.

Proof 1. Since T (L) is generated by i L (L) and jL = νoi L , it follows that U (L) is
generated by jL (L).
2. Evidently, the map jˆL from L/A to U (L)/ Â given by jˆL (x + A) = jL (x) + Â
is a linear map. Also
jˆL ([x + A, y + A])
= jˆL ([x, y] + A)
1.2 Universal Enveloping Algebras: PBW Theorem 25

= jL ([x, y]) + Â
= ( jL (x) jL (y) − jL (y) jL (x)) + Â
= [ jL (x) + Â, jL (y) + Â]
= [ jˆL (x + A), jˆL (y + A)].
This shows that jˆL is a Lie algebra homomorphism from L/A to (U (L)/ Â) L . Next,
let B be an associative algebra and let μ be a Lie algebra homomorphism from
L/A to B L . Then μ induces a Lie algebra homomorphism ν from L to B L given by
ν(x) = μ(x + A). Since (U (L), jL ) is a universal enveloping algebra of L, we have
a unique Lie algebra homomorphism μ̂ from U (L) to B such that μ̂ojL = ν. Hence
(μ̂ojL )(x) = ν(x) = μ(x + A) = 0 whenever x ∈ A. Hence μ̂( jL (A)) = 0.
Consequently, μ̂( Â) = 0. Thus, μ̂ induces a unique homomorphism μ from U (L)/ Â
to B such that μo jˆL = μ. It follows that (U (L)/ Â, jˆL ) is the universal enveloping
algebra of L/A.
3. Consider the universal enveloping algebra (U (L), jL ) of L. U (L) is also an
associative algebra with respect to the new product  given by u  v = v · u, where
· is the product in U (L). Let us denote this new associative Lie algebra by U (L) .
Let [, ] denote the associated Lie algebra structure. Consider the map η from L to
U (L) given by η(x) = − jL (x). Then η([x, y]) = − jL [x, y] = [ jL (y), jL (x)] =
[η(y), η(x)] = [η(x), η(y)] . Thus, η is a Lie algebra homomorphism from L to
(U (L) ) L . From the universal property of (U (L), jL ), we get a homomorphism ρ from
U (L) to U (L) such that ρoJ L = η = − jL . Evidently, ρ is the anti-automorphism
of U (L) such that ρ2 = I L .
4. Consider the associative algebra U (L) ⊗ F U (L) and the map φ from L to
U (L) ⊗ K U (L) given by φ(x) = J L (x) ⊗ 1 + 1 ⊗ jL (x). Clearly, φ is a linear
map. Further,

φ([x, y]) = jL ([x, y]) ⊗ 1 + 1 ⊗ jL ([x, y]) = ( jL (x) jL (y) − jL (y) jL (x))⊗
1 + 1 ⊗ ( jL (x) jL (y) − jL (y) jL (x)) = φ(x)φ(y) − φ(y)φ(x).

This means that φ is a Lie algebra homomorphism from L to (U (L) ⊗ F U (L)) L .

From the universal property of (U (L), jL ), we have a unique algebra homomorphism
 with the required property.
5. Let d be a derivation of the Lie algebra L. Consider the algebra M2 (U (L)) of
2 × 2 matrices with entries in the universal enveloping algebra U (L) of L. Consider
the map φ from L to M2 (U (L)) which is given by
 
jL (x) jL (d(x))
φ(x) = .
0 jL (x)

Clearly φ is a linear map. Also using the fact that d is a derivation and jL is a linear
map, it can be easily seen that φ[x, y] = φ(x)φ(y) − φ(y)φ(x). Hence from the
universal property of (U (L), jL ), there is a unique homomorphism φ from U (L)
to M2 (U (L)) such that φojL = φ. Since jL (L) generates U (L) and since the set
26 1 Lie Algebras

of upper triangular matrices with same diagonal entries is closed under the matrix
product, we have a unique map d̂ from U (L) to itself such that for u ∈ U (L),
 
u d̂(u)
φ(u) = .
0 u

The fact that φ is a homomorphism implies that d̂ is a derivation of U (L). Clearly,

d̂ojL = jL od. 

Our next aim is to determine the structure of U (L) (PBW theorem). We have the
filtration

T 0 (L) ⊆ T 1 (L) ⊆ T 2 (L) ⊆ · · · · · · T n (L) ⊆ T n+1 (L) ⊆ · · ·

m
of T (L) considered as a vector space, where T m (L) = ⊕ i=0 ⊗i L. Consequently,
we get a filtration

U 0 (L) ⊆ U 1 (L) ⊆ U 2 (L) ⊆ · · · · · · U n (L) ⊆ U n+1 (L) ⊆ · · ·

of U (L) considered as a vector space, where U m (L) = ν(T m ), ν being the quo-
tient map from T (L) to U (L). Evidently, U p (L)U q (L) ⊆ U p+q (L). Let m (L)
denote the vector space U m (L)/U m−1 (L). Put (L) = ⊕ ∞ m=0  (L). The
m

product in U (L) induces a bi-linear product from  (L) ×  (L) to  p+q (L)..
p q

These bi-linear products can be extended by linearity to an associative product

on (L) which makes (L) an associative algebra. For each m, we have a lin-
ear map μm from ⊗m L to m (L) given by μm (x) = ν(x) + U m−1 (L). Given
any u ∈ U m (L), there is an element x = xo + x1 + · · · + xm of T m (L) such that
ν(x) = u, where xi ∈ ⊗i L for each i. Evidently, μm (xm ) = ν(xm ) + U m−1 (L) =
ν(x) + U m−1 (L) = u + U m−1 (L). This shows that μm is a surjective linear map
from ⊗m (L) to m (L). This gives us a unique linear map μ from T (L) to (L) sub-
ject to μ|⊗m L = μm for each m. If x ∈ ⊗ p (L) and y ∈ ⊗q (L), then x y ∈ ⊗ p+q (L)
and μ(x y) = μ p+q (x y) = ν(x y) + U p+q−1 (L) = ν(x)ν(y) + U p+q−1 (L) =
(ν(x) + U p−1 (L))(ν(y) + U q−1 (L)) = μ(x)μ(y). This shows that μ is a surjective
algebra homomorphism from T (L) to (L). Consider the element x ⊗ y − y ⊗ x ∈
⊗2 L. Then ν(x ⊗ y − y ⊗ x) ∈ 2 (L). Also x ⊗ y − y ⊗ x − [x, y] ∈ J for all
x, y ∈ L. Further, μ(x ⊗ y − y ⊗ x) = μ2 (x ⊗ y − y ⊗ x) + U1 (L) = U1 (L),
since x ⊗ y − y ⊗ x = [x, y] ∈ U1 (L). It follows that x ⊗ y − y ⊗ x ∈ ker μ for
each pair x, y ∈ L. Hence I ⊆ ker μ. In turn, μ induces a unique surjective alge-
bra homomorphism μ from the symmetric algebra S(L) of L to (L) such that
μoπ = μ, where π is the natural quotient map from T (L) to S(L).
Theorem 1.2.4 (Poincaré–Birkhoff–Witt) The algebra homomorphism μ from S(L)
to (L) is an isomorphism.
Before proving the theorem, we derive some of its important consequences as
corollaries.
1.2 Universal Enveloping Algebras: PBW Theorem 27

 m of ⊗ L such that the quotient map ηm from

m
Corollary 1.2.5 Let W be a subspace
⊗ L to S (L) = ⊗ L/(I ⊗ L) restricted to W is an isomorphism. Then
m m m

U m (L) = ν(W ) ⊕ U m−1 (L).


Proof Under the hypothesis of the corollary, it is clear that ⊗m L = W ⊕ (I ⊗m L).
Thus, it follows from the PBW theorem that μm |W is an isomorphism from W to
m (L) = U m (L)/U m−1 (L). Consequently, U m (L) = ν(W ) ⊕ U m−1 (L). 

Corollary 1.2.6 The linear map jL from L to U (L) is an injective linear map which
is also a Lie homomorphism from L to U (L) L .

Proof The map η1 is a tautological isomorphism from L = ⊗1 L to S 1 (L). The

result follows from the above corollary, since U0 = {0}. 

Let L be a Lie algebra. Let {xλ | λ ∈ } be a basis of L, where (, ≤) is a well-

ordered index set. Such a basis exists because of the well-ordering theorem. For
each member  = (λ1 , λ2 , · · · , λm ) ∈ m , let x denote the element xλ1 ⊗ xλ2 ⊗
· · · ⊗ xλm of ⊗m L. We put x∅ = 1 for ∅ ∈ 0 , and for (λ) ∈ 1 , we denote x(λ)
by xλ . Evidently, X m = {x |  ∈ m } is a basis of ⊗m L. Next, let m denote the
subset {(λ1 , λ2 , · · · , λm ) ∈ m | λ1 ≤ λ2 ≤ · · · ≤ λm } of m . For m = 0, 0 =
{∅} and for m = 1, 1 = 1 . For each m and for each ordered  m-tuple  =
(λ1 , λ2 , · · · , λm ) ∈ m , the element xλ1 ⊗ xλ2 ⊗ · · · ⊗ xλm + (I ⊗m L) of S m (L)
is denoted by x . We put x(λ) = xλ . Thus, x = xλ1 xλ2 · · · xλm . By the convention,
x∅ = 1. More generally, the image of the element x ∈ L under the embedding i L
of L to S1 (L) ⊆ S(L) is denoted by x.
Let Wm be the subspace of ⊗m L generated by {x |  ∈ m }. Since {x |  ∈
m } is a basis of S m (L), the quotient map ηm |Wm from Wm to S m (L) is an iso-
morphism. It follows from Corollary 1.2.5 that U m (L) = ν(Wm ) ⊕ U m−1 (L) and
ν(X m ) = {x + J |  ∈ m } is a basis of ν(Wm ). Since

U 0 (L) ⊆ U 1 (L) ⊆ U 2 (L) ⊆ · · · ⊆ U n (L) ⊆ U n+1 (L) ⊆ · · ·

is a filtration of U (L), we obtain the following corollary:

∞ 
Corollary 1.2.7 m=1 ν(X m ) {1} is a basis
 of U (L). More explicitly, {xλ1 ⊗
xλ2 ⊗ · · · ⊗ xλm + J | λ1 ≤ λ2 ≤ · · · ≤ λm } {1} is a basis of U (L). 
The basis of U (L) described in the above corollary is termed as the P BW basis
of U (L).
Corollary 1.2.8 Let M be a Lie subalgebra of a Lie algebra L. Then U (i) is injec-
tive algebra homomorphism from U (M) to U (L), where i is the inclusion homomor-
phism. Further, U (L) is a free U (M)-module.

Proof Let X = {xλ | λ ∈ } be a basis of M, where (, ≤) is a well-ordered set.

Embed it into a basis {xλ | λ ∈  }, where  ⊆  . By the well-ordering theorem,
we have a well order on  − . In turn, we get a well order ≤ on  such that
≤ | =  and λ < μ for all λ ∈  and μ ∈  − . From the above corollary,
28 1 Lie Algebras

   
( ∞ 
m=1 ν(X m )) {1} is a basis of U (L) whereas ( ∞ m=1 ν(X m )) {1} is a basis
of U (M). Evidently, ν(X m ) ⊆ ν(X m ). This shows that U (M) can be realized as a
∞ 
 of U (L) and U (L) is a free U (M)-module with basis m=1 (ν(X m ) −
subalgebra
ν(X m )) {1}. 
Corollary 1.2.9 If L is a Lie algebra, then U (L) is without zero divisors.
Proof Since S(L) ≈ (L) and the symmetric algebra S(L), being isomorphic to
the polynomial algebra, is without zero divisors, it follows that the algebra (L)
is without zero divisors. We show that U (L) is without zero divisors. Let x, y
be members of U (L) such that x y = 0 with y = 0. Then there is the small-
est r ≥ 1 such that y ∈ U r (L) but y ∈ / U r −1 (L). Suppose that x ∈ U m (L). Then
r −1
(x + U m−1
(L))(y + U (L)) = 0 in (L). Since (L) is without zero divi-
sors and y + U r −1 (L) is a nonzero element, x + U m−1 (L) is zero in m (L). Hence
x ∈ U m−1 (L). Proceeding inductively, we see that x ∈ U 0 (L) = F. Since x y = 0,
x = 0. 
Consider the universal enveloping algebra (U (L), jL ) of L. For each u ∈ U (L),
we have the left multiplication map lu from U (L) to U (L) given by lu (v) = uv.
Clearly, lu ∈ End K (U (L)) = gl(U (L)). The map φ from U (L) to End K (U (L))
defined by φ(u) = lu is easily seen to be an algebra homomorphism. Since U (L)
has no nonzero divisors, φ is injective. Consequently, φojL is an injective Lie algebra
homomorphism from L to gl(U (L)) L . We have established the following analogue
of Cayley’s theorem for Lie algebras.
Corollary 1.2.10 Every Lie algebra is isomorphic to a Lie subalgebra of gl(V ) for
some vector space V . In the language of the representation theory, every Lie algebra
has a faithful representation. 
Observe that even if L is finite dimensional, the algebra U (L) need not be
finite dimensional. However, the theorem of Ado–Iwasava asserts that every finite-
dimensional Lie algebra is the Lie subalgebra of gl(V ) for some finite-dimensional
space V . Ado proved it for the characteristic 0 case, and Iwasava proved it for the
positive characteristic. The proofs of these results will follow in Chap. 3.
We still postpone the proof of the PBW theorem and use it further to introduce
the concept of free Lie algebras and that of the presentations of Lie algebras.
Definition 1.2.11 Let X be a set. A pair (L(X ), i X ), where L(X ) is a Lie algebra
over F and i X is a map from X to L(X ), is called a free Lie algebra on X if it satisfies
the following universal property: For any pair (B, j), where B is a Lie algebra over
F and j is a map from X to B, there is a unique Lie algebra homomorphism j from
L(X ) to B such that joi X = j.
As usual, the pair (L(X ), i X ) is unique up to isomorphism. For the existence, let V be
a vector space over F with X as a basis. Consider the tensor algebra (T (V ), i V ) on V .
Let L(X ) denote the Lie subalgebra of the Lie algebra (T (V )) L generated by i V (X ).
Observe that (T (V ), i X ) is the free associative algebra on X , where i X = i V | X .
We show that (L(X ), i X ) is a free Lie algebra on X . Let M be a Lie algebra and
1.2 Universal Enveloping Algebras: PBW Theorem 29

j be a map from X to M. Then j M oj is a map from X to the universal enveloping

algebra U (M) of M. Since T (V ) is free associative algebra on X , we get an algebra
homomorphism φ from T (V ) to U (M) such that φoi X = j M oj. In turn, φ is also a
Lie algebra homomorphism from T (V ) L to U (M) L . Since j M (M) is a Lie subalgebra
of U (M) L containing ( j M oj)(X ) and L(X ) is the Lie subalgebra of T (V ) L generated
by i X (X ), η = ( j M )−1 oφ| L(X ) is a Lie algebra homomorphism from L(X ) to M
such that ηoi X = j. This ensures that (L(X ), i X ) is a free Lie algebra on X . The
members of L(X ) are called the Lie elements in X .
In particular, if X is a set, W is a vector space over F, and · is a map from X × W to
W such that x · (αw + βw  ) = α(x · w) + β(x · w  ) for all w, w ∈ W and α, β ∈
F, then we have a unique L(X )-module structure on W subject to i X (x) · w = x · w.
Presentations of Lie Algebras
Let F be a fixed field. A presentation of a Lie algebra is a pair < X ; R >, where X
is a set and R is a set of Lie elements in X . Let L be a Lie algebra. A presentation
< X ; R > together with a map φ from X to L is called a presentation of L if the
induced homomorphism φ from L(X ) to L is surjective and the kernel of φ is the
ideal of L(X ) generated by R. The members of R are called the defining relations of
the presentation. Thus, < X ; ∅ > together with the map i X is a presentation of the
free Lie algebra L(X ) on X . The presentation < {x, y}; {x ⊗ y − y ⊗ x − x} >
is a presentation of a non-abelian two-dimensional Lie algebra. As in the case of
the presentation theory of groups, one can have the analogue of the word problem
and also the isomorphism problems in the theory of presentations of Lie algebras. In
general, a finite presentation may represent an infinite-dimensional Lie algebra. One
of the most important aims of the book is to have a faithful classification of finite-
dimensional semi-simple Lie algebras over an algebraically closed characteristic 0
in terms of presentations.
We state without proof the following analogue of the Nielsen–Schreier theorem.
Theorem 1.2.12 ( Širšov–Witt Theorem) Every subalgebra of a free Lie algebra is
free. 
Proof of the PBW Theorem
Now, we proceed for a proof of the PBW theorem. We establish a few lemmas for
this purpose.
We have a filtration

S0 (L) ⊆ S1 (L) ⊆ S2 (L) ⊆ · · · Sm (L) ⊆ Sm+1 (L) ⊆ · · ·

m
of S(L), where Sm (L) = ⊕ i=0 S i (L). Thus, S0 (L) = F and S1 (L) = F ⊕
S (L), where S (L) = L = {x | x ∈ L} is isomorphic to L as a vector space.
1 1

Lemma 1.2.13 Given aLie algebra L, there is a unique sequence { f m ∈ H om(L ⊗ F

Sm (L), S(L)) | m ∈ N {0}} of linear maps satisfying the following conditions:
1. f m | L⊗ F Sm−1 (L) = f m−1 for each m.
2. f m (xλ ⊗ x ) = xλ x , where  = (λ1 , λ2 , · · · , λm ) ∈ m and λ ≤ λ1 .
30 1 Lie Algebras

3. If  ∈ r , r ≤ m, then f m (xλ ⊗ x ) = xλ x + u r for some u r ∈ Sr (L).

4. If  ∈ m−1 , then f m ([xλ , xμ ] ⊗ x ) = f m (xλ ⊗ ( f m (xμ ⊗ x ))) − f m (xμ ⊗
( f m (xλ ⊗ x ))).

Proof We construct the sequence { f m | m ≥ 0} by the induction on m. The unique-

ness is also established by the induction. Indeed, for each m, we construct the
finite sequence { fr | r ≤ m} satisfying the conditions of the lemma. For m = 0,
S0 (L) = F, 0 = {∅}, x∅ = 1. We have the unique linear map f 0 from
L ⊗ S0 (L) to S(L) such that f 0 (xλ ⊗ x∅ ) = xλ x∅ = xλ . Indeed, it is given by
f 0 (x ⊗ α) = αx. Conditions 3 and 4 are vacuously satisfied. Uniqueness is evident
because of condition 2. For more clarity in the arguments, we construct f 1 also.
We need to define f 1 (xλ ⊗ xμ ) for all λ, μ ∈  which we can further extend by
linearity. Condition 2 forces us to define f 1 (xλ ⊗ xμ ) = xλ xμ whenever λ ≤ μ
and also f 1 (xλ ⊗ 1) = f 1 (xλ ⊗ x∅ ) = xλ x∅ = xλ . Thus, f 1 | L⊗S0 (L) = f 0 .
Suppose that μ < λ. Condition 3 requires the existence of a u 1 ∈ S1 (L) such that
f 1 (xλ ⊗ xμ ) = xλ xμ + u 1 . We use condition 4 to determine u 1 . By 2 and 4,

f 1 (xλ ⊗ xμ ) = f 1 (xλ ⊗ f 1 (xμ ⊗ x∅ )) = f 1 (xμ ⊗ f 1 (xλ ⊗ x∅ ) + f 1 ([xλ , xμ ] ⊗ x∅ ) =

f 1 (xμ ⊗ xλ ) + [xλ , xμ ] = xμ xλ + [xλ , xμ ] = xμ xλ + xλ xμ − xμ xλ = xλ xμ .

Thus, u 1 = 0 and f 1 (xλ ⊗ xμ ) = xλ xμ for all λ, μ ∈ . This completes the

construction of the sequence { f 0 , f 1 }.
Assume that the sequence { f 0 , f 1 , · · · , f m }, m ≥ 1 with the required conditions
has already been constructed. We construct f m+1 . Condition 1 forces us to define
f m+1 (x ⊗ u) = fr (x ⊗ u) whenever u ∈ Sr (L), r ≤ m. Thus, it is sufficient to
define f m+1 (xλ ⊗ x ) for all  = (λ1 , λ2 , · · · , λm+1 } ∈ m+1 and then extend it to
a linear map from L ⊗ Sm+1 (L) to S(L). If λ ≤ λ1 , then condition 2 forces us to put
f m+1 (xλ ⊗ x ) = xλ x . In general to ensure condition 3, f m+1 (xλ ⊗ x ) should be
xλ x + u m+1 for some u m+1 ∈ Sm+1 (L) to be determined. Suppose that λ1 < λ.
Let us denote the element (λ2 , λ3 , · · · , λm+1 ) of m by   . Since x ∈ Sm (L),
f m+1 (xλ1 ⊗ x ) = f m (xλ1 ⊗ x ) = xλ1 x = x . Again, since x ∈ Sm (L), by
the induction hypothesis, there is a unique u m ∈ Sm (L) such that f m+1 (xλ ⊗ x ) =
f m (xλ ⊗ x ) = xλ x + u m . Condition 4 forces us to put

f m+1 (xλ ⊗ x ) = xλ1 xλ x + f m (xλ1 ⊗ u m ) + f m ([xλ , xλ1 ] ⊗ x ).

This completes the forced construction of f m+1 . Conditions 1, 2, and 3 hold by the
construction. If   = (λ2 , λ3 , · · · λm+1 ) ∈ m and μ = λ1 < λ then also condition
4 follows from the construction of f m+1 as above. Since [xμ , xλ ] = −[xλ , xμ ],
condition 4 also holds when λ < μ ≤ λ2 . Since [xλ , xλ ] = 0, condition 4 also holds
for λ = μ. Suppose that λ2 < λ and also λ2 < μ. Denote (λ3 , λ4 , · · · λm+1 ) ∈ m−1
by   . Since λ2 ≤ λi for all i ≥ 3, by the induction assumption,

f m+1 (xλ2 ⊗ x ) = f m−1 (xλ2 ⊗ x ) = xλ2 x = x .

1.2 Universal Enveloping Algebras: PBW Theorem 31

Again using the induction hypothesis,

f m+1 (xμ ⊗ x ) = f m (xμ ⊗ f m (xλ2 ⊗ x )) =

f m (xλ2 ⊗ f m (xμ ⊗ x )) + f m ([xμ , xλ2 ] ⊗ x ) · · · (1.1)

Thus, since λ2 < μ,

f m+1 (xλ ⊗ f m+1 (xμ ⊗ x ))

= f m+1 (xλ ⊗ ( f m (xλ2 ⊗ f m (xμ ⊗ x )) + f m ([xμ , xλ2 ] ⊗ x )))
= f m+1 (xλ ⊗ ( f m (xλ2 ⊗ f m (xμ ⊗ x )))) + f m+1 (xλ ⊗ ( f m ([xμ , xλ2 ] ⊗ x )))
= f m+1 (xλ2 ⊗ f m (xλ ⊗ f m (xμ ⊗ x ))) + f m ([xλ , xλ2 ] ⊗ f m (xμ ⊗ x ))+
f m ([xμ , xλ2 ] ⊗ f m (xλ ⊗ x )) + f m ([xλ , [xμ , xλ2 ]] ⊗ x ) · · · (1.2)

Since λ2 < λ, we may interchange λ and μ to get the equation

f m+1 (xμ ⊗ f m+1 (xλ ⊗ x ))

= f m+1 (xλ2 ⊗ f m (xμ ⊗ f m (xλ ⊗ x ))) + f m ([xμ , xλ2 ] ⊗ f m (xλ ⊗ x ))+
f m ([xλ , xλ2 ] ⊗ f m (xμ ⊗ x )) + f m ([xμ , [xλ , xλ2 ]] ⊗ x ) · · · (1.3)

Subtracting 3 from 2, applying condition 4 for f m again, and using the Jacobi identity
for L, we arrive at

f m+1 (xλ ⊗ f m+1 (xμ ⊗ x )) − f m+1 (xμ ⊗ f m+1 (xλ ⊗ x )) = f m ([xλ , xμ ]⊗
f m (xλ2 x )) + f m (([xλ2 , [xλ , xμ ]] + [xλ , [xμ , xλ2 ]] + [xμ , [xλ2 , xλ ]]) ⊗ x ) =
f m ([xλ , xμ ] ⊗ f m (xλ2 x )).

Since λ2 ≤ λi for all i ≥ 3, f m (xλ2 ⊗ x ) = xλ2 x = x . Thus,

f m+1 (xλ ⊗ f m+1 (xμ ⊗ x  )) − f m+1 (xμ ⊗ f m+1 (xλ ⊗ x  )) = f m+1 ([xλ , xμ ] ⊗ x  ).

This verifies condition 4 also for f m+1 . The proof of the lemma is complete. 

As a consequence of the above lemma, we have the following:

Lemma 1.2.14 We have a representation ρ of L on S(L) such that the following
hold:
For each  = (λ1 , λ2 , · · · , λm ) in m ,

ρ(xλ )(x ) = xλ x

whenever λ ≤ λ1 , and in general

32 1 Lie Algebras

ρ(xλ )(x ) = xλ x + u m

for some u m ∈ Sm (L).

Proof Let x ∈ L and u ∈ S(L). Then there is a m ≥ 0 such that u ∈ Sm (L). If

u ∈ Sm (L) and also u ∈ Sn (L) with m ≤ n, then from Lemma 1.2.13, f n (x ⊗ u) =
f m (x ⊗ u). Define ρ(x)(u) = f m (x ⊗ u). This gives us a linear map ρ from L to
End F (S(L)). Condition 4 of Lemma 1.2.13 ensures that it is a Lie algebra homo-
morphism from L to (End F (S(L))) L = gl(S(L)) L . It is also clear that ρ satisfies
the required condition. 

Lemma 1.2.15 Let x = x0 + x1 + · · · + xm be a member of Tm (L) = ⊕ rm=0

⊗r L, where xr ∈ ⊗r L , r ≤ m. Suppose that x ∈ J , where J is the kernel of the
natural quotient map ν from T (L) to U (L). Then xm ∈ I , where I is the kernel of
the natural quotient map π from T (L) to S(L).

Proof For each  = (λ1 , λ2 , · · · , λm ) ∈ m , let x denote the member xλ1 ⊗ xλ2 ⊗
· · · ⊗ xλm of T (L). Evidently, π(x ) is the monomial xλ1 xλ2 · · · xλm in xλi (note that
S(L) is commutative polynomial algebra in {xλ | λ ∈ }). Clearly, {x |  ∈ m }
is a basis of ⊗m L. From Lemma 1.2.14, we have a Lie algebra homomorphism
ρ from L to gl(S(L)) L = (End F (S(L))) L satisfying the conditions described in
Lemma 1.2.14. Again, from the universal property of U (L), we have a unique algebra
homomorphism ρ from U (L) = T (L)/J to End F (S(L)) such that ρojL = ρ. In
turn, we obtain an algebra homomorphism ρ̂ from T (L) to End F (S(L)) such that
ρ̂oi L = ρ. It follows that J ⊆ K er ρ̂. By Lemma 1.2.14, ρ(xλ )(1) = ρ(xλ )(x∅ ) =
xλ and ρ(xλ )(x ) = xλ x + u m for some u m ∈ Sm (L). It follows that ρ(x)(1) is a
linear combination of monomials in xλ with the highest m degree terms representing
ρ(xm ). Since x ∈ J ⊆ K er ρ̂, ρ(x) = 0. In turn, ρ(x)(1) = 0. Consequently,
π(xm ) = 0. This shows that xm ∈ I . 

Proof of the PBW Theorem. We need to show that μ from S(L) to (L) is an
isomorphism. Recall the discussion before the statement of the PBW theorem. We
have already seen that μ is a surjective homomorphism. By the definition, μoπ = μ,
where μ is the surjective homomorphism from T (L) to (L). Let x be a member
of T (L) such that μ(x) = 0. We need to show that x ∈ I . Clearly, x = x0 + x1 +
· · · + xm belong to Tm (L) for some m, where xi ∈ ⊗i (L) for each i ≤ m. Recall that
μ(x) = μm (xm ) = ν(xm ) + Um−1 (L). Hence ν(xm ) ∈ Um−1 (L). Consequently,
there is a member xm ∈ Tm−1 (L) such that ν(xm ) = ν(xm ). In turn, xm − xm belongs
to J . From the previous lemma, xm ∈ I . Hence μ(x) = μm (xm ) = 0. This shows
that x ∈ I . 
Exercises

1.2.1. Suppose that two Lie algebras L and L  are isomorphic. Show that U (L) is
isomorphic to U (L  ). Is the converse true? Support.
1.2.2. Describe the structure of U (sl(2, C)).
1.2 Universal Enveloping Algebras: PBW Theorem 33

1.2.3. Suppose that L is a finitely generated Lie algebra. Show that U (L) is a
Noetherian ring. What about the converse? Hint: Use the Hilbert basis the-
orem.
1.2.4. Let L be a free Lie algebra. Show that U (L) is a free associative algebra.
1.2.5. Characterize Lie algebras for which U (L) is a semi-simple ring.
1.2.6. Describe the free Lie algebra on a singleton.
1.2.7. Let L(X ) be a free Lie algebra on X over a field of characteristic
 0. Sup-
pose that X contains m elements. Show that Dim (L(X ) U (L)n ) =
n
d|n μ(d)m , where μ is the Möbius function.
1 d
n
In the following exercises (1.2.8-1.2.12), we give another description of free
Lie algebras and also of presentations.
1.2.8. Let X = {xλ | λ ∈ } be a set and (X ) denote the set of all bracket
arrangements in X of different weights with 1 representing the empty bracket
arrangement. Define a product · in (X ) by taking 1 as the identity and by
putting

β r (xλ1 , xλ2 , · · · , xλr ) · β s (xδ1 , xδ2 , · · · , xδs ) =

β r +s (xλ1 , xλ2 , · · · , xλr , xδ1 , xδ2 , · · · , xδs ),

where β r +s = (βr βs ), λi , δ j ∈ . Show that (X ) is a free magma with

identity on X .
1.2.9. Let F((X )) denote the vector space over F with (X ) as a basis. Extend
the multiplication in (X ) to F((X )) by bi-linearity and then F((X ))
becomes an algebra over F with identity (recall that an algebra over F is a
vector space with a bi-linear product). Show that F((X )) is a free algebra
(with identity) on X .
1.2.10. An ideal of F((X )) is a subalgebra of F((X )) which is invariant under
left and right multiplications by members of (X ). Describe the ideal of
F((X )) generated by a single element of F((X )).
1.2.11. Let A be the ideal of F((X )) generated by the set

 = {u · u | u ∈ F((X ))} {u · (v · w) + v · (w · u) + w · (u · v) | u, v, w ∈ F((X ))}.

Show that the quotient F((X ))/A together with the map i from X to
F((X ))/A given by i(x) = x + A is a free Lie algebra on X . The mem-
bers of  are called the trivial relations.
1.2.12. Consider a pair < X ; R >, where X is a set and R ⊂ F((X )). The pair
< X ; R > together with a map f from X to a Lie algebra L is called a
presentation of L if the induced homomorphism f from F((X )) to L
is a surjective
 algebra homomorphism whose kernel is the ideal generated
by A R. Thus, < X, ∅ > is a presentation of the free Lie algebra on X .
The members of R are called the relators of the presentation. Describe the
Lie algebra having the presentation < X ; R >, where X = {x, y} and
R = {(x y) − x}.
34 1 Lie Algebras

1.3 Solvable and Nilpotent Lie Algebras

The terminology of solvability and nilpotency in Lie algebra has been borrowed
from the corresponding concepts of solvability and nilpotency in groups. Indeed,
there is an intrinsic relationship between solvability/nilpotency of groups and the
solvability/nilpotency of corresponding Lie algebras. We develop the theory in an
analogous manner.
Definition 1.3.1 Let L be a Lie algebra. Define the ideals L n of L inductively as
follows. Put L 0 = L, L 1 = [L , L]. Assuming that L n has already been defined,
define L n+1 = [L n , L n ]. Evidently, each L n is an ideal of L such that L n+1 ⊆ L n
and L n /L n+1 is an abelian Lie algebra. Indeed, each L n is a fully invariant subalgebra
of L in the sense that f (L n ) ⊆ L n for all endomorphism f of L. The series

L = L 0  L 1  · · ·  L n  L n+1  · · ·

is called the derived series of the Lie algebra L. L n is called the nth term of the
derived series. A Lie algebra L is said to be a solvable Lie algebra if L n = {0}
for some n. If L n = {0} but L n−1 = {0}, then we say that L is solvable of derived
length n.
Thus, a nontrivial abelian Lie algebra is solvable of derived length 1. A non-abelian
Lie algebra of dimension 2 is solvable of derived length 2. If the characteristic of F
is 2, then sl(2, F) is solvable. However, if the characteristic of F is different from
2, then sl(2, F) is not solvable. The classical Lie algebras An , Bn , Cn , and Dn are
non-solvable. The Lie algebra t (n, F) is solvable of derived length n, and n(n, F) is
solvable of derived length n − 1. Indeed, t (n, F)r is the Lie algebra of n × n matrices
[ai j ] for which ai j = 0 whenever j < i + r .
Proposition 1.3.2 A nontrivial solvable Lie algebra L is simple if and only if L is
a cyclic Lie algebra.

Proof Let L be a nontrivial solvable Lie algebra. Then [L , L] = L. Suppose that

L is simple, then [L , L] is {0}. Consequently, every subspace of L is an ideal.
Since L is simple, it has no nontrivial proper subspace. This means that L is one
dimensional. 

Corollary 1.3.3 If A is a maximal ideal of a nontrivial solvable Lie algebra L, then

A is of co-dimension 1.

Proof A is a maximal ideal of a solvable Lie algebra L if and only if L/A is simple
and solvable. Consequently, the dimension of L/A is 1. 

The proof of the following proposition is easy, and it is similar to the proof in the
case of groups. As such, it is left as an exercise.
1.3 Solvable and Nilpotent Lie Algebras 35

Proposition 1.3.4 (i) A Lie algebra L is solvable if and only if there is a finite chain

L = L 0  L 1  · · ·  L n−1  {0}

of subalgebras of L terminating to {0}, where L i+1 is an ideal of L i and L i /L i+1 is

abelian for each i ≤ n − 1.
(ii) Subalgebras and homomorphic images of solvable Lie algebras are solvable.
In particular, the quotient of a solvable Lie algebra is solvable.
(iii) If A is an ideal of L such that A and L/A are solvable, then L is solvable. In
particular, L is solvable if and only if ad(L) is solvable.
(iv) If A and B are solvable ideals of L, then A + B is also a solvable ideal of
L. In particular, every finite-dimensional Lie algebra contains the largest solvable
ideal. 
The largest solvable ideal of L is called the radical of L, and it is denoted by
Rad(L). A nontrivial finite-dimensional Lie algebra L is called a semi-simple Lie
algebra if the radical Rad(L) of L is trivial. Thus, a non-abelian simple Lie algebra
is semi-simple. If L is not solvable, then L/Rad(L) is semi-simple. We have a short
exact sequence
i ν
0 −→ Rad(L) → L → L/Rad(L) −→ 0.

We shall see (Theorem 1.5.16) that L is a semi-direct product Rad(L)  L/Rad(L).

As such, the structure theory of Lie algebras depends on the structure theory of
solvable Lie algebras, and even more importantly on the structure theory of semi-
simple Lie algebras.
Theorem 1.3.5 Let L be a solvable Lie subalgebra of gl(V ), where V is a nonzero
finite-dimensional vector space over an algebraically closed field F of characteristic
0. Then there is a nonzero vector v ∈ V which is an eigenvector of each member of
L.

Proof The proof is by the induction of dim L. If dim L = 0, then there is nothing to
do. Assume the induction hypothesis. Let A be a maximal ideal of L. From Corollary
1.3.3, A is of co-dimension 1 in L. If A = {0}, then L = F x is one dimensional,
where x is a nonzero endomorphism of V . Since F is algebraically closed, there
is an eigenvector v0 of x. Evidently, v0 is the eigenvector of each member of L,
and we are done. Suppose that A = {0}, and L = A ⊕ F x0 , where x0 is a nonzero
member of L. From the induction hypothesis, there is a nonzero vector v ∈ V such
that v is an eigenvector of each member x ∈ A. We get a linear functional λ on A
such that x(v) = λ(x)v for all x ∈ A. Let Vλ(x) denote λ(x)-eigenspace  of x in
V . Evidently, v ∈ Vλ(x) for all x ∈ A. Consider the subspace W = x∈A λ(x) of
V
V . Clearly, v ∈ W , and so W = {0}. Suppose that we are able to show that W is
invariant under all endomorphisms in L. Then, in particular, W is invariant under the
endomorphism x0 . Since F is algebraically closed, there is an eigenvector v0 of x0
in W . Since L = A ⊕ F x0 , it follows that v0 is an eigenvector of all members of L.
36 1 Lie Algebras

Now, we show that W is invariant under all transformations  in L. Let w ∈

W and y ∈ L. We need to show that y(w) ∈ W = x∈A λ(x) . Let x ∈ A.
V
Then x(y(w)) = y(x(w)) − [x, y](w) = y(λ(x)w) − [x, y](w) = λ(x)y(w) −
λ([x, y])w (note that [x, y] ∈ A). Thus, it is sufficient to show that λ([x, y]) = 0
for all y ∈ L and x ∈ A. Let n be the smallest positive integer such that the set
{w, y(w), y 2 (w), · · · , y n−1 (w)} is linearly independent. For each z ∈ A, z(w) =
λ(z)w, z(y(w)) = y(z(w)) − [y, z](w) = λ(z)y(w) − λ([y, z])w. Proceeding
inductively, we can show that z(y i (w)) is a linear combination of {w, y(w), y 2 (w),
· · · , y i (w)}. Again, for each i, 0 ≤ i ≤ n − 1, We show that
i−1
x(y i (w)) = λ(x)y i (w) + αij y j (w)
j=0

for some αij ∈ F. The proof is by the induction on i. It is evident for i = 0. Assume
it to be true for i. Now,

x(y i+1 (w)) = x(y(y i )(w)) = y(x(y i (w))) − [y, x](y i (w)) =
i−1
y(λ(x)y i (w) + αij y j (w)) − [y, x](y i (w)) =
j=0
i−1
λ(x)y i+1 (w) + αij y j+1 (w)) − [y, x](y i (w)).
j=0

Since [y, x] ∈ A, the result for i + 1 follows from the previous observation. It follows
that the subspace U of V generated by {w, y(w), y 2 (w), · · · , y n−1 (w)} is invariant
under x and the matrix representation of x on U is lower triangular with all diagonal
entries λ(x). Thus, the trace of x on U is nλ(x). Since U is invariant under x as well
as y, it is invariant under [x, y]. As such, the trace of [x, y] on U is zero. But already
the trace of [x, y] is nλ([x, y]). Since the field is of characteristic 0, it follows that
λ([x, y]) = 0 for all x ∈ A and y ∈ L. 

A Flag on a finite-dimensional vector space V of dimension n is a chain

{0} ⊆ V1 ⊆ V2 ⊆ · · · Vn−1 ⊆ Vn = V

of subspaces of V such that dimension of Vi is i for each i. The above flag is said to
be invariant under an endomorphism x in gl(V ) if x(Vi ) ⊆ Vi for all i.
Corollary 1.3.6 (Theorem of Lie) Let L be a solvable Lie subalgebra of gl(V ), where
V is a finite-dimensional vector space over an algebraically closed field. Then there
is a flag of V which is invariant under all members of L.

Proof The proof is by the induction on the dimension of V . If dim V = 1, then

there is nothing to do. Assume that the result is true for vector spaces of dimension
n. Let V be a vector space of dimension n + 1 and L is a solvable Lie subalgebra
of gl(V ). From the above theorem, there is a common eigenvector v0 of elements
1.3 Solvable and Nilpotent Lie Algebras 37

of L. Evidently, the subspace V1 generated by v0 is invariant under all elements

of L. Consider the vector space V  = V /V1 . Each x ∈ L determines an element
x̂ ∈ gl(V  ) given by x̂(v + V1 ) = x(v) + V1 . The map φ from L to gl(V  ) defined
by φ(x) = x̂ is easily seen to be a Lie algebra homomorphism. Evidently, φ(L) is
a solvable Lie subalgebra of gl(V  ). By the induction hypothesis, there is a flag

V1 /V1 ⊆ V2 /V1 ⊆ · · · ⊆ Vn /V1 ⊆ V  = V /V1

of V  which is invariant under φ(L). Clearly,

{0} ⊆ V1 ⊆ V2 ⊆ · · · Vn ⊆ Vn+1 = V

is a flag of V which is invariant under L. 

Corollary 1.3.7 Let L be a finite-dimensional solvable Lie algebra over an alge-
braically closed field F of characteristic 0. Then there is a chain

{0} ⊆ L 1 ⊆ L 2 ⊆ · · · L n−1 ⊆ L n = L

of ideals of L such that dim L i = i for each i.

Proof We have the Lie homomorphism ad from L to ad(L) ⊆ gl(L). Evidently,
ad(L) is solvable. From the previous corollary, there is a flag

{0} ⊆ L 1 ⊆ L 2 ⊆ · · · L n−1 ⊆ L n = L

of subspaces of L which is invariant under ad(L). This means that each L i is also
an ideal of L. 
Recall that t (n, F) is a solvable Lie algebra of derived length n. We have the
following corollary.
Corollary 1.3.8 Let V be an n-dimensional vector space over an algebraically
closed field F of characteristic 0. Then every solvable Lie subalgebra L of gl(V )
is isomorphic to a Lie subalgebra of t (n, F). Indeed, from Ado’s theorem, it further
follows that a Lie algebra is solvable if and only if it is isomorphic to a Lie subalgebra
of t (n, F) for some n.
Proof Let L be a solvable Lie subalgebra of gl(V ), where V is an n-dimensional
vector space over an algebraically closed field K . From Corollary 1.3.6, there is a
flag
{0} ⊆ V1 ⊆ V2 ⊆ · · · Vn−1 ⊆ Vn = V

of V which is invariant under the members of L. Select a basis {v1 , v2 , · · · , vn } of

V such that {v1 , v2 , · · · , vi } is a basis of Vi . Evidently, the matrix representation
of each member of L with respect to this basis is a member of t (n, K ). The result
follows. 
Random documents with unrelated
content Scribd suggests to you:
The Project Gutenberg eBook of The Story of a
Donkey
This ebook is for the use of anyone anywhere in the United States
and most other parts of the world at no cost and with almost no
restrictions whatsoever. You may copy it, give it away or re-use it
under the terms of the Project Gutenberg License included with this
ebook or online at www.gutenberg.org. If you are not located in the
United States, you will have to check the laws of the country where
you are located before using this eBook.

Title: The Story of a Donkey

Author: comtesse de Sophie Ségur

Editor: Charles F. Dole

Illustrator: E. H. Saunders

Translator: Charles Welsh

Release date: August 4, 2017 [eBook #55256]

Most recently updated: October 23, 2024

Language: English

Credits: Produced by Turgut Dincer and the Online Distributed

Proofreading Team at http://www.pgdp.net (This file was
produced from images generously made available by The
Internet Archive)

*** START OF THE PROJECT GUTENBERG EBOOK THE STORY OF A

DONKEY ***
“I ate up several cabbages.”
 THE
STORY OF A DONKEY
ABRIDGED FROM THE FRENCH OF

MADAME LA COMTESSE DE SÉGUR

By CHARLES WELSH

EDITED BY

CHARLES F. DOLE

ILLUSTRATED BY E. H. SAUNDERS

BOSTON, U.S.A.
D. C. HEATH & CO., PUBLISHERS
1904

Copyright, 1901,
By D. C. Heath & Co.
 Dedication.
TO MY PRESENT LITTLE MASTER, HARRY.

My Dear Little Master,

You have been kind to me, but you have spoken contemptuously of donkeys in general. I
want you to know better what sort of animals donkeys really are, and so I have written for
you this story of my life. You will see, my dear little Master, that we donkeys have been,
and still are, often badly treated by human beings. We are often very nice indeed; but I
must also confess that in my youth I sometimes behaved very badly, and you will see how I
was punished for it, and how unhappy I was, and how at last I repented, and how at last
my repentance changed me and gained for me the forgiveness of my friends and masters.
So, when you have read my history, you won’t say any more “as stupid as a donkey,” or “as
obstinate as a donkey,” but “as sensible as a donkey,” “as clever as a donkey,” or “as gentle
as a donkey.”

Hee-haw! my dear little Master, hee-haw! I hope you will never be as I was when I was
young.

I remain,

Your obedient servant,

NEDDY.
 PREFACE.
I do not recollect my childhood; I was probably as unhappy as the
rest of the little donkeys are; and no doubt as pretty and as graceful.
Certainly I was full of wit and intelligence, for, old as I am now, I
have more of both than most donkeys possess.
I have often outwitted some of my poor masters, who, being only
men, could not be expected to have the intelligence of a donkey,—
and I will begin my Memoirs with the story of a trick I once played in
the days of my youth.
 INTRODUCTION.
The author of this book was the daughter of that Count Rostopchine
who was governor of Moscow when it was burned in 1812, and
Napoleon was obliged in consequence to make his disastrous retreat
from that city. Born in 1799, Sophie de Rostopchine married, in
1821, the Count de Ségur, a son of one of the oldest and proudest
families of France. She was a very accomplished and lovable person,
and, as her writings attest, she was thoroughly in sympathy with the
ways and feelings of children.
She did not begin to publish her stories until she was fifty-seven
years of age, but between that date and the time of her death in
1874, she wrote and brought out a great many books for children.
The “Memoirs of a Donkey,” published in 1860, is one of the most
popular, wholesome, and entertaining of her books. It is longer in
the original than the version here given, as it contains a great
number of scenes that could interest only the boys and girls of
France; and there are many incidents in which the donkey scarcely
figures. We have, therefore, given in this book the story of Neddy,
the donkey. His adventures are interesting and amusing enough by
themselves, and as there has been nothing quite like them originally
written in English, we have included this retelling of the story in our
Home and School Classics.
 LIST OF ILLUSTRATIONS.
“I ate up several cabbages” Frontispiece
“I jumped clean over the hedge” 5
“I galloped along” 11
“We were placed along the wall” 16
“I galloped to the other end of the meadow” 24
“I drank up a bowl of cream” 27
“The boys shouted” 29
“I was ahead of all” 33
“I followed him all the way” 37
“Jack and Janie took the greatest care of me” 42
“The gentlemen and boys formed a broad line across the field” 45
“A sad procession” 49
“Muffles took the bunch of flowers” 54
“In order to clear a path” 58
“Along the edge of the ditch” 61
“I took the hat in my teeth” 67

AND TEN SMALLER ONES IN THE TEXT.

The Story of a Donkey.

 CHAPTER I.
Men, poor things, can’t be expected to be as wise as donkeys, and
therefore you probably do not know that there was a market in our
country-town every Tuesday. At this market vegetables were sold,
and butter, and eggs, and cheese, and fruit, and many other nice
things.
Tuesday was a miserable day for the poor donkeys, and especially
for me. I belonged to a farmer’s wife, and she was very severe and
ill-tempered. Just think! every week she used to load up my back
with all the eggs her hens laid, all the butter and cheese she made
from the milk of her cows, all the vegetables and fruit that were
ready for market out of her garden. Then she would get on the top
of all this and beat me with a hard, knotty stick because my poor
thin legs didn’t carry her to market with all that load as fast as she
liked. I trotted, I almost galloped, but that farmer’s wife whipped me
all the same. I used to get very angry at such cruelty and injustice. I
tried to kick her off, but I was loaded down too heavily, and so I
could only wobble about from side to side; but I did have the
satisfaction of knowing that she was well jolted. Then she would
growl, “Ah, you wretched animal! see if I don’t teach you to wobble!”
and she would beat me again till I could scarcely keep on my legs.
One day we reached the market-town in this way, and the baskets
with which my poor back had been nearly crushed were taken off
and set down upon the ground. My mistress hitched me to a post,
and went away to get her dinner. I was dying of hunger and thirst,
but nobody thought of offering me a single blade of grass or a drop
of water. While the farmer’s wife was away, I managed to get my
head close to the basket of vegetables, and made a dinner of the
cabbages and lettuces. I never tasted anything so good.
I had just finished the last cabbage and the last lettuce in that
basket when my mistress came back. She cried out when she saw
the empty basket, and I looked at her with such an impudent and
self-satisfied air, that she at once guessed that I was the culprit. I
won’t repeat to you the mean things she said to me. When she was
angry she used language which was enough to make me blush,
donkey as I am. So after heaping me with abuse, of which I took no
notice, she seized her stick and began to beat me so severely, that
at last I lost patience and launched out three kicks. The first kick
broke her nose and two teeth, the second sprained her wrist, and
the third knocked her flat.
A score of people at once set upon me and knocked me about. They
picked up my mistress and carried her away, leaving me fastened to
the post, by the side of which were spread out the things I had
brought to be sold in the market. I remained there a long while, and
finding that no one paid any more attention to me, I ate a second
basketful of excellent vegetables, and then with my teeth I gnawed
through the cord that tied me up, and quietly took the road home.
The people I passed on the way were astonished to see me all
alone.
“Look,” said one, “see that ass with the broken nose! He has run
away.”
“Then he has run away from prison,” said the other, and they all
began to laugh.
“He doesn’t carry a heavy load upon his back,” said a third.
“Certainly he has done some mischief,” a fourth one said.
“Catch him and we will put the little one upon his back,” said a
woman.
“He will carry you as well as the little boy,” answered her husband.
I, wishing to give a good opinion of my kindness and good will,
came gently towards the country woman and stopped near her to let
her mount upon my back.
“He doesn’t seem a bad sort!” said the man, helping his wife to the
saddle.
I smiled with pity on hearing this remark. Bad! as if a donkey kindly
treated were ever bad! We become angry, disobedient, and
obstinate only to revenge ourselves for the blows and injuries we
receive. When we are well treated we are good,—much better, in
fact, than many other animals.
I took the young woman and her little child of two years back to
their home; they stroked me, were very much pleased with me, and
would willingly have kept me.
But it was, I thought, not honest to stay with them. My masters had
bought me and I belonged to them. I had already broken my
mistress’s nose, teeth, and wrist, and had kicked her in the stomach.
I was sufficiently revenged.

“I jumped clean over the Hedge.

Seeing that the mother was going to give in to her little boy (who I
noticed was a spoiled child), I jumped to one side, and before the
mother could catch my bridle again, I ran away at a gallop and came
back to my home.
Mary, my mistress’s little girl, saw me come back.
“Hallo, here’s Neddy,” she said; “how early he is! Jim, come and take
off his pack-saddle.”
“That wretched donkey!” growled Jim; “always something to be
done for him! Why is he alone? I suspect he has run away from
mother.”
My saddle and bridle were taken off, and I galloped away to the
meadow. Suddenly I heard shrieks. I looked over the hedge, and
saw some men carrying my mistress home. Then I heard Jim say:—
“I say, father, I’m going to take the cart-whip, and I shall tie that
donkey to a tree, and then whip him till he can’t stand.”
“All right, my lad,” said my master, “but mind and don’t kill him, for
he cost money. I’ll sell him next fair-day.”
I shuddered when I heard this. There wasn’t a moment to be lost.
This time I did not care whether they lost their money or not. I
made a run and jumped clean over the hedge. Then I ran till I was
out of sight and hearing in the depths of a beautiful large forest,
where there was plenty of soft grass to eat, and plenty of sparkling
brooks to drink from.
 CHAPTER II.
Next day after, I thought over my good fortune. “Here I am saved,”
thought I; “they never will find me, and in a couple of days, when I
am quite rested, I will go farther on.”
Just then I heard the far-off barking of a dog; then of a second one;
and several minutes afterwards the yelling of a whole pack. Restless
and frightened, I got up and went towards a little brook that I had
noticed in the morning. I had hardly ventured into the water, when I
heard Jules saying to the dogs, “Go on, go on, dogs, search him out,
find this miserable donkey, and bring him back to me.”
I nearly fell down with fright, but I quickly remembered that if I
walked in the water the dogs could not follow my scent. So I began
to run in the brook which was fortunately bordered on both sides
with thick bushes.
I went on for a long time without stopping. The barking of the dogs
as well as the voice of Jules became fainter, until at last I heard
nothing more.
Breathless and exhausted I rested a minute to drink. I ate a few
leaves off the bushes. My legs were stiff with cold, but I did not dare
to get out of the water lest the dogs might come upon my scent
again. When I had rested a little, I set off again, always following
the brook until I got out of the forest. I then found myself in a
meadow where over fifty cattle were grazing. I lay down in the sun
in a corner of the field. The cattle paid no attention to me, so that I
could rest at my ease.
Towards evening two men came into the meadow. “Brother,” said the
taller of the two, “shall we take the cattle in to-night? They say there
are wolves in the woods.”
“Wolves! Who told you that nonsense?”
“People say that the donkey from the farm has been taken away and
eaten in the forest.”
“Bah! don’t believe it; the people of that farm are so wicked that
they have killed their donkey with bad treatment.”
“Then why do they say that the wolves have eaten him?”
“So that people won’t know that they have killed him.”
“We had better take in our cattle, all the same.”
“Do as you wish, brother; it is all one to me.”
I was in such fear of being seen that I lay in my corner and did not
stir; fortunately the grass was long and hid me; the cattle were not
on the side where I was. The men drove them towards the gate, and
then to the farm where their masters lived.
I was not afraid of wolves, because the donkey of whom they spoke
was myself, and because I had not seen the tail of a wolf in the
forest where I passed the night. So I slept beautifully and was
finishing my breakfast when the cattle came back to the meadow,
guarded by two large dogs.
I was looking at them, when one of the dogs saw me, and, barking
fiercely, ran towards me, his companion following. What should I do?
how could I escape them?
I flew towards the hedges surrounding the meadow, through which
ran the brook I had followed. I was fortunate enough to jump over
it, and I heard the voice of one of the men I had seen yesterday,
calling off his dogs.
I went on my way at my ease, and walked as far as another forest,
the name of which I don’t know. I must have gone more than ten
miles. I was saved; nobody knew me; and I could show myself
without fear of being taken back to my former masters.
But it began to grow cold, for winter was coming on, and I thought
it high time to look out for a comfortable home. I trotted on right
through the forest, and out at the other side, and after some days’
travelling, I arrived at a village that I had never seen or heard of
before. Here I felt I should be safe.
Just outside the village there stood a little cottage in a garden quite
by itself. It was very clean and neat. An old woman was sitting by
the door doing some needlework. I thought she looked both kind
and sad; so I went up to her, and put my head on her shoulder.
The good woman gave a shriek, and jumped up quickly.
I did not move, but lifted my face towards hers with a gentle and
pleading look.
“Poor thing!” she said at last; “you don’t look like a bad creature. If
you don’t belong to any one, you shall take the place of my poor
Greycoat, who died the other day of old age, and I shall still be able
to earn my living by taking my vegetables to market to sell. But,” she
added, with a sigh, “you’ve got a master somewhere, I’ll be bound.”
“Granny, whom are you talking to?” said a pleasant voice from the
house, and a nice little boy came out of the door. He was six or
seven years old, poorly but very neatly dressed. He looked at me,
half admiring, half afraid.
 “I galloped along.” P. 12.

“Granny, may I stroke him?” he said.

“Of course you may, George, my dear; but take care he doesn’t bite
you.”
The little boy stretched up his hand, but he was so short that he had
to stand on tiptoe before he could reach my back. I didn’t move, for
fear of frightening him; I only turned my head round, and licked his
hand.
“Oh, granny, granny! just see! what a dear donkey! he licked my
hand!”
“It’s very strange,” said George’s grandmother, “that he should be
here all by himself. Go to the village, my dear, and ask whether
anybody has lost a donkey. Perhaps his master is very anxious about
him.”
George set off at a run, and I trotted after him. When he saw me
come up, and then stand still by a mound on the roadside, he
climbed up on my back, and said, “Gee up!”
I galloped along, and George was enchanted. When we got to the
village inn, George cried, “Whoa back!” and I stopped immediately.
“What do you want, laddie?” said the innkeeper.
“Please, sir, do you know whose donkey this is?”
The innkeeper came out, and looked me all over. “No, my boy, he
isn’t mine, and he doesn’t belong to any one I know. Go and ask
farther on.”
So George went through the village asking the same question, but
nobody had ever seen me before. At last we went back to the good
old woman, who was still sitting with her work at the cottage door.
“So you can’t find his master, my dear? Very well, then, we may
keep him till he is claimed. He mustn’t stay out all night. Take him to
Greycoat’s shed, and give him some hay and a pail of water.”
The next morning George came to fetch me out of the shed, and
gave me some breakfast. Then he put on the halter, and took me
round to the cottage door. The old woman put a light pack-saddle on
my back and mounted. Then George brought her a basket of
vegetables, which she took on her knee, and we set off to market.
Nobody in this market-town had ever seen or heard of me, and I
came back joyfully to my new home.
I lived there for four years, and was very happy. I did my work well
and never did anybody any harm. I loved my good old mistress and
my little master. They never beat me or overworked me, and they
gave me the best food they could. We donkeys are not dainty. The
outside leaves of vegetables and plants that cows and horses won’t
eat, and hay and potato-peel and carrots and turnips, are all we
need.
 CHAPTER III.
On some days, however, I was not so happy, for my mistress, now
and then, hired me out to the children in the neighborhood. She was
not well off, and on the days when she had nothing for me to do she
was glad to earn something in this way.
And the people who hired me were not always good to me, as the
following story will show:—
There were six donkeys in a row in the courtyard; I was the
strongest and one of the most beautiful. Three little girls brought us
oats in a bucket; as I ate I listened to the children talking.
Charles.—Come along, let us choose our donkeys. I’ll begin by
taking this one (pointing me out with his finger).
“Yes, you always take what you think is the best,” said the six
children all at once. “We must draw lots.”
Charles.—How can we draw lots, Caroline? Can we put the donkeys
in a bag and draw them out like marbles?
Anthony.—Ha! ha! ha! The idea of donkeys in a bag! As if one could
not number them 1, 2, 3, 4, 5, 6. Put the numbers in a bag and
draw them out as they come.
“That is so, that is so,” cried the five others. “Ernest, make the
number slips while we write them on the backs of the donkeys.”
“These children are stupid,” said I to myself; “if they had the sense
of a donkey, instead of giving themselves the trouble of writing the
numbers on our backs, they would simply place us along the wall;
then the first would be 1, the second 2, and so on.”
Meanwhile Anthony had brought a large piece of charcoal. I was the
first. He wrote a large 1 on my flank; while he was writing 2 on that
of my comrade, I shook myself to show him that his invention was
not very clever. In a moment the 1 had disappeared. “Stupid,” cried
he, “I must begin again.”
While he was doing his number 1 over again, my comrade, who had
seen me, and who was clever, shook himself in his turn. There was
the 2 gone. Then Anthony began to get angry, and the others
laughed and teased him. I made a sign to my comrades to let them
go on, and then not one of us moved after being marked. Ernest
returned with the numbers in his pocket-handkerchief. Each one
drew.
While they were looking at their numbers, I made another sign to
my comrades, and we all shook ourselves vigorously. Charcoal and
numbers disappeared and all must be done over again! The children
were angry; Charles was triumphant and sneering; Ernest, Albert,
Caroline, Cecil and Louise, crying out against Anthony, who stamped
his feet. They began to quarrel with each other, and my comrades
and I began to bray.

"We were placed along the Wall."

The noise brought out the fathers and the mothers, to whom the
matter was explained. One of the fathers at last thought of placing
us in order along the wall. Then they made the children draw the
numbers. “One!” cried Ernest. It was I. “Two!” said Cecil. It was one
of my friends. “Three!” said Anthony, and so on until the last donkey
was drawn. “Now, let us go,” said Charles; “I will start first.” “Oh! I
shall catch up with you,” quickly answered Ernest. “I’ll wager you
won’t,” said Charles. “I’ll wager I will,” replied Ernest. Charles struck
his donkey and started at a gallop. Before Ernest had time to strike
me with his whip, I started also, and at a rate which enabled us
soon to overtake Charles and his donkey.
Ernest was delighted; Charles was furious and beat his donkey
repeatedly. Ernest had no need to beat me; I ran like the wind, and
passed Charles in a minute. I heard the others following, laughing
and shouting.
“Bravo! donkey Number 1! He runs like a horse.” Pride gave me
courage. I continued to gallop until we reached a bridge, where I
stopped suddenly, for I saw that one of the large boards of the
bridge was rotten. I did not wish to fall into the water with Ernest,
so I decided to return to the others, who were far, far behind us.
“Gee up! Gee up! Donkey!” said Ernest, “over the bridge, my friend,
over the bridge.” I would not go on; he hit me with his stick, but I
continued to walk towards the others. “Obstinate, stupid brute! will
you turn round and cross the bridge?” said he. I walked on towards
my comrades, and joined them in spite of the insults and blows of
this wicked boy.
“Why do you beat your donkey, Ernest?” cried Caroline; “he is very
good; he took you like lightning, and made you pass Charles.” “I
beat him because he would not cross the bridge,” said Ernest; “he
took it into his head to turn back.” “Nonsense! that was because he
was alone; now that we are all together, he will cross the bridge like
the others.”
“Unhappy children,” thought I, “they all will tumble into the river. I
must try to show them that there is danger;”—and again I started at
a gallop, running towards the bridge, to the great satisfaction of
Ernest and the other children, who shouted with joy. I galloped to
the bridge, but as soon as I got there, I stopped suddenly as if I
were afraid. Ernest was astonished, and urged me to go on. I drew
back with a frightened look which still more surprised Ernest. The
silly fellow saw nothing: the rotten board was, nevertheless, in plain
sight. Presently the others rejoined us and looked on laughingly at
Ernest’s attempts to make me cross. Then they got off their donkeys
and each one pushed me and beat me without pity. But I did not stir.
“Pull him by his tail,” cried Charles; “donkeys are so stubborn, that
when you want them to go backwards they go forwards.” Then they
tried to catch hold of my tail.
I defended myself by kicking, upon which they all beat me at once,
but in spite of this I would not move.
“Wait, Ernest,” said Charles, “I will go over first; your donkey will
certainly follow me.”
He started to go on; I put myself across the entrance to the bridge.
He made me turn by dint of blows.
“All right,” said I, “if this naughty boy wishes to drown himself, let
him. I did what I could to save him; let him drown if he wishes so
much to do so.”
No sooner had his donkey put his foot upon the rotten board than it
broke, and there was Charles and his donkey in the water!
There was no danger for my comrade, because, like all donkeys, he
could swim.
Charles struggled in frantic attempts to get out. “A stick! a stick!” he
cried. The children screamed and ran here and there. At last Caroline
found a long stick, picked it up and gave it to Charles, who seized it;
but his weight dragged down Caroline, who called out for help.
Ernest, Anthony, and Albert ran to her. At last the unhappy Charles,
who had by this time got more than he bargained for, was pulled out
of the water soaked from head to foot. When he was safe the
children began to laugh at his doleful appearance. Charles growing
angry, the children jumped upon their donkeys and advised him to
return to his home to change his clothes. Dripping wet he
remounted his donkey. I laughed to myself at his ridiculous
appearance.

The current had swept away his hat and his shoes; the water ran in
streams from his clothes; his soaked, wet hair stuck to his face, and
his furious look made him a thoroughly comical sight. The children
laughed; my comrades jumped and ran to express their joy. I must
add that Charles’s donkey was detested by all of us, because, unlike
most donkeys, he was quarrelsome, greedy, and stupid.
At last, Charles having disappeared, the children and my comrades
were calmed down. Every one stroked me and admired my
cleverness. We all set out again, I at the head of the party.
But these lively times were coming to an end. One day, George’s
father, who was a soldier, came home from the army and bought a
house in town. His mother and his little boy went to live with him,
and I was sold to a neighboring farmer.
 CHAPTER IV.
My new master was not a bad sort of man, but he had what I
thought was an unpleasant habit of making everybody work very
hard. He used to harness me to a little cart, and make me carry
earth and apples and wood and many other things. I began to grow
lazy; I didn’t enjoy going in harness, and I disliked market-days very
much. It wasn’t that they made me draw too heavy a load or that
they beat me, but I had to go without anything to eat from morning
till three or four o’clock in the afternoon. When the weather was hot
I nearly died of thirst, and yet I had to wait till everything was sold,
and my master had got all his money.
I wasn’t always good in those days. I wanted them to treat me
kindly, and as they didn’t, I began to think of revenge. You see that
donkeys are not always stupid, but you also see that I was growing
bad.
On market-days in the summer the people at the farm always got up
very early to cut the vegetables and gather the eggs and churn the
butter, while I was still lying out in the meadow. I used to watch all
this going on, knowing that at eight o’clock they would come and
fetch me to be harnessed to the cart.
One day I determined to play them a trick.
In the meadow I had noticed a deep ditch filled with thistles and
blackberry bushes. “Now,” I said to myself, “I’ll hide in that ditch, so
that when they come to fetch me there’ll be no donkey anywhere to
be seen.” So, as soon as I saw the cart being filled and the people
bustling about, I ran off to the side of the field, and lay down very
softly in the ditch, so that I was quite hidden by the bushes.
In a little while I heard one of the farm boys call me, and then run
looking about for me everywhere, and at last go back to the farm. In
a few minutes I heard the farmer himself say, “He must have got
Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.

More than just a book-buying platform, we strive to be a bridge

connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.

Join us on a journey of knowledge exploration, passion nurturing, and

personal growth every day!

ebookbell.com