Functional Analysis Master PDF

F UNCTIONAL A NALYSIS N OTES
(2011)
Mr. Andrew Pinchuck
Department of Mathematics (Pure & Applied)
Rhodes University
Contents
Introduction 1
1 Linear Spaces 2
1.1 Introducton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Subsets of a linear space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Subspaces and Convex Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Quotient Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.5 Direct Sums and Projections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.6 The Holder and Minkowski Inequalities . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2 Normed Linear Spaces 13

2.1 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Quotient Norm and Quotient Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3 Completeness of Normed Linear Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4 Series in Normed Linear Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.5 Bounded, Totally Bounded, and Compact Subsets of a Normed Linear Space . . . . . . . 26
2.6 Finite Dimensional Normed Linear Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.7 Separable Spaces and Schauder Bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3 Hilbert Spaces 36
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.2 Completeness of Inner Product Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.3 Orthogonality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.4 Best Approximation in Hilbert Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.5 Orthonormal Sets and Orthonormal Bases . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4 Bounded Linear Operators and Functionals 62

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.2 Examples of Dual Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.3 The Dual Space of a Hilbert Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5 The Hahn-Banach Theorem and its Consequences 81

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.2 Consequences of the Hahn-Banach Extension Theorem . . . . . . . . . . . . . . . . . . . 85
5.3 Bidual of a normed linear space and Reflexivity . . . . . . . . . . . . . . . . . . . . . . . 88
5.4 The Adjoint Operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
5.5 Weak Topologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
1
2011 F UNCTIONAL A NALYSIS ALP
6 Baires Category Theorem and its Applications 99

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.2 Uniform Boundedness Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
6.3 The Open Mapping Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
6.4 Closed Graph Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
2
Introduction
These course notes are adapted from the original course notes written by Prof. Sizwe Mabizela when
he last gave this course in 2006 to whom I am indebted. I thus make no claims of originality but have made
several changes throughout. In particular, I have attempted to motivate these results in terms of applications
in science and in other important branches of mathematics.
Functional analysis is the branch of mathematics, specifically of analysis, concerned with the study of
vector spaces and operators acting on them. It is essentially where linear algebra meets analysis. That is,
an important part of functional analysis is the study of vector spaces endowed with topological structure.
Functional analysis arose in the study of tansformations of functions, such as the Fourier transform, and in
the study of differential and integral equations. The founding and early development of functional analysis
is largely due to a group of Polish mathematicians around Stefan Banach in the first half of the 20th century
but continues to be an area of intensive research to this day. Functional analysis has its main applications in
differential equations, probability theory, quantum mechanics and measure theory amongst other areas and
can best be viewed as a powerful collection of tools that have far reaching consequences.
As a prerequisite for this course, the reader must be familiar with linear algebra up to the level of a
standard second year university course and be familiar with real analysis. The aim of this course is to
introduce the student to the key ideas of functional analysis. It should be remembered however that we
only scratch the surface of this vast area in this course. We examine normed linear spaces, Hilbert spaces,
bounded linear operators, dual spaces and the most famous and important results in functional analysis
such as the Hahn-Banach theorem, Baires category theorem, the uniform boundedness principle, the open
mapping theorem and the closed graph theorem. We attempt to give justifications and motivations for the
ideas developed as we go along.
Throughout the notes, you will notice that there are exercises and it is up to the student to work through
these. In certain cases, there are statements made without justification and once again it is up to the student
to rigourously verify these results. For further reading on these topics the reader is referred to the following
texts:
G. B ACHMAN , L. NARICI , Functional Analysis, Academic Press, N.Y. 1966.

E. K REYSZIG, Introductory Functional Analysis, John Wiley & sons, New York-Chichester-Brisbane-
Toronto, 1978.
G. F. S IMMONS , Introduction to topology and modern analysis, McGraw-Hill Book Company, Sin-
gapore, 1963.
A. E. TAYLOR , Introduction to Functional Analysis, John Wiley & Sons, N. Y. 1958.
I have also found Wikipedia to be quite useful as a general reference.
1
Chapter 1
Linear Spaces
1.1 Introducton
In this first chapter we review the important notions associated with vector spaces. We also state and prove
some well known inequalities that will have important consequences in the following chapter.
Unless otherwise stated, we shall denote by R the field of real numbers and by C the field of complex
numbers. Let F denote either R or C.
1.1.1 Definition
A linear space over a field F is a nonempty set X with two operations
C W X X !X (called addition); and
W FX !X (called multiplication)
satisfying the following properties:
[1] x C y 2 X whenever x; y 2 X ;
[2] x C y D y C x for all x; y 2 X ;
[3] There exists a unique element in X , denoted by 0, such that x C 0 D 0 C x D x for all x 2 X ;
[4] Associated with each x 2 X is a unique element in X , denoted by x , such that x C . x/ D
x C x D 0;
[5] .x C y/ C z D x C .y C z/ for all x; y; z 2 X ;
[6] x 2 X for all x 2 X and for all 2 F;
[7] .x C y/ D x C y for all x; y 2 X and all 2 F;
[8] . C / x D x C x for all x 2 X and all ; 2 F;
[9] ./ x D . x/ for all x 2 X and all ; 2 F;
[10] 1 x D x for all x 2 X .
We emphasize that a linear space is a quadruple .X; F; C; / where X is the underlying set, F a field, C
addition, and multiplication. When no confusion can arise we shall identify the linear space .X; F; C; /
with the underlying set X . To show that X is a linear space, it suffices to show that it is closed under
addition and scalar multiplication operations. Once this has been shown, it is easy to show that all the other
axioms hold.
2
1.1.2 Definition
A real (resp. complex) linear space is a linear space over the real (resp. complex) field.
A linear space is also called a vector space and its elements are called vectors.
1.1.3 Examples
[1] For a fixed positive integer n, let X D Fn D fx D .x1 ; x2 ; : : : ; xn / W xi 2 F; i D
1; 2; : : : ; ng the set of all n-tuples of real or complex numbers. Define the operations
of addition and scalar multiplication pointwise as follows: For all x D .x1 ; x2 ; : : : ; xn /;
y D .y1 ; y2 ; : : : ; yn / in Fn and 2 F,
xCy D .x1 C y1 ; x2 C y2 ; : : : ; xn C yn /
x D .x1 ; x2; : : : ; xn/:
Then Fn is a linear space over F.
[2] Let X D Ca; b D f x W a; b ! F j x is continuous g. Define the operations of addition and
scalar multiplication pointwise: For all x; y 2 X and all 2 R, define

.x C y/.t/ D x.t/ C y.t/ and
for all t 2 a; b:
. x/.t/ D x.t/
Then Ca; b is a real vector space.
Sequence Spaces: Informally, a sequence in X is a list of numbers indexed by N. Equivalently,
a sequence in X is a function x W N ! X given by n 7! x.n/ D xn . We shall denote a
sequence x1 ; x2 ; : : : by
x D .x1 ; x2 ; : : :/ D .xn /1
1 :
[3] The sequence space s. Let s denote the set of all sequences x D .xn /11 of real or complex
numbers. Define the operations of addition and scalar multiplication pointwise: For all x D
.x1 ; x2 ; : : :/, y D .y1 ; y2 ; : : :/ 2 s and all 2 F, define
xCy D .x1 C y1 ; x2 C y2 ; : : :/
x D .x1 ; x2; : : :/:
Then s is a linear space over F.
[4] The sequence space `1 . Let `1 D `1 .N/ denote the set of all bounded sequences of real or
complex numbers. That is, all sequences x D .xn /1
1 such that
sup jxi j < 1:

i2N
Define the operations of addition and scalar multiplication pointwise as in example (3). Then
`1 is a linear space over F.
[5] The sequence space `p D `p .N/; 1 p < 1. Let `p denote the set of all sequences
x D .xn /1
1 of real or complex numbers satisfying the condition
1
X
jxi jp < 1:
iD1
Define the operations of addition and scalar multiplication pointwise: For all x D .xn /, y D
.yn / in `p and all 2 F, define
xCy D .x1 C y1 ; x2 C y2 ; : : :/
x D .x1 ; x2; : : :/:
3
Then `p is a linear space over F.

Proof. Let x D .x1 ; x2 ; : : :/, y D .y1 ; y2 ; : : :/ 2 `p . We must show that x C y 2 `p . Since,
for each i 2 N,
jxi C yi jp 2 maxfjxi j; jyi jgp 2p maxfjxi jp ; jyi jp g 2p .jxi jp C jyi jp / ;
it follows that !
1
X 1
X 1
X
p p p p
jxi C yi j 2 jxi j C jyi j < 1:
iD1 iD1 iD1
Thus, x C y 2 `p . Also, if x D .xn / 2 `p and 2 F, then

1
X 1
X
jxi jp D jjp jxi jp < 1:
iD1 iD1
That is, x 2 `p .
[6] The sequence space c D c.N/. Let c denote the set of all convergent sequences x D .xn /1
1 of
real or complex numbers. That is, c is the set of all sequences x D .xn /1
1 such that lim xn n!1
exists. Define the operations of addition and scalar multiplication pointwise as in example
(3). Then c is a linear space over F.
[7] The sequence space c0 D c0 .N/. Let c0 denote the set of all sequences x D .xn /11 of real
or complex numbers which converge to zero. That is, c0 is the space of all sequences
x D .xn /1
1 such that lim xn D 0. Define the operations of addition and scalar multiplication
n!1
pointwise as in example (3). Then c0 is a linear space over F.
[8] The sequence space `0 D `0 .N/. Let `0 denote the set of all sequences x D .xn /11 of real or
complex numbers such that xi D 0 for all but finitely many indices i . Define the operations
of addition and scalar multiplication pointwise as in example (3). Then `0 is a linear space
over F.
4
1.2 Subsets of a linear space

Let X be a linear space over F; x 2 X and A and B subsets of X and 2 F. We shall denote by
xCA WD fx C a W a 2 Ag;
ACB WD fa C b W a 2 A; b 2 Bg;
A WD fa W a 2 Ag:
1.3 Subspaces and Convex Sets

1.3.1 Definition
A subset M of a linear space X is called a linear subspace of X if
(a) x C y 2 M for all x; y 2 M , and
(b) x 2 M for all x 2 M and for all 2 F.
Clearly, a subset M of a linear space X is a linear subspace if and only if M C M M and M M

for all 2 F.
1.3.2 Examples
[1] Every linear space X has at least two distinguished subspaces: M D f0g and M D X .
These are called the improper subspaces of X . All other subspaces of X are called the
proper subspaces.
[2] Let X D R2 . Then the nontrivial linear subspaces of X are straight lines through the origin.
[3] M D fx D .0; x2; x3; : : : ; xn / W xi 2 R; i D 2; 3; : : : ; ng is a subspace of Rn .
[4] M D fx W 1; 1 ! R; x continuous and x.0/ D 0g is a subspace of C 1; 1.
[5] M D fx W 1; 1 ! R; x continuous and x.0/ D 1 g is not a subspace of C 1; 1.
[6] Show that c0 is a subspace of c.
1.3.3 Definition
Let K be a subset of a linear space X . The linear hull of K , denoted by lin.K/ or span.K/, is the
intersection of all linear subspaces of X that contain K .
The linear hull of K is also called the linear subspace of X spanned (or generated) by K.
It is easy to check that the intersection of a collection of linear subspaces of X is a linear subspace of
X . It therefore follows that the linear hull of a subset K of a linear space X is again a linear subspace of X .
In fact, the linear hull of a subset K of a linear space X is the smallest linear subspace of X which contains
K.
1.3.4 Proposition
Let K be a subset of a linear space X . Then the linear hull of K is the set of all finite linear combinations
of elements of K . That is,
8 9
<X n =
lin.K/ D j xj j x1 ; x2 ; : : : ; xn 2 K; 1; 2 ; : : : ; n 2 F; n 2 N :
: ;
jD1
5
Proof. Exercise.
1.3.5 Definition
[1] A subset fx1 ; x2; : : : ; xn g of a linear space X is said to be linearly independent if the equation
1 x 1 C 2 x 2 C C n x n D 0
only has the trivial solution 1 D 2 D D n D 0. Otherwise, the set fx1 ; x2 ; : : : ; xn g is

linearly dependent.
[2] A subset K of a linear space X is said to be linearly independent if every finite subset fx1 ; x2 ; : : : ; xn g
of K is linearly independent.
1.3.6 Definition
If fx1 ; x2 ; : : : ; xn g is a linearly independent subset of X and
X D linfx1; x2 ; : : : ; xn g, then X is said to have dimension n. In this case we say that fx1; x2 ; : : : ; xn g
is a basis for the linear space X . If a linear space X does not have a finite basis, we say that it is infinite-
dimensional.
1.3.7 Examples
[1] The space Rn has dimension n. Its standard basis is fe1 ; e2 ; : : : ; en g, where, for each
j D 1; 2; : : : ; n, ej is an n-tuple of real numbers with 1 in the j -th position and zeroes
elsewhere; i.e.,
ej D .0; 0; : : : ; 1; 0; : : : ; 0/; where 1 occurs in the j -th position.
[2] The space Pn of polynomials of degree at most n has dimension n C 1. Its standard basis is
f1; t; t 2 ; : : : ; t n g.
[3] The function space Ca; b is infinite-dimensional.
[4] The spaces `p , with 1 p 1, are infinite-dimensional.
1.3.8 Definition
Let K be a subset of a linear space X . We say that
(a) K is convex if x C .1 /y 2 K whenever x; y 2 K and 2 0; 1;
(b) K is balanced if x 2 K whenever x 2 K and jj 1;
(c) K is absolutely convex if K is convex and balanced.
1.3.9 Remark
[1] It is easy to verify that K is absolutely convex if and only if x C y 2 K whenever x; y 2 K
and jj C jj 1.
[2] Every linear subspace is absolutely convex.
1.3.10 Definition
Let S be a subset of the linear space X . The convex hull of S, denoted co.S/, is the intersection
of all convex sets in X which contain S.
Since the intersection of convex sets is convex, it follows that co.S/ is the smallest convex set which
contains S. The following result is an alternate characterization of co.S/.
6
1.3.11 Proposition
Let S be a nonempty subset of a linear space X . Then co.S/ is the set of all convex combinations of
elements of S . That is,
8 9
<X n n
X =
co.S/ D j xj j x1 ; x2 ; : : : ; xn 2 S; j 0 8 j D 1; 2; : : : ; n; j D 1; n 2 N :
: ;
jD1 jD1
Proof. Let C denote the set of all convex combinations of elements of S. That is,
8 9
<X n n
X =
C D j xj j x1 ; x2; : : : ; xn 2 S; j 0 8 j D 1; 2; : : : ; n; j D 1; n 2 N :
: ;
jD1 jD1
n
X m
X n
X
Let x; y 2 C and 0 1. Then x D i xi ; y D i yi , where i ; i 0, i D 1,
1 1 1
m
X
i D 1, and xi ; yi 2 S. Thus
1
n
X m
X
x C .1 /y D i xi C .1 /i yi
1 1
is a linear combination of elements of S, with nonnegative coefficients, such that

n
X m
X n
X m
X
i C .1 /i D i C .1 / i D C .1 / D 1:
1 1 1 1
That is, x C .1 /y 2 C and C is convex. Clearly S C . Hence co.S/ C .
We now prove the inclusion C co.S/. Note that, by definition, S co.S/. Let x1 ; x2 2 S,
1 0; 2 0 and 1 C 2 D 1. Then, by convexity of co.S/, 1x1 C 2 x2 2 co.S/. Assume that
n 1
X n
X1
i xi 2 co.S/ whenever x1 ; x2; : : : ; xn 1 2 S, j 0, j D 1; 2; : : : ; n 1 and j D 1. Let
1 jD1
Xn
x1 ; x2 ; : : : ; xn 2 S and 1 ; 2; : : : ; n be such that j 0, j D 1; 2; : : : ; n and j D 1. If
jD1
n 1
X n
X n
X1 j
j D 0, then n D 1. Hence j xj D n xn 2 co.S/. Assume that D j > 0. Then 0
jD1 1 jD1
n 1
X X j n 1
j
for all j D 1; 2; : : : ; n 1 and D 1. By the induction assumption, xj 2 co.S/. Hence

jD1 jD1
0 1
n
X n 1
X j
j xj D @ xj A C nxn 2 co.S/:

jD1 jD1
Thus C co.S/.
1.4 Quotient Space

Let M be a linear subspace of a linear space X over F. For all x; y 2 X , define
x y.mod M / x y 2 M:
7
It is easy to verify that defines an equivalence relation on X .

For x 2 X , denote by
x WD fy 2 X W x y.mod M /g D fy 2 X W x y 2 M g D x C M;
the coset of x with respect to M . The quotient space X=M consists of all the equivalence classes x,
x 2 X . The quotient space is also called a factor space.
1.4.1 Proposition
Let M be a linear subspace of a linear space X over F. For x; y 2 X and 2 F, define the operations
x C y D x C y and x D x:
Then X=M is a linear space with respect to these operations.
Proof. Exercise.
Note that the linear operations on X=M are equivalently given by: For all x; y 2 X and 2 F,
.x C M / C .y C M / D x C y C M and .x C M / D x C M:
1.4.2 Definition
Let M be a linear subspace of a linear space X over F. The codimension of M in X is defined as the
dimension of the quotient space X=M . It is denoted by codim.M / D dim.X=M /.
Clearly, if X D M , then X=M D f0g and so codim.X / D 0.
1.5 Direct Sums and Projections

1.5.1 Definition
Let M and N be linear subspaces of a linear space X over F. We say that X is a direct sum of M and N
if
X D M C N and M \ N D f0g:
If X is a direct sum of M and N , we write X D M N . In this case, we say that M (resp. N ) is an
algebraic complement of N (resp. M ).
1.5.2 Proposition
Let M and N be linear subspaces of a linear space X over F. If X D M N , then each x 2 X has a
unique representation of the form x D m C n for some m 2 M and n 2 N .
Proof. Exercise.
Let M and N be linear subspaces of a linear space X over F such that X D M N . Then
codim.M / Ddim.N /. Also, since X D M N , dim.X / Ddim.M /Cdim.N /. Hence
dim.X / D dim.M / C codim.M /:
It follows that if dim.X / < 1, then codim.M / Ddim.X / dim.M /.
The operator P W X ! X is called an algebraic projection if P is linear (i.e., P .x Cy/ D P x CPy

for all x; y 2 X and 2 F) and P 2 D P , i.e., P is idempotent.
8
1.5.3 Proposition
Let M and N be linear subspaces of a linear space X over F such that X D M N . Define P W X ! X
by P .x/ D m, where x D m C n, with m 2 M and n 2 N . Then P is an algebraic projection of X onto
M along N . Moreover M D P .X / and N D .I P /.X / D ker.P /.
Conversely, if P W X ! X is an algebraic projection, then X D M N , where M D P .X / and

N D .I P /.X / D ker.P /.
Proof. Linearity of P: Let x D m1 C n1 and y D m2 C n2, where m1 ; m2 2 M and n1 ; n2 2 N . For

2 F,
P .x C y/ D P ..m1 C m2 / C . n1 C n2 // D m1 C m2 D P x C Py:
Idempotency of P: Since m D m C 0, with m 2 M and 0 2 N , we have that P m D m and hence
P 2 x D P m D m D P x. That is, P 2 D P .
Finally, n D x m D .I P /x. Hence N D .I P /.X /. Also, P x D 0 if and only if x 2 N , i.e.,
ker.P / D N .
Conversely, let x 2 X and set m D P x and n D .I P /x. Then x D m C n, where m 2 M and

n 2 N . We show that this representation is unique. Indeed, if x D m1 C n1 where m1 2 M and n1 2 N ,
then m1 D P u and n1 D .I P /v for some u; v 2 X . Since P 2 D P , it follows that P m1 D m1 and
P n1 D 0. Hence m D P x D P m1 C P n1 D P m1 D m1 . Similarly n D n1.
1.6 The Holder and Minkowski Inequalities

We now turn our attention to three important inequalities. The first two are required mainly to prove the
third which is required for our discussion about normed linear spaces in the subsequent chapter.
1.6.1 Definition 1 1
Let p and q be positive real numbers. If 1 < p < 1 and C D 1, or if p D 1 and q D 1, or if p D 1
p q
and q D 1, then we say that p and q are conjugate exponents.
1.6.2 Lemma
(Youngs Inequality). Let p and q be conjugate exponents, with 1 < p; q < 1 and ; 0. Then
p q
C :
p q
Proof. If p D 2 D q, then the inequality follows from the fact that . /2 0. Notice also, that if D 0
or D 0, then the inequality follows trivially.
If p 6D 2, then consider the function f W 0; 1/ ! R given by

p q
f ./ D C ; for fixed > 0:
p q
1 q
Then, f 0 ./ D p 1 D 0 when p 1 D . That is, when D p 1 D p > 0. We now apply
q
the second derivative test to the critical point D p .
f 00 ./ D .p 1/ p 2
> 0; for all 2 .0; 1/:
9
q
Thus, we have a global minimum at D p : It is easily verified that
q p q p q
0 D f . p / f ./ D C , C ;
p q p q
for each 2 0; 1/.

1.6.3 Theorem

(Holders Inequality for sequences). Let .xn / 2 `p and .yn / 2 `q , where p > 1 and 1=p C 1=q D 1.
Then
1 1
! p1 1 ! 1q
X X X
jxk yk j jxk jp jyk jq :
kD1 kD1 kD1
1
X 1
X 1
X
Proof. If jxk jp D 0 or jyk jq D 0, then the inequality holds. Assume that jxk jp 6D 0 and
kD1 kD1 kD1
1
X
jyk jq 6D 0. Then for k D 1; 2; : : :, we have, by Lemma 1.6.2, that
kD1
jxk j jyk j 1 jxk jp 1 jyk jq

P1 1 P1 1q p P1 jx jp C q P1 jy jq :
jxk j p p q k k
kD1 jyk j
kD1 kD1
kD1
Hence, P1
kD1 jxk yk j 1 1
P1 p1 P1 q1 p C q D 1:
jxk j p q
kD1 kD1 jyk j
That is,
1 1
! p1 1
! 1q
X X X
p q
jxk yk j jxk j jyk j :
kD1 kD1 kD1
1.6.4 Theorem
(Minkowskis Inequality for sequences). Let p > 1 and .xn / and .yn / sequences in `p . Then
1
! p1 1
! p1 1
! p1
X X X
jxk C yk jp jxk jp C jyk jp :
kD1 kD1 kD1
10
1
X
p
Proof. Let q D . If jxk C yk jp D 0, then the inequality holds. We therefore assume that
p 1
kD1
1
X
jxk C yk jp 6D 0. Then
kD1
1
X 1
X
jxk C yk jp D jxk C yk jp 1
jxk C yk j
kD1 kD1
X1 1
X
jxk C yk jp 1
jxk j C jxk C yk jp 1
jyk j
kD1 kD1
1
! q1 2 1 ! p1 1
! p1 3
X X X
jxk C yk j.p 1/q 4 jxk jp C jyk jp 5
kD1 kD1 kD1
1
! q1 2 1 ! p1 1
! p1 3
X X X
p
4 p p 5:
D jxk C yk j jxk j C jyk j
kD1 kD1 kD1
1
! 1q
X
p
Dividing both sides by jxk C yk j , we have
kD1
1
! p1 1
!1 1
1
! p1 1
! p1
X X q
X X
p p p p
jxk C yk j D jxk C yk j jxk j C jyk j :
kD1 kD1 kD1 kD1
1.6.5 Exercise
[1] Show that the set of all n m real matrices is a real linear space.
[2] Show that a subset M of a linear space X is a linear subspace if and only if x C y 2 M
for all x; y 2 M and all ; 2 F.
[3] Prove Proposition 1.3.4

[4] Prove Proposition 1.4.1.
[5] Prove Proposition 1.5.2.
[6] Show that c0 is a linear subspace of the linear space `1 .
[7] Which of the following subsets are linear subspaces of the linear space C 1; 1?
(a) M1 D fx 2 C 1; 1 W x. 1/ D x.1/g.
Z1
(b) M2 D fx 2 C 1; 1 W x.t/dt D 1g.
1
(c) M3 D fx 2 C 1; 1 W jx.t2 / x.t1 /j jt2 t1 j for all t1 ; t2 2 1; 1g.
11
[8] Show that if fM g is a family of linear subspaces of a linear space X , then M D \ M is a
linear subspace of X .
If M and N are linear subspaces of a linear space X , under what condition(s) is M [ N a
linear subspace of X ?
12
Chapter 2
Normed Linear Spaces
2.1 Preliminaries
For us to have a meaningful notion of convergence it is necessary for the Linear space to have a notion
distance and therefore a topology defined on it. This leads us to the definition of a norm which induces a
metric topology in a natural way.
2.1.1 Definition
A norm on a linear space X is a real-valued function k k W X ! R which satisfies the following properties:
For all x; y 2 X and 2 F,
N1. kxk 0;
N2. kxk D 0 x D 0;
N3. kxk D jjkxk;
N4. kx C yk kxk C kyk (Triangle Inequality).
A normed linear space is a pair .X; k k/, where X is a linear space and k k a norm on X . The number
kxk is called the norm or length of x .
Unless there is some danger of confusion, we shall identify the normed linear space .X; k k/ with the
underlying linear space X .
2.1.2 Examples
(Examples of normed linear spaces.)
[1] Let X D F. For each x 2 X , define kxk D jxj. Then .X; k k/ is a normed linear space.
We give the proof for X D C. Properties N1 -N3 are easy to verify. We only verify N4. Let
x; y 2 C. Then
kx C yk2 D jx C yj2 D .x C y/.x C y/ D .x C y/.x C y/ D xx C yx C xy C yy

D jxj2 C xy C xy C jyj2 D jxj2 C 2Re.xy/ C jyj2
jxj2 C 2jxyj C jyj2 D jxj2 C 2jxjjyj C jyj2
D jxj2 C 2jxjjyj C jyj2
D .jxj C jyj/2 D .kxk C kyk/2 :
Taking the positive square root both sides yields N4.
13
[2] Let n be a natural number and X D Fn . For each x D .x1 ; x2 ; : : : ; xn/ 2 X , define
n
! p1
X
p
kxkp D jxi j ; for 1 p < 1; and
iD1
kxk1 D max jxi j:
1in
Then .X; k kp / and .X; k k1 / are normed linear spaces. We give a detailed proof that
.X; k kp / is a normed linear space for 1 p < 1.
N1. For each 1 i n,
n n
! p1
X X
jxi j 0 ) jxi jp 0 ) jxi jp 0 ) kxkp 0:
iD1 iD1
N2. For any x 2 X ,

n
! p1
X
p
kxkp D 0 jxi j D0
iD1
jxi jp D 0 for all i D 1; 2; 3; : : : ; n
xi D 0 for all i D 1; 2; 3; : : : ; n x D 0:
N3. For any x 2 X and any 2 F,

n
! p1 n
! p1
X X
kxkp D jxi jp D jjp jxi jp
iD1 iD1
n
! p1
X
D jj jxi jp D jjkxkp:
iD1
N4. For any x; y 2 X ,

n
! p1
X
p
kx C ykp D jxi C yi j
iD1
n
! p1 n
! p1
X X
p p
jxi j C jyi j .by Minkowski0 s Inequality/
iD1 iD1
D kxkp C kykp :
[3] Let X D Ba; b be the set of all bounded real-valued functions on a; b. For each x 2 X ,
define
kxk1 D sup jx.t/j:
at b
Then .X; k k1 / is a normed linear space. We prove the triangle inequality: For any t 2 a; b
and any x; y 2 X ,
jx.t/ C y.t/j jx.t/j C jy.t/j sup jx.t/j C sup jy.t/j D kxk1 C kyk1 :
at b at b
14
Since this is true for all t 2 a; b, we have that
kx C yk1 D sup jx.t/ C y.t/j kxk1 C kyk1 :

at b
[4] Let X D Ca; b. For each x 2 X , define
kxk1 D sup jx.t/j

at b
0 1 21
Zb
kxk2 D @ jx.t/j2 dt A :
a
Then .X; k k1 / and .X; k k2 / are normed linear spaces.
[5] Let X D `p ; 1 p < 1. For each x D .xi /1

1 2 X , define
! p1
X
kxkp D jxi jp :
i2N
Then .X; k kp / is a normed linear space.
[6] Let X D `1 ; c or c0 . For each x D .xi /1

1 2 X , define
kxk D kxk1 D sup jxi j:

i2N
Then X is a normed linear space.
[7] Let X D L.Cn / be the linear space of all n n complex matrices. For A 2 L.Cn /, let
n
X
.A/ D .A/ii be the trace of A. For A 2 L.Cn /, define
iD1
v v
u n n u n n
p uX X uX X
kAk2 D
.A A/ D t .A/ki .A/ki D t j.A/ki j2 ;
iD1 kD1 iD1 kD1
where A is the conjugate transpose of the matrix A.
Notation
Let a be an element of a normed linear space.X; k k/ and r > 0.
B.a; r / D fx 2 X j kx ak < r g .Open ball with centre a and radius r /I

Ba; r D fx 2 X j kx ak r g .Closed ball with centre a and radius r /I
S.a; r / D fx 2 X j kx ak D r g .Sphere with centre a and radius r /:
15
y y y
1 1 1
1 1 x 1 1 x 1 1 x
1 1 1
k.x; y/k1 D 1 k.x; y/k2 < 1 k.x; y/k1 1
Equivalent Norms
2.1.3 Definition
Let k k and k k0 be two different norms defined on the same linear space X . We say that k k is equivalent
to k k0 if there are positive numbers and such that
kxk kxk0 kxk; for all x 2 X:
2.1.4 Example
Let X D Fn . For each x D .x1 ; x2 ; : : : ; xn/ 2 X , let
n n
! 21
X X
kxk1 D jxi j; kxk2 D jxi j2 ; and kxk1 D max jxi j:
1in
iD1 iD1
We have seen that k k1 ; k k2 and k k1 are norms on X . We show that these norms are
equivalent.
Equivalence of k k1 and k k1 : Let x D .x1 ; x2 ; : : : ; xn / 2 X . For each k D 1; 2; : : : ; n,

n
X n
X
jxk j jxi j ) max jxk j jxi j kxk1 kxk1:
1kn
iD1 iD1
Also, for k D 1; 2; : : : ; n,
n
X n
X
jxk j max jxk j D kxk1 ) jxi j kxk1 D nkxk1 kxk1 nkxk1:
1kn
iD1 iD1
Hence, kxk1 kxk1 nkxk1.
We now show that k k2 is equivalent to k k1 . Let x D .x1 ; x2 ; : : : ; xn / 2 X . For each

k D 1; 2; : : : ; n,
n
X n
X
jxk j kxk1 ) jxk j2 .kxk1 /2 ) jxi j2 .kxk1 /2 D n.kxk1/2
iD1 iD1
p
kxk2 nkxk1 :
16
Also, for each k D 1; 2; : : : ; n,

n
!1=2
X
2
jxk j jxi j D kxk2 ) max jxk j kxk2 kxk1 kxk2 :
1kn
iD1
p
Consequently, kxk1 kxk2 nkxk1, which proves equivalence of the norms k k2 and k k1 .
It is, of course, obvious now that all the three norms are equivalent to each other. We shall
see later that all norms on a finite-dimensional normed linear space are equivalent.
2.1.5 Exercise
Let N .X / denote the set of norms on a linear space X . For k k and k k0 in N .X /, define a relation ' by
k k ' k k0 if and only if k k is equivalent to k k0 :
Show that ' is an equivalence relation on N .X /, i.e., ' is reflexive, symmetric, and transitive.
Open and Closed Sets
2.1.6 Definition
A subset S of a normed linear space .X; k k/ is open if for each s 2 S there is an > 0 such that
B.s; / S .
A subset F of a normed linear space .X; k k/ is closed if its complement X n F is open.
2.1.7 Definition
Let S be a subset of a normed linear space .X; k k/. We define the closure of S, denoted by S, to be the
intersection of all closed sets containing S.
It is easy to show that S is closed if and only if S D S.
Recall that a metric on a set X is a real-valued function d W X X ! R which satisfies the following
properties: For all x; y; z 2 X ,
M1. d.x; y/ 0;
M2. d.x; y/ D 0 x D y;
M3. d.x; y/ D d.y; x/;
M4. d.x; z/ d.x; y/ C d.y; z/.
2.1.1 Theorem
(a) If .X; k k/ is a normed linear space, then
d.x; y/ D kx yk
defines a metric on X . Such a metric d is said to be induced or generated by the norm k k. Thus,
every normed linear space is a metric space, and unless otherwise specified, we shall henceforth
regard any normed linear space as a metric space with respect to the metric induced by its norm.
(b) If d is a metric on a linear space X satisfying the properties: For all x; y; z 2 X and for all 2 F,
.i/ d.x; y/ D d.x C z; y C z/ (Translation Invariance)
.ii/ d.x; y/ D jjd.x; y/ (Absolute Homogeneity);
then
kxk D d.x; 0/
defines a norm on X .
17
Proof. (a) We show that d.x; y/ D kx yk defines a metric on X . To that end, let x; y; z 2 X .
M1. d.x; y/ D kx yk 0 by N1.
M2.
d.x; y/ D 0 kx yk D 0 x y D 0 by N2
x D y:
M3.
d.x; y/ D kx yk D k. 1/.y x/k D j 1jky xk by N3
D ky xk D d.y; x/:
M4.
d.x; z/ D kx zk D k.x y/ C .y z/k kx yk C ky zk by N4
D d.x; y/ C d.y; z/:
(b) Exercise.
It is clear from Theorem 2.1.1, that a metric d on a linear space X is induced by a norm on X if and
only if d is translation-invariant and positive homogeneous.
2.2 Quotient Norm and Quotient Map

We now want to introduce a norm on a quotient space. Let M be a closed linear subspace of a normed
linear space X over F. For x 2 X , define
kxk WD inf kyk:
y2x
If y 2 x, then y x 2 M and hence y D x C m for some m 2 M . Hence

kxk D inf kyk D inf kx C mk D inf kx mk D d.x; M /:
y2x m2M m2M
2.2.1 Proposition
Let M be a closed linear subspace of a normed linear space X over F. The quotient space X=M is a
normed linear space with respect to the norm
kxk WD inf kyk; where x 2 X=M:
y2x
Proof.
N1. It is clear that for any x 2 X , kxk D d.x; M / 0.
N2. For any x 2 X ,
kxk D 0 d.x; M / D 0 x 2 M D M x C M D M D 0:
N3. For any x; y 2 X and 2 F n f0g,

y

kxk D kxk D d.x; M / D inf kx yk D inf x
y2M y2M
D jj inf kx zk D jjd.x; M / D jjkxk:
z2M
18
N4. Let x; y 2 X . Then
kx C yk D kx C yk D d.x C y; M / D inf kx C y zk
z2M
D inf kx C y .z1 C z2 /k
z1 ;z2 2M
D inf k.x z1 / C .y z2 /k
z1 ;z2 2M
inf kx z1 k C ky z2 k
z1 ;z2 2M
D inf kx z1 k C inf ky z2 k
z1 2M z2 2M
D d.x; M / C d.y; M / D kxk C kyk:
The norm on X=M as defined in Proposition 2.2.1 is called the quotient norm on X=M .
Let M be a closed subspace of the normed linear space X . The mapping QM from X ! X=M defined
by
QM .x/ D x C M; x 2 X;
is called the quotient map (or natural embedding) of X onto X=M .
2.3 Completeness of Normed Linear Spaces
Now that we have established that every normed linear space is a metric space, we can deploy on a normed
linear space all the machinery that exists for metric spaces.
2.3.1 Definition
Let .xn /1
nD1 be a sequence in a normed linear space .X; k k/.
(a) .xn /1
nD1 is said to converge to x if given > 0 there exists a natural number N D N ./ such that
kxn xk < for all n N:
Equivalently, .xn /1
nD1 converges to x if
lim kxn xk D 0:
n!1
If this is the case, we shall write
xn ! x or lim xn D x:
n!1
Convergence in the norm is called norm convergence or strong convergence.

(b) .xn /1
nD1 is called a Cauchy sequence if given > 0 there exists a natural number N D N ./ such
that
kxn xmk < for all n; m N:
Equivalently, .xn / is Cauchy if
lim kxn xmk D 0:
n;m!1
19
In the following lemma we collect some elementary but fundamental facts about normed linear spaces.
In particular, it implies that the operations of addition and scalar multiplication, as well as the norm and
distance functions, are continuous.
2.3.2 Lemma
Let C be a closed set in a normed linear space .X; k k/ over F, and let .xn / be a sequence contained in C
such that lim xn D x 2 X . Then x 2 C .
n!1
Proof. Exercise.
2.3.3 Lemma
Let X be a normed linear space and A a nonempty subset of X .
[1] jd.x; A/ d.y; A/j kx yk for all x; y 2 X ;
[2] j kxk kyk j kx yk for all x; y 2 X ;
[3] If xn ! x , then kxn k ! kxk;
[4] If xn ! x and yn ! y , then xn C yn ! x C y ;
[5] If xn ! x and n ! , then n xn ! x ;
[6] The closure of a linear subspace in X is again a linear subspace;
[7] Every Cauchy sequence is bounded;
[8] Every convergent sequence is a Cauchy sequence.
Proof. (1). For any a 2 A,
d.x; A/ kx ak kx yk C ky ak;
so d.x; A/ kx yk C d.y; A/ or d.x; A/ d.y; A/ kx yk: Interchanging the roles of x and y gives
the desired result.
(2) follows from (1) by taking A D f0g.
(3) is an obvious consequence of (2).
(4), (5) and (8) follow from the triangle inequality and, in the case of (5), the absolute homogeneity.
(6) follows from (4) and (5).
(7). Let .xn / be a Cauchy sequence in X . Choose n1 so that kxn xn1 k 1 for all n n1 . By (2),
kxn k 1 C kxn1 k for all n n1 . Thus
kxnk maxf kx1k; kx2k; kx3k; : : : ; kxn1 1 k; 1 C kxn1 kg
for all n.
(8) Let .xn / be a sequence in X which converges to x 2 X and let > 0. Then there is a natural
number N such that kxn xk < 2 for all n N . For all n; m N ,

kxn xm k kxn xk C kx xm k < C D :
2 2
Thus, .xn / is a Cauchy sequence in X .
2.3.4 Proposition
Let .X; kk/ be a normed linear space over F. A Cauchy sequence in X which has a convergent subsequence
is convergent.
20
Proof. Let .xn / be a Cauchy sequence in X and .xnk / its subsequence which converges to x 2 X . Then,
for any > 0, there are positive integers N1 and N2 such that

kxn xm k < for all n; m N1
2
and

kxnk for all k N2 :
xk <
2
Let N D maxfN1 ; N2 g. If k N , then since nk k,

kxk xk kxk xnk k C kxnk xk < C D :
2 2
Hence xn ! x as n ! 1.
2.3.5 Definition
A metric space .X; d / is said to be complete if every Cauchy sequence in X converges in X .
2.3.6 Definition
A normed linear space that is complete with respect to the metric induced by the norm is called a Banach
space.
2.3.1 Theorem
Let .X; k k/ be a Banach space and let M be a linear subspace of X . Then M is complete if and only if
the M is closed in X .
Proof. Assume that M is complete. We show that M is closed. To that end, let x 2 M . Then there
is a sequence .xn / in M such that kxn xk ! 0 as n ! 1. Since .xn / converges, it is Cauchy.
Completeness of M guarantees the existence of an element y 2 M such that kxn yk ! 0 as n ! 1.
By uniqueness of limits, x D y. Hence x 2 M and, consequently, M is closed.
Assume that M is closed. We show that M is complete. Let .xn / be a Cauchy sequence in M . Then
.xn / is a Cauchy sequence in X . Since X is complete, there is an element x 2 X such that kxn xk ! 0
as n ! 1. But then x 2 M since M is closed. Hence M is complete.
2.3.7 Examples
[1] Let 1 p < 1. Then for each positive integer n, .Fn ; k kp / is a Banach space.
[2] For each positive integer n, .Fn ; k k1 / is a Banach space.
[3] Let 1 p < 1. The sequence space `p is a Banach space. Because of the importance of
this space, we give a detailed proof of its completeness.
The classical sequence space `p is complete.

Proof. Let .xn /1
1 be a Cauchy sequence in `p . We shall denote each member of this
sequence by
xn D .xn .1/; xn .2/; : : :/:
Then, given > 0, there exists an N ./ D N 2 N such that
1
! p1
X
p
kxn xm kp D jxn .i / xm .i /j < for all n; m N:
iD1
For each fixed index i , we have
jxn .i / xm .i /j < for all n; m N:
21
That is, for each fixed index i , .xn .i //1

1 is a Cauchy sequence in F. Since F is complete,
there exists x.i / 2 F such that
xn .i / ! x.i / as n ! 1:
Define x D .x.1/; x.2/; : : :/. We show that x 2 `p , and xn ! x. To that end, for each k 2 N,
k
! p1 1
! p1
X X
p p
jxn .i / xm.i /j kxn xm kp D jxn .i / xm .i /j < :
iD1 iD1
That is,
k
X
jxn .i / xm.i /jp < p ; for all k D 1; 2; 3; : : : :
iD1
Keep k and n N fixed and let m ! 1. Since we are dealing with a finite sum,
k
X
jxn .i / x.i /jp p :
iD1
Now letting k ! 1, then for all n N ,

1
X
jxn .i / x.i /jp p ; .2:3:7:1/
iD1
which means that xn x 2 `p . Since xn 2 `p , we have that x D .x xn / C xn 2 `p . It also

follows from (2.3.7.1) that xn ! x as n ! 1.
[4] The space `0 of all sequences .xi /11 with only a finite number of nonzero terms is an in-
complete normed linear space. It suffices to show that `0 is not closed in `2 (and hence not
complete). To that end, consider the sequence .xi /1
1 with terms
x1 D .1; 0; 0; 0; : : :/
1
x2 D .1; ; 0; 0; 0; : : :/
2
1 1
x3 D .1; ; 2 ; 0; 0; 0; : : :/
2 2
::
:
1 1 1
xn D .1; ; 2 ; : : : ; n 1 ; 0; 0; 0; : : :/
2 2 2
::
:
This sequence .xi /1

1 converges to
1 1 1 1 1
x D .1; ; 2 ; : : : ; n 1 ; n ; nC1 ; : : :/:
2 2 2 2 2
Indeed, since x xn D .0; 0; 0; : : : ; 0; 21n ; 2nC1

1
; : : :/, it follows that
1
X 1
kxn xk2 D ! 0 as n ! 1:
22k
kDn
That is, xn ! x as n ! 1, but x 62 `0 :
22
[5] The space C2 1; 1 of continuous real-valued functions on 1; 1 with the norm

0 1 1=2
Z1
kxk2 D @ x 2 .t/ dt A
1
is an incomplete normed linear space.

To see this, it suffices to show that there is a Cauchy sequence in C2 1; 1 which converges
to an element which does not belong to C2 1; 1. Consider the sequence .xn /1 1 2 C2 1; 1
defined by 8

0 if 1t 0

<
xn .t/ D nt if 0 t n1

:
1 if n1 t 1:
y
xn .t/
1
1
1 0 n 1 t
We show that .xn /1

1 is a Cauchy sequence in C2 1; 1. To that end, for positive integers m
and n such that m > n,
Z1
kxn xmk22 D xn .t/ xm.t/2 dt
1
1=m
Z Z1=n
2
D nt mt dt C 1 nt2 dt
0 1=m
1=m
Z Z1=n
2 2 2 2 2
D m t 2mnt C n t dt C 1 2nt C n2t 2 dt
0 1=m
1=m 1=n
t3 t 3
D .m2 2mn C n2/ C t nt 2 C n2
3 0 3 1=m
m2 2mn C n2 .m n/2
D D ! 0 as n; m ! 1:
3m2 n 3m2 n
Define
0 if 1t 0
x.t/ D
1 if 0 < t 1:
23
Then x 62 C2 1; 1, and
1
Z1 Zn
1
kxn xk22 D xn .t/ x.t/2 dt D nt 12 dt D !0 as n ! 1:
3n
1 0
That is, xn ! x as n ! 1.
2.4 Series in Normed Linear Spaces
Let .xn / be a sequence in a normed linear space .X; k k/. To this sequence we associate another sequence
Xn
.sn / of partial sums, where sn D xk .
kD1
2.4.1 Definition
Let .xn / be a sequence in a normed linear space .X; k k/. If the sequence .sn / of partial sums converges to
1
X 1
X
s , then we say that the series xk converges and that its sum is s . In this case we write xk D s .
kD1 kD1
1
X 1
X
The series xk is said to be absolutely convergent if kxk k < 1.
kD1 kD1
We now give a series characterization of completeness in normed linear spaces.
2.4.1 Theorem
A normed linear space .X; k k/ is a Banach space if and only if every absolutely convergent series in X is
convergent.
1
X 1
X
Proof. Let X be a Banach space and suppose that kxj k < 1. We show that the series xj converges.
jD1 jD1
n
X
To that end, let > 0 and for each n 2 N, let sn D xj . Let K be a positive integer such that
jD1
1
X
kxj k < . Then, for all m > n > K, we have
jDK C1
m m
X n
X X X m 1
X 1
X

ksm sn k D xj xj D xj kxj k kxj k kxj k < :

1 1 nC1 nC1 nC1 K C1
Hence the sequence .sn / of partial sums forms a Cauchy sequence in X . Since X is complete, the sequence
1
X
.sn / converges to some element s 2 X . That is, the series xj converges.
jD1
Conversely, assume that .X; k k/ is a normed linear space in which every absolutely convergent series
converges. We show that X is complete. Let .xn / be a Cauchy sequence in X . Then there is an n1 2 N
such that kxn1 xmk < 12 whenever m > n1 . Similarly, there is an n2 2 N with n2 > n1 such that
kxn2 xm k < 212 whenever m > n2. Continuing in this way, we get natural numbers n1 < n2 < such
24
that kxnk xm k < 21k whenever m > nk . In particular, we have that for each k 2 N, kxnkC1 xnk k < 2 k
.
For each k 2 N, let yk D xnkC1 xnk . Then
n
X n
X n
X 1
kyk k D kxnkC1 xnk k < :
2k
kD1 kD1 kD1
1
X 1
X
Hence, kyk k < 1. That is, the series yk is absolutely convergent, and hence, by our assumption,
kD1 kD1
1 j
X X
the series yk is convergent in X . That is, there is an s 2 X such that sj D yk ! s as j ! 1. It
kD1 kD1
follows that
j
X j
X j!1
sj D yk D xnkC1 xnk D xnj C1 xn1 ! s:
kD1 kD1
j!1
Hence xnj C1 ! s C xn1 . Thus, the subsequence xnk of .xn / converges in X . But if a Cauchy
sequence has a convergent subsequence, then the sequence itself also converges (to the same limit as the
subsequence). It thus follows that the sequence .xn / also converges in X . Hence X is complete.
We now apply Theorem 2.4.1 to show that if M is a closed linear subspace of a Banach space X , then
the quotient space X=M , with the quotient norm, is also a Banach space.
2.4.2 Theorem
Let M be a closed linear subspace of a Banach space X . Then the quotient space X=M is a Banach space
when equipped with the quotient norm.
1
X
Proof. Let .xn / be a sequence in X=M such that kxj k < 1. For each j 2 N, choose an element
jD1
yj 2 M such that
j
kxj yj k kxj k C 2 :
1
X 1
X
It now follows that kxj yj k < 1, i.e., the series .xj yj / is absolutely convergent in X . Since
jD1 jD1
1
X 1
X
X is complete, the series .xj yj / converges to some element z 2 X . We show that the series xj
jD1 jD1
converges to z. Indeed, for each n 2 N,
2 3 2 3
n n n
X X X
xj z D 4 x 5 z D 4 x z 5
j j
jD1 jD1 jD1

n
X
D inf
x j z m

m2M
jD1

n n
X X
xj z y
j

jD1 jD1

X
n
D .xj yj / z ! 0 as n ! 1:

jD1
25
Hence, every absolutely convergent series in X=M is convergent, and so X=M is complete.
2.5 Bounded, Totally Bounded, and Compact Subsets of a Normed

Linear Space
2.5.1 Definition
A subset A of a normed linear space .X; k k/ is bounded if A Bx; r for some x 2 X and r > 0.
It is clear that A is bounded if and only if there is a C > 0 such that kak C for all a 2 A.
2.5.2 Definition
Let A be a subset of a normed linear space .X; k k/ and > 0. A subset A X is called an -net for A
if for each x 2 A there is an element y 2 A such that kx yk < . Simply put, A X is an -net for A
if each element of A is within an distance to some element of A .
A subset A of a normed linear space .X; k k/ is totally bounded (or precompact) if for any > 0 there
is a finite -net F X for A. That is, there is a finite set F X such that
[
A B.x; /:
x2F
The following proposition shows that total boundedness is a stronger property than boundedness.
2.5.3 Proposition
Every totally bounded subset of a normed linear space .X; k k/ is bounded.
Proof. This follows from the fact that a finite union of bounded sets is also bounded.
The following example shows that boundedness does not, in general, imply total boundedness.
2.5.4 Example
Let X D `2 and consider B D B.X / D fx 2 X j kxk 1g, the closed unit ball in X . Clearly, B
is bounded. We show that B is not totally bounded. Consider the elements of B of the form: p for
j 2 N, ej D .0; 0; : : : ; 0; 1; 0; : : :/, where 1 occurs in the j -th position.
p
Note that ke i ej k 2 D 2
for all i j . Assume that an -net B X existed for 0 < < 22 . Then for each j 2 N,
there is an element yj 2 B such that kej yj k < . This says that for each j 2 N, there is an
element yj 2 B such that yj 2 B.ej ; /. But the balls B.ej ; / are disjoint. Indeed, if i 6D j , and
z 2 B.ei ; / \ B.ej ; /, then by the triangle inequality
p p
2 D kei ej k2 kei zk C kz ej k < 2 < 2;
which is absurd. Since the balls B.ej ; / are (at least) countably infinite, there can be no finite
-net for B.
In our definition of total boundedness of a subset A X , we required that the finite -net be a subset
of X . The following proposition suggests that the finite -net may actually be assumed to be a subset of A
itself.
26
2.5.5 Proposition
A subset A of a normed linear space .X; k k/ is totally bounded if and only if for any > 0 there is a finite
set F A such that [
A B.x; /:
x2F
Proof. Exercise.
We now give a characterization of total boundedness.
2.5.1 Theorem
A subset K of a normed linear space .X; k k/ is totally bounded if and only if every sequence in K has a
Cauchy subsequence.
Proof. Assume that K is totally bounded and let .xn / be an infinite sequence in K. There is a finite set of
points fy11 ; y12 ; : : : ; y1r g in K such that
r
[ 1
K B.y1j ; /:
2
jD1
At least one of the balls B.y1j ; 21 /; j D 1; 2; : : : ; r , contains an infinite subsequence .xn1 / of .xn /. Again,
there is a finite set fy21 ; y22 ; : : : ; y2s g in K such that
s
[ 1
K B.y2j ; /:
22
jD1
At least one of the balls B.y2j ; 212 /; j D 1; 2; : : : ; s, contains an infinite subsequence .xn2 / of .xn1 /.
at the m-th
Continuing in this way, step, we obtain a subsequence .xnm / of .xn.m 1/ / which is contained in
a ball of the form B ymj ; 21m .
Claim: The diagonal subsequence .xnn / of .xn / is Cauchy. Indeed, if m > n, then both xnn and xmm are
in the ball of radius 2 n . Hence, by the triangle inequality,
kxnn xmm k < 21 n

! 0 as n ! 1:
Conversely, assume that every sequence in K has a Cauchy subsequence and that K is not totally
bounded. Then, for some > 0, no finite -net exists for K. Hence, if x1 2 K, then there is an x2 2 K
such that kx1 x2 k . (Otherwise, kx1 yk < for all y 2 K and consequently fx1g is a finite -net
for K, a contradiction.) Similarly, there is an x3 2 K such that
kx1 x3k and kx2 x3 k :
Continuing in this way, we obtain a sequence .xn / in K such that kxn xm k for all m n. Therefore
.xn / cannot have a Cauchy subsequence, a contradiction.
2.5.6 Definition
A normed linear space .X; k k/ is sequentially compact if every sequence in X has a convergent subse-
quence.
2.5.7 Remark
It can be shown that on a metric space, compactness and sequential compactness are equivalent. Thus, it
follows, that on a normed linear space, we can use these terms interchangeably.
27
2.5.2 Theorem
A subset of a normed linear space is sequentially compact if and only if it is totally bounded and complete.
Proof. Let K be a sequentially compact subset of a normed linear space .X; k k/. We show that K is
totally bounded. To that end, let .xn / be a sequence in K. By sequential compactness of K, .xn / has a
subsequence .xnk / which converges in K. Since every convergent sequence is Cauchy, the subsequence
.xnk / of .xn / is Cauchy. Therefore, by Theorem 2.5.1, K is totally bounded.
Next, we show that K is complete. Let .xn / be a Cauchy sequence in K. By sequential compactness
of K, .xn / has a subsequence .xnk / which converges in K. But if a subsequence of a Cauchy sequence
converges, so does the full sequence. Hence .xn / converges in K and so K is complete.
Conversely, assume that K is a totally bounded and complete subset of a normed linear space .X; k k/.
We show that K is sequentially compact. Let .xn / be a sequence in K. By Theorem 2.5.1, .xn / has a Cauchy
subsequence .xnk /. Since K is complete, .xnk / converges in K. Hence K is sequentially compact.
2.5.8 Corollary
A subset of a Banach space is sequentially compact if and only if it is totally bounded and closed.
Proof. Exercise.
2.5.9 Corollary
A sequentially compact subset of a normed linear space is closed and bounded.
Proof. Exercise.
We shall see that in finite-dimensional spaces the converse of Corollary 2.5.9 also holds.
2.5.10 Corollary
A closed subset F of a sequentially compact normed linear space .X; k k/ is sequentially compact.
Proof. Exercise.
2.6 Finite Dimensional Normed Linear Spaces
The theory for finite-dimensional normed linear spaces turns out to be much simpler than that of their
infinite-dimensional counterparts. In this section we highlight some of the special aspects of finite-dimensional
normed linear spaces.
The following Lemma is crucial in the analysis of finite-dimensional normed linear spaces.
2.6.1 Lemma
Let .X; k k/ be a finite-dimensional normed linear space with basis fx1 ; x2 ; : : : ; xn g. Then there is a
constant m > 0 such that for every choice of scalars 1 ; 2 ; : : : ; n , we have

n n
X X
m
jj j j xj :
jD1 jD1
28
n
X
Proof. If jj j D 0, then j D 0 for all j D 1; 2; : : : ; n and the inequality holds for any m > 0.
jD1
n
X
Assume that jj j 0. We shall prove the result for a set of scalars f1 ; 2 ; : : : ; n g that satisfy the
jD1
n
X
condition jj j D 1. Let
jD1
n
X
A D f.1 ; 2 ; : : : ; n / 2 Fn j jj j D 1g:
jD1
Since A is a closed and bounded subset of Fn , it is compact. Define f W A ! R by

n
X

f .1 ; 2; : : : ; n / D j xj
:
jD1
Since for any .1 ; 2 ; : : : ; n / and .1 ; 2 ; : : : ; n / in A

n n
X X
jf .1 ; 2 ; : : : ; n / f .1 ; 2 ; : : : ; n /j D x
j j

x
j j
jD1 jD1

n n
X X
j xj j xj

jD1 jD1

n n
X X
D .j j /xj

jj j jkxj k
jD1 jD1
n
X
max kxj k jj j j;
1jn
jD1
f is continuous on A. Since f is a continuous function on a compact set A, it attains its minimum on A,

i.e., there is an element .1 ; 2; : : : ; n / 2 A such that
f .1 ; 2; : : : ; n / D infff .1 ; 2 ; : : : ; n / j .1 ; 2 ; : : : ; n / 2 Ag:
Let m D f .1 ; 2 ; : : : ; n /. Since f 0, it follows that m 0. If m D 0, then

n n
X X
x
j j
D 0 ) j xj D 0:

jD1 jD1
Since the set fx1 ; x2 ; : : : ; xn g is linearly independent, j D 0 for all j D 1; 2; : : : ; n. This is a

contradiction since .1 ; 2 ; : : : ; n / 2 A. Hence m > 0 and consequently for all .1 ; 2; : : : ; n / 2 A,

n n
X X
0 < m f .1 ; 2 ; : : : ; n / m
jj j j xj
:
jD1 jD1
n
X
Now, let f1 ; 2 ; : : : ; n g be any collection of scalars and set D jj j. If D 0, then the
jD1
29

1 2 n
inequality holds vacuously. If > 0, then ; ; ::: ; 2 A and consequently

n n n
X X j 1 2 n X
j xj
D xj
D f ; ; : : : ; m D m jj j:

jD1 jD1
jD1

n n
X X
That is, m jj j
x
j j .
jD1 jD1
2.6.1 Theorem
Let X be a finite-dimensional normed linear space over F. Then all norms on X are equivalent.
Proof. Let fx1; x2 ; : : : ; xn g be a basis for X and k k0 and k k be any two norms on X . For any x 2 X
n
X
there is a set of scalars f1; 2 ; : : : ; n g such that x D j xj . By Lemma 2.6.1, there is an m > 0 such
jD1
that
n n
X X
m
jj j j xj
D kxk:
jD1 jD1
By the triangle inequality
n
X n
X
kxk0 jj jkxj k0 M jj j;
jD1 jD1
where M D max kxj k0 . Hence

1jn

1 m m
kxk0 M kxk ) kxk0 kxk kxk0 kxk where D :
m M M
Interchanging the roles of the norms k k0 and k k, we similarly get a constant such that kxk kxk0.
Hence, kxk0 kxk kxk0 for some constants and .
2.6.2 Theorem
Every finite-dimensional normed linear space .X; k k/ is complete.
Proof. Let fx1 ; x2 ; : : : ; xn g be a basis for X and let .zk / be a Cauchy sequence in X . Then, given any
> 0, there is a natural number N such that
kzk z` k < for all k; ` > N:

n
X
Also, for each k 2 N, zk D kj xj . By Lemma 2.6.1, there is an m > 0 such that
jD1
n
X
m jkj `j j kzk z` k:
jD1
Hence, for all k; ` > N and all j D 1; 2; : : : ; n,

1
jkj `j j kzk z` k < :
m m
30
That is, for each j D 1; 2; : : : ; n, .kj /k is a Cauchy sequence of numbers. Since F is complete, kj ! j
Xn
as k ! 1 for each j D 1; 2; : : : ; n. Define z D j xj . Then z 2 X and
jD1

n n n n
X X X X
kzk zk D
kj xj j xj
D .kj j /xj
jkj j jkxj k ! 0
jD1 jD1 jD1 jD1
as k ! 1. That is, the sequence .zk / converges to z 2 X . hence X is complete.
2.6.2 Corollary
Every finite-dimensional normed linear space X is closed.
Proof. Exercise.
2.6.3 Theorem
In a finite-dimensional normed linear space .X; k k/, a subset K X is sequentially compact if and only
if it is closed and bounded.
Proof. We have seen (Corollary 2.5.9), that a compact subset of a normed linear space is closed and
bounded.
Conversely, assume that a subset K X is closed and bounded. We show that K is compact. Let
n
X
fx1 ; x2 ; : : : ; xng be a basis for X and let .zk / be any sequence in K. Then for each k 2 N, zk D kj xj .
jD1
Since K is bounded, there is a positive constant M such that kzk k M for all k 2 N. By Lemma 2.6.1,
there is an m > 0 such that
n X
X n
m jkj j

kj j D kzk k M:
x
jD1 jD1
It now follows that jkj j M m

for each j D 1; 2; : : : ; n, and for all k 2 N. That is, for each fixed j D
1; 2; : : : ; n, the sequence .kj /k of numbers is bounded. Hence the sequence .kj /k has a subsequence
Xn
.kr j / which converges to j for j D 1; 2; : : : ; n. Setting z D j xj , we have that
jD1

n n n
X X X
kzkr zk D
kr j xj j xj
jkr j j jkxj k ! 0 as r ! 1:
jD1 jD1 jD1
That is, zkr ! z as r ! 1. Since K is closed, z 2 K. Hence K is compact.
2.6.3 Lemma
(Rieszs Lemma). Let M be a closed proper linear subspace of a normed linear space .X; k k/. Then for
each 0 < < 1, there is an element z 2 X such that kzk D 1 and
ky zk > 1 for all y 2 M:
Proof. Choose x 2 X n M and define
d D d.x; M / D inf kx mk:

m2M
31
Since M is closed, d > 0. By definition of infimum, there is a m 2 M such that
d kx mk < d C d D d.1 C /:

m x
Take z D . Then kzk D 1 and for any y 2 M ,
km xk

m x
ky zk D y C D ky.km xk/ C m xk
km xk km xk
d d 1
> D D1 >1 :
km xk d.1 C / 1C 1C
We now give a topological characterization of the algebraic concept of finite dimensionality.

2.6.4 Theorem
A normed linear space .X; k k/ is finite-dimensional if and only its closed unit ball B.X / D fx 2
X j kxk 1g is compact.
Proof. Assume that .X; k k/ is finite-dimensional normed linear space. Since the ball B.X / is closed and
bounded, it is compact.
Assume that the closed unit ball B.X / D fx 2 X j kxk 1g is compact. Then B.X / is totally
bounded. Hence there is a finite 21 -net fx1 ; x2 ; : : : ; xn g in B.X /. Let M Dlinfx1; x2 ; : : : ; xn g. Then
M is a finite-dimensional linear subspace of X and hence closed.
Claim: M D X . If M is a proper subspace of X , then, by Rieszs Lemma there is an element x0 2 B.X /
such that d.x0; M / > 12 . In particular, kx0 xk k > 12 for all k D 1; 2; : : : ; n: However this contradicts
the fact that fx1 ; x2; : : : ; xn g is a 21 -net in B.X /. Hence M D X and, consequently, X is finite-
dimensional.
We now give another argument to show that boundedness does not imply total boundedness. Let X D `2
and B.X / D fx 2 X j kxk2 1g. It is obvious that B.X / is bounded. We show that B.X / is not totally
bounded. Since X is complete and B.X / is a closed subset of X , B.X / is complete. If B.X / were totally
bounded, then B.X / would, according to Theorem 2.26, be compact. By Theorem 2.6.4, X would be
finite-dimensional. But this is false since X is infinite-dimensional.
2.7 Separable Spaces and Schauder Bases

2.7.1 Definition
(a) A subset S of a normed linear space .X; k k/ is said to be dense in X if S D X ; i.e., for each x 2 X
and > 0, there is a y 2 S such that kx yk < .
(b) A normed linear space .X; k k/ is said to be separable if it contains a countable dense subset.
2.7.2 Examples
[1] The real line R is separable since the set Q of rational numbers is a countable dense subset
of R.
[2] The complex plane C is separable since the set of all complex numbers with rational real
and imaginary parts is a countable dense subset of C.
[3] The sequence space `p , where 1 p < 1, is separable. Take M to be the set of all
sequences with rational entries such that all but a finite number of the entries are zero. (If
32
the entries are complex, take for M the set of finitely nonzero sequences with rational real
and imaginary parts.) It is clear that M is countable. We show that M is dense in `p . Let
> 0 and x D .xn / 2 `p . Then there is an N such that
1
X
jxk jp < :
2
kDN C1

Now, for each 1 k N , there is a rational number qk such that jxk q k jp < 2N
. Set
q D .q1 ; q2 ; : : : ; qN ; 0; 0; : : :/. Then q 2 M and
N
X 1
X
kx qkpp D jxk p
qk j C jxk jp < :
kD1 kDN C1
Hence M is dense in `p .
[4] The sequence space `1 , with the supremum norm, is not separable. To see this, consider
the set M of elements x D .xn /, in which xn is either 0 or 1. This set is uncountable since
we may consider each element of M as a binary representation of a number in the interval
0; 1. Hence there are uncountably many sequences of zeroes and ones. For any two
distinct elements x; y 2 M , kx yk1 D 1. Let each of the elements of M be a centre of
a ball of radius 41 . Then we get uncountably many nonintersecting balls. If A is any dense
subset of `1 , then each of these balls contains a point of A. Hence A cannot be countable
and, consequently, `1 is not separable.
2.7.1 Theorem
A normed linear space .X; k k/ is separable if and only if it contains a countable set B such that lin.B/ D
X.
Proof. Assume that X is separable and let A be a countable dense subset of X . Since the linear hull of A,
lin.A/, contains A and A is dense in X , we have that lin.A/ is dense in X , that is, lin.A/ D X .
Conversely, assume that X contains a countable set B such that lin.B/ D X . Let B D fxn j n 2 Ng.
Assume first that D R, and put
8 9
<X n =
C D j xj j j 2 Q; j D 1; 2; : : : ; n; n 2 N :
: ;
jD1
We first show that C is a countable subset of X . The set Q B is countable and consequently, the family
F of all finite subsets of Q B is also countable. The mapping
n
X
f.1 ; x1/; .2 ; x2 /; : : : ; .n ; xn/g 7! j xj
jD1
maps F onto C . Hence C is countable.

Next, we show that C is dense in X . Let x 2 X and > 0. Since lin.B/ D X , we can find an n 2 N,
points x1 ; x2; : : : ; xn 2 B and 1 ; 2; : : : ; n 2 F such that

n
X
x
j j < :
x

jD1 2
33
Since Q is dense in R, for each i 2 R, we can find a i 2 Q such that

ji i j < for all i D 1; 2; : : : ; n:
2n.1 C kxi k/
Hence,

n n X
X X n n
X
x j xj x j xj j xj
C j xj
jD1 jD1 jD1 jD1
n
X
< C jj j jkxj k
2
jD1
n
X kxj k
< C < C D :
2 2n.1 C kxj k/ 2 2
jD1
This shows that C is dense in X .

If F D C, the set C is that of finite linear combinations with coefficients being those complex numbers with
rational real and imaginary parts.
We now give another argument based on Theorem 2.7.1 to show that the sequence space `p , where
1 p < 1, is separable. Let en D .nm /m2N , where

1 if n D m
nm D
0 otherwise:
Clearly, en 2 `p . Let > 0 and x D .xn / 2 `p . Then there is a natural number N such that
1
X
jxk jp < p for all n N:
kDnC1
Now, if n N , then
0 11=p
n 1
X X
x xj ej @ jxk jp A
D < :
jD1 kDnC1
p
Hence lin.fen j n 2 Ng/ D `p . Of course, the set fen j n 2 Ng is countable.
2.7.3 Definition
A sequence .bn / in a Banach space .X; k k/ is called a Schauder basis if for any x 2 X , there is a unique
sequence .n / of scalars such that
Xn

lim x
j j D 0:
b
n!1
jD1
1
X
In this case we write x D j bj :
jD1
2.7.4 Remark
It is clear from Definition 2.7.3 that .bn / is a Schauder basis if and only if X D linfbn j n 2 Ng and
1
X
every x 2 X has a unique expansion x D j bj :
jD1
Uniqueness of this expansion clearly implies that the set fbn j n 2 Ng is linearly independent.
34
2.7.5 Examples
[1] For 1 p < 1, the sequence .en /, where en D .nm /m2N , is a Schauder basis for `p .
[2] .en / is a Schauder basis for c0 .
[3] .en / [ feg, where e D .1; 1; 1; : : :/ (the constant 1 sequence), is a Schauder basis for c.
[4] `1 has no Schauder basis.
2.7.6 Proposition
If a Banach space .X; k k/ has a Schauder basis, then it is separable.
Proof. Let .bn / be a Schauder basis for X . Then fbn j n 2 Ng is countable and
lin.fbn j n 2 Ng/ D X .
Schauder bases have been constructed for most of the well-known Banach spaces. Schauder conjectured
that every separable Banach space has a Schauder basis. This conjecture, known as the Basis Problem,
remained unresolved for a long time until Per Enflo in 1973 answered it in the negative. He constructed a
separable reflexive Banach space with no basis.
2.7.7 Exercise
[1] Let X be a normed linear space over F. Show that X is finite-dimensional if and only if every
bounded sequence in X has a convergent subsequence.
[2] Complete the proof of Theorem 2.1.1.
[3] Prove Lemma 2.3.2.

[4] Prove the claims made in [1] and [2] of Example 2.3.7.
[5] Prove Theorem 2.5.5.
[6] Prove Corollary 2.5.8.
[10] Is .Ca; b; k k1 / complete? What about .Ca; b; k k1 /? Fully justify both answers.
35
Chapter 3
Hilbert Spaces
3.1 Introduction
In this chapter we introduce an inner product which is an abstract version of the dot product in elementary
vector algebra. Recall that if x D .x1 ; x2; x3/ and y D .y1 ; y2 ; y3 / are any two vectors in R3 , then the
q product of x and
dot
p
y is x y D x1 y1 C x2y2 C x3 y3 . Also, the length of the vector x is kxk D
2 2 2
x1 C x2 C x3 D x x.
It turns out that Hilbert spaces are a natural generalization of finite-dimensional Euclidean spaces.
Hilbert spaces arise naturally and frequently in mathematics, physics, and engineering, typically as infinite-
dimensional function spaces.
3.1.1 Definition
Let X be a linear space over a field F. An inner product on X is a scalar-valued function h; i W X X ! F
such that for all x; y; z 2 X and for all ; 2 F, we have
IP1. hx; xi 0;
IP2. hx; xi D 0 x D 0;
IP3. hx; yi D hy; xi (The bar denotes complex conjugation.);
IP4. hx; yi D hx; yi;
IP5. hx C y; zi D hx; zi C hy; zi.
An inner product space .X; h; i/ is a linear space X together with an inner h; i product defined on it. An
inner product space is also called pre-Hilbert space.
3.1.2 Examples
Examples of inner product spaces.
[1] Fix a positive integer n. Let X D Fn . For x D .x1 ; x2; : : : ; xn/ and y D .y1 ; y2 ; : : : ; yn / in X ,
define
n
X
hx; yi D xi yi :
iD1
Since this is a finite sum, h; i is well-defined. It is easy to show that .X; h; i/ is an inner
product space. The space Rn (resp. Cn ) with this inner product is called the Euclidean
n-space (resp. unitary n-space) and will be denoted by `2 .n/.
36
[2] Let X D `0 , the linear space of finitely non-zero sequences of real or complex numbers. For
x D .x1 ; x2 ; : : :/ and y D .y1 ; y2 ; : : :/ in X , define
1
X
hx; yi D xi yi :
iD1
Since this is essentially a finite sum, h; i is well-defined. It is easy to show that .X; h; i/ is
an inner product space.
[3] Let X D `2 , the space of all sequences x D .x1 ; x2; : : :/ of real or complex numbers with
X1
jxi j2 < 1. For x D .x1 ; x2; : : :/ and y D .y1 ; y2 ; : : :/ in X , define
1
1
X
hx; yi D xi yi :
iD1
In order to show that h; i is well-defined we first observe that if a and b are real numbers,
then
1
0 .a b/2 ; whence ab .a2 C b 2/:
2
Using this fact, we have that
!
1 2 X1 1 1
2 1 X 2
X
2
jxi yi j D jxi jjyi j jxi j C jyi j ) jxi yi j jxi j C jyi j < 1:
2 2
iD1 iD1 iD1
Hence, h; i is well-defined (i.e., the series converges).

[4] Let X D Ca; b, the space of all continuous complex-valued functions on a; b. For x; y 2 X ,
define
Zb
hx; yi D x.t/y.t/ dt:
a
We shall denote by C2 a; b the linear space Ca; b equipped with this inner product.
[5] Let X D L.Cn / be the linear space of all n n complex matrices. For A 2 L.Cn /, let
n
X
.A/ D .A/ii be the trace of A. For A; B 2 L.Cn /, define
iD1
hA; Bi D .B A/; where B denotes conjugate transpose of matrix B:

Show that .L.Cn /; h; i/ is an inner product space.
It should be mentioned that we could consider real Hilbert spaces but there are powerful methods that
can be applied by using the more general complex Hilbert spaces.
3.1.1 Theorem
(Cauchy-Bunyakowsky-Schwarz Inequality). Let .X; h; i/ be an inner product space over a field F.
Then for all x; y 2 X , p p
jhx; yij hx; xi hy; yi:
Moreover, given any x; y 2 X , the equality
p p
jhx; yij D hx; xi hy; yi
holds if and only if x and y are linearly dependent.
37
Proof. If x D 0 or y D 0, then the result holds vacuously. Assume that x 6D 0 and y 6D 0. For any 2 F,
we have
0 hx y; x yi D hx; xi hy; xi hx; yi C hy; yi

D hx; xi hx; yi hy; xi hy; yi:
hx; yi
Now choosing D , we have
hy; yi
hy; xihx; yi jhx; yij2

0 hx; xi D hx; xi ;
hy; yi hy; yi
whence p p
jhx; yij hx; xi hy; yi:
p p
Assume that jhx; yij D hx; xi hy; yi. We show that x and y are linearly dependent. If x D 0 or
y D 0, then x and y are obviously linearly dependent. We therefore assume that x 6D 0 and y 6D 0. Then
hx; yi
hy; yi 6D 0. With D , we have that
hy; yi
jhx; yij2
hx y; x yi D hx; xi D 0:
hy; yi
That is,
hx y; x yi D 0; ) x D y:
That is, x and y are linearly dependent.
Conversely, assume that x and y are linearly dependent. Without loss of generality, x D y for some
2 F. Then
jhx; yij D jhy; yij D jjjhy; yij D jjhy; yi

p p q p p p
D jj hy; yi hy; yi D jj2hy; yi hy; yi D hyyi hy; yi
p p
D hx; xi hy; yi:
3.1.2 Theorem
Let .X; h; i/ be an inner product space over a field F. For each x 2 X , define
p
kxk WD hx; xi: .3:1:2:1/
Then k k defines a norm on X . That is, .X; k k/ is a normed linear space over F.
Proof. Let x; y 2 X and 2 F. Then

p
N1. kxk D hx; xi 0;
p
N2. kxk D 0 hx; xi D 0 hx; xi D 0 x D 0; by IP2.
p p p
N3. kxk D hxxi D jj2hx; xi D jj hx; xi D jjkxk.
N4.
38
kx C yk2 D hx C y; x C yi D hx; xi C hx; yi C hy; xi C hy; yi

D hx; xi C hx; yi C hx; yi C hy; yi
D kxk2 C 2Rehx;
p yipC kyk2 kxk2 C 2jhx; yij C kyk2
kxk C 2 hx; xi hy; yi C kyk2 (by Theorem 3.1.1)
2
D kxk2 C 2kxkkyk C kyk2 D .kxk C kyk/2 :

Taking the positive square root both sides yields
kx C yk kxk C kyk:
In view of (3.1.2.1), the Cauchy-Bunyakowsky-Schwarz Inequality now becomes
jhx; yij kxkkyk:
Any inner product space can thus be made into a normed linear space in a natural way: by defining the norm
as in (3.1.2.1). The norm advertised in (3.1.2.1) is called the inner product norm or a norm induced or
generated by the inner product.
A natural question arises: Is every normed linear space an inner product space? If the answer is NO,
how then does one recognise among all normed linear spaces those that are inner product spaces in disguise,
i.e., those whose norms are induced by an inner product?
These questions will be examined later.
3.1.3 Theorem
(Polarization Identity). Let .X; h; i/ be an inner product space over a field F. Then for all x; y 2 X ,
kx C yk2 kx yk2
hx; yi D if F D R; and
4 4

kx C yk2 kx yk2 kxCyik2 kx yi k2
hx; yi D Ci 4
if F D C:
4 4 4
Proof. Assume that F D R. Then
kx C yk2 kx yk2 D hx C y; x C yi hx y; x yi
D hx; xi C hx; yi C hy; xi C hy; yi hx; xi C hx; yi C hy; xi hy; yi
D 4hx; yi; since hx; yi D hy; xi:
The case when F D C is proved analogously and is left as an exercise.
3.1.4 Theorem
(Parallelogram Identity). Let .X; h; i/ be an inner product space over a field F. Then for all x; y 2 X ,
kx yk2 C kx C yk2 D 2kxk2 C 2kyk2 : .3:1:4:1/
Proof.
kx yk2 C kx C yk2 D hx y; x yi C hx C y; x C yi
D hx; xi hx; yi hy; xi C hy; yi C hx; xi C hx; yi C hy; xi C hy; yi
D 2kxk2 C 2kyk2 :
The geometric interpretation of the Parallelogram Identity is evident: the sum of the squares of the
lengths of the diagonals of a parallelogram is equal to the sum of the squares of the lengths of the four
39
sides.
The following theorem asserts that the Parallelogram Identity (Theorem 3.1.4) distinguishes inner prod-
uct spaces among all normed linear spaces. It also answers the question posed after Theorem 3.1.2. That is,
a normed linear space is an inner product space if and only if its norm satisfies the Parallelogram Identity.
3.1.5 Theorem
A normed linear space X over a field F is an inner product space if and only if the Parallelogram Identity
kx yk2 C kx C yk2 D 2kxk2 C 2kyk2 .PI /
holds for all x; y 2 X .
Proof. ). We have already shown (Theorem 3.1.4) that if X is an inner product space, then the parallel-
ogram identity (PI) holds in X .
(. Let X be a normed linear space in which the parallelogram identity (PI) holds. We shall only
consider the case F D R. The polarization identity (Theorem 3.1.3) gives us a hint as to how we should
define an inner product: For all x; y 2 X , define
x
x C y 2 y
2
hx; yi D
2
:
2
We claim that h; i is an inner product on X .

x C x 2
x x 2
IP1. hx; xi D 2
D kxk2 0.
2
IP2. hx; xi D 0 kxk2 D 0 x D 0.

2
x C y 2
x y 2 y x
IP3. hx; yi D D y C x

2
D hy; xi D hy; xi since F D R.
2 2 2 2
IP5. Replace x by u C v and y by w C v in the parallelogram identity:
ku C w C 2vk2 C ku wk2 D 2ku C vk2 C 2kw C vk2 : .3:1:5:1/
Replace x by u v and y by w v in the parallelogram identity:
ku C w 2vk2 C ku wk2 D 2ku vk2 C 2kw vk2 : .3:1:5:2/
Subtract (3.1.5.2) from (3.1.5.1):

h i
ku C w C 2vk2 ku C w 2vk2 D 2 ku C vk2 ku vk2 C kv C wk2 kv wk2 :
40
Use the definition of h; i,

1
4hu C w; 2vi D 8hu; vi C hw; vi ) hu C w; 2vi D hu; vi C hw; vi: .3:1:5:3/
2
Take w D 0:
1
hu; 2vi D hu; vi: .3:1:5:4/
2
Now replace u by x C y and v by z in (3.1.5.4) and use (3.1.5.3) to get
1
hx C y; zi D hx C y; 2zi D hx; zi C hy; zi:
2
IP4. We show that hx; yi D hx; yi for all 2 R and all x; y 2 X . If D n is a nonzero integer, then
using IP5, D x E D nx E
hnx; yi D nhx; yi ) n ; y D ; y D hx; yi:
n n
That is,
Dx E 1
; y D hx; yi:
n n
If is a rational number, D pq , say. Then

p x p
x; y D p ; y D hx; yi:
q q q
If 2 R, then there is a sequence .rk / of rational numbers such that rk ! as k ! 1. Using
continuity of the norm, we have that
2 2
1
1
hx; yi D h lim rk x; yi D lim rk x C y
lim rk x y

k!1 4 k!1 4 k!1
1 1
D lim krk x C yk2 lim krk x yk2
4 k!1 4 k!1
!
rk x C y 2 r x y 2
k
D lim

k!1 2 2
D lim hrk x; yi
k!1
D lim rk hx; yi D hx; yi:
k!1
Thus, hx; yi D hx; yi for all 2 R and all x; y 2 X .
3.1.3 Corollary
Let .X; k k/ be a normed linear space over a field F. If every two-dimensional linear subspace of X is an
inner product space over F, then X is an inner product space.
3.1.4 Examples
[1] Let X D `p ; for p 6D 2. Then X is not an inner product space. We show that the norm
on `p ; p 6D 2 does not satisfy the parallelogram identity. Take x D .1; 1; 0; 0; : : :/ and
y D .1; 1; 0; 0; : : :/ in `p . Then
1
kxk D 2 p D kyk and kx C yk D 2 D kx yk:
Thus,
2
kx C yk2 C kx yk2 D 8 6D 2kxk2 C 2kyk2 D 4 2 p :
41
[2] The normed linear space X D Ca; b, with the supremum norm k k1 is not an inner product
space. We show that the norm
kxk1 D max jx.t/j
at b
does not satisfy the parallelogram identity. To that end, take

t a
x.t/ D 1 and y.t/ D :
b a
Since
t a t a
x.t/ C y.t/ D 1 C and x.t/ y.t/ D 1 ;
b a b a
we have that
kxk D 1 D kyk; and kx C yk D 2; kx yk D 1:
Thus,
kx C yk2 C kx yk2 D 5 6D 2kxk2 C 2kyk2 D 4:
3.2 Completeness of Inner Product Spaces

The mathematical concept of a Hilbert space, named after David Hilbert, generalizes the notion of Euclidean
space. Hilbert spaces, as the following definition states, are inner product spaces which in addition are
required to be complete, a property that stipulates the existence of enough limits in the space to allow the
techniques of calculus to be used.
The earliest Hilbert spaces were studied from this more abstract point of view in the first decade of
the 20th century by David Hilbert, Erhard Schmidt, and Frigyes Riesz. They are indispensable tools in
the theories of partial differential equations, quantum mechanics, Fourier analysis which includes applica-
tions to signal processing, and ergodic theory which forms the mathematical underpinning of the study of
thermodynamics.
3.2.1 Definition
Let .X; h; i/ be an inner product space. If X is complete with respect to the norm induced by the inner
product h; i, then we say that X is a Hilbert space.
3.2.2 Examples
[1] The classical space `2 is a Hilbert space.
[2] `0 is an incomplete inner product space.
[3] The space C 1; 1 is an incomplete inner product space.
3.3 Orthogonality
3.3.1 Definition
Two elements x and y in an inner product space .X; h; i/ are said to be orthogonal, denoted by x ? y , if
hx; yi D 0:
The set M X is called orthogonal if it consists of non-zero pairwise orthogonal elements.

If M is a subset of X such that hx; mi D 0 for all m 2 M , then we say that x is orthogonal to M and write
x ? M . We shall denote by
M ? D fx 2 X W hx; mi D 0 8 m 2 M g
42
the set of all elements in X that are orthogonal to M . The set M ? is called the orthogonal complement
of M .
3.3.2 Proposition
Let M and N be subsets of an inner product space .X; h; i/. Then
[1] f0g? D X and X ? D f0g;

[2] M ? is a closed linear subspace of X ;
[3] M .M ? /? D M ?? ;
[4] If M is a linear subspace, then M \ M ? D f0g;
[5] If M N , then N ? M ? ;
[6] M ? D .linM /? D .linM /? .
Proof.
[1] Exercise.
[2] Let x; y 2 M ?, and ; 2 F. Then for each z 2 M ,
hx C y; zi D hx; zi C hy; zi D 0:
Hence, x C y 2 M ? . That is, M ? is a subspace of X . To show that M ? is closed, let x 2 M ?.

Then there exists a sequence .xn / in M ? such that xn ! x as n ! 1. Thus, for all y 2 M ,
hx; yi D limhxn; yi D 0;
n
whence x 2 M ?.
[3] Exercise.
[4] Exercise.
[5] Let x 2 N ? . Then hx; yi D 0 for all y 2 N . In particular, hx; yi D 0 for all y 2 M since M N .
Thus, x 2 M ? .
[6] Since M linM linM , we have, by [5], that .linM /? .linM /? M ? . It remains to show
that M ? .linM /? . To that end, let x 2 M ? . Then hx; yi D 0 for all y 2 M , and consequently
hx; yi D 0 for all y 2 linM . If z 2 linM , then there exists a sequence .zn / in linM such that zn ! z
as n ! 1. Thus,
hx; zi D limhx; zni D 0;
n
?
whence x 2 .linM / .
3.3.3 Examples
Let X D R3 . The vectors . 3; 0; 2/ and .4; 1; 6/ are orthogonal since
h. 3; 0; 2/; .4; 1; 6/i D . 3/.4/ C 0.1/ C .2/.6/ D 0:
43
If M D `0 , the linear subspace of `2 consisting of all scalar sequences .xi /1

1 with only a finite
number of nonzero terms, then M ? D f0g. Indeed, suppose that y D .yi /1 iD1 2 M ?. Let
8
< 1 if i D j
ij D
:
0 if i 6D j:
and en D .nj /1
n;jD1 . Then en 2 M for each n 2 N, and hence,
1
X
0 D hy; ei i D yj ij D yi for all i D 1; 2; : : : :
jD1
That is, y D 0, whence M ? D f0g.

3.3.1 Theorem
(Pythagoras). Let .X; h; i/ be an inner product space over a field F and let x; y 2 X .
[1] If F D R, then x ? y if and only if
kx C yk2 D kxk2 C kyk2 :
[2] If F D C, then x ? y if and only if

kx C yk2 D kxk2 C kyk2 and kx C iyk2 D kxk2 C kyk2 :
Proof. [1] ). If x ? y, then

kx C yk2 D hx C y; x C yi D hx; xi C 2hx; yi C hy; yi D kxk2 C kyk2 :
(. Suppose that kx C yk2 D kxk2 C kyk2 : Then
hx C y; x C yi D hx; xi C hy; yi
) hx; xi C 2hx; yi C hy; yi D hx; xi C hy; yi
) 2hx; yi D 0 ) hx; yi D 0:
[2] ). Assume that x ? y. Then
kx C yi k2 D hx C yi ; x C yi i D hx; xi C hx; yi i C hyi ; xi C hyi ; yi i
D hx; xi i hx; yi C i hy; xi C hy; yi D kxk2 C kyk2 :
(. Assume that kx C yk2 D kxk2 C kyk2 and kx C iyk2 D kxk2 C kyk2 : Then
hx C y; x C yi D hx; xi C hy; yi
) hx; xi C hx; yi C hy; xi C hy; yi D hx; xi C hy; yi
) hx; yi C hy; xi D 0 ) 2Rehx; yi D 0 ) Rehx; yi D 0:
Also,
hx C yi ; x C yi i D hx; xi C hy; yi
) hx; xi i hx; yi C i hy; xi C hy; yi D hx; xi C hy; yi
) i hx; yi C i hy; xi D 0
) i hx; yi hy; xi D 0
h i
) i hx; yi hx; yi D 0
) i 2i Imhx; yi D 0 ) Imhx; yi D 0:
Since Rehx; yi D 0 D Imhx; yi, we have that hx; yi D 0.
44
3.3.4 Corollary
If M D fx1; x2 ; : : : ; xng is an orthogonal set in an inner product space .X; h; i/ then
2
Xn n
X

xi D kxi k2 :

iD1 iD1
Proof. Exercise.
3.4 Best Approximation in Hilbert Spaces
3.4.1 Definition
Let K be a closed subset of an inner product space .X; h; i/. For a given x 2 X nK , a best approximation
or nearest point to x from K is any element y0 2 K such that
kx y0 k kx yk for all y 2 K:
Equivalently, y0 2 K is a best approximation to x from K if
kx y0 k D inf kx yk D d.x; K/:

y2K
The (possibly empty) set of all best approximations to x from K is denoted by PK .x/. That is,
PK .x/ D fy 2 K W kx yk D d.x; K/g:
The (generally set-valued) map PK which associates each x in X with its best approximations in K is
called the metric projection or the nearest point map. The set K is called
[1] proximinal if each x 2 X has a best approximation in K ; i.e., PK .x/ 6D ; for each x 2 X ;
[2] Chebyshev if each x 2 X has a unique best approximation in K ; i.e., the set PK .x/ consists of a
single point.
The following important result asserts that if K is a complete convex subset of an inner product space
.X; h; i/, then each x 2 X has one and only one element of best approximation in K.
3.4.1 Theorem
Every nonempty complete convex subset K of an inner product space .X; h; i/ is a Chebyshev set.
Proof. Existence: Without loss of generality, x 2 X n K. Let
D inf kx yk:
y2K
By definition of the infimum, there exists a sequence .yn /1

1 in K such that
kx yn k ! as n ! 1:
We show that .yn /1

1 is a Cauchy sequence. By the Parallelogram Identity (Theorem 3.1.3),
kym yn k2 D k.x yn / .x ym /k2

D 2kx yn k2 C 2kx ym k2 k2x .yn C ym /k2
2
yn C ym
D 2kx yn k2 C 2kx ym k2 4
x

2
2kx yn k2 C 2kx ym k2 4 2;
45
yn C ym
since 2 K by convexity of K. Thus,
2
kym yn k2 2kx yn k2 C 2kx ym k2 4 2 ! 0 as n; m ! 1:
That is, .yn /1

1 is a Cauchy sequence in K. Since K is complete, there exists y 2 K such that yn !
y as n ! 1. Since the norm is continuous,
kx yk D kx lim yn k D k lim .x yn /k D lim kx yn k D :

n!1 n!1 n!1
Thus,
kx yk D D d.x; K/:
Uniqueness: Assume that y; y0 2 K are two best approximations to x from K. That is,
kx y0 k D kx yk D D d.x; K/:
By the Parallelogram Identity,
0 ky y0 k2 D k.y x/ C .x y0 /k2
D 2kx yk2 C 2kx y0 k2 k2x .y C y0 /k2
2
y C y0
2
D 2 C 2 2
4 x
2
4 2 4 2 D 0:
Thus, y0 D y.
3.4.2 Corollary
Every nonempty closed convex subset of a Hilbert space is Chebyshev.
The following theorem characterizes best approximations from a closed convex subset of a Hilbert
space.
3.4.2 Theorem
Let K be a nonempty closed convex subset of a Hilbert space .H; h; i/, x 2 H n K and y0 2 K . Then y0
is the best approximation to x from K if and only if
Rehx y0 ; y y0 i 0 for all y 2 K:
Proof. The existence and uniqueness of the best approximation to x in K are guaranteed by Theorem 3.4.1.
Let y0 be the best approximation to x in K. Then, for any y 2 K and any 0 < < 1, y C .1 /y0 2 K
since K is convex. Thus,
kx y0 k2 kx y C .1 /y0 k2 D k.x y0 / .y y0 /k2

D hx y0 / .y y0 /; x y0 / .y y0 /i
D hx y0 ; x y0 i hx y0 ; y y0 i C hy y0 ; x y0 i
C2hy y0 ; y y0 i
D kx y0 k2 2Re.hx y0 ; y y0 i/ C 2ky y0 k2
) 2Re.hx y0 ; y y0 i/ 2 ky y0 k2

) Re.hx y0 ; y y0 i/ ky y0 k2 :
2
46

As ! 0, ky y0 k2 ! 0, and consequently Rehx y0 ; y y0 i 0.
2
Conversely, assume that for each y 2 K, Rehx y0 ; y y0 i 0. Then, for any y 2 K,
kx yk2 D k.x y0 / .y y0 /k2

D h.x y0 / .y y0 /; .x y0 / .y y0 /i
D hx y0 ; x y0 i hx y0 ; y y0 i hy y0 ; x y0 i C hy y0 ; y y0 i
D hx y0 ; x y0 i hx y0 ; y y0 i C hy y0 ; x y0 i C hy y0 ; y y0 i
D hx y0 ; x y0 i hx y0 ; y y0 i C hx y0 ; y y0 i C hy y0 ; y y0 i
D kx y0 k2 2Rehx y0 ; y y0 i C ky y0 k2
kx y0 k2 :
Taking the positive square root both sides, we have that kx y0 k kx yk for all y 2 K.
As a corollary to Theorem 3.4.2, one gets the following characterization of best approximations from a
closed linear subspace of a Hilbert space.
3.4.3 Corollary
(Characterization of Best Approximations from closed subspaces). Let M be a closed subspace of a
Hilbert space H and let x 2 H n M . Then an element y0 2 M is the best approximation to x from M if
and only if hx y0 ; yi D 0 for all y 2 M (i.e., x y0 2 M ?/.
Corollary 3.4.3 says that if M is a closed linear subspace of a Hilbert space H, then y0 D PM .x/
(i.e., y0 is the best approximation to x from M ) if and only if x PM .x/ ? M . That is, the unique best
approximation is obtained by dropping the perpendicular from x onto M . It is for this reason that the
map PM W x ! PM .x/ is also called the orthogonal projection of H onto M .
PM x
3.4.4 Example
Let X D C2 1; 1, M D P2 D linf1; t; t 2g, and x.t/ D t 3 . Find PM .x/.
Solution. Note that C2 1; 1 is an incomplete inner product space. Since M is finite-dimensional, it
is complete, and consequently proximinal in C2 1; 1. Uniqueness of best approximations follows
from the Parallelogram Identity.
47
2
X
Let y0 D i t i 2 M . By Corollary 3.4.3,
iD0
y0 D PM .x/ x y0 2 M ?
hx y0 ; t j i D 0 for all j D 0; 1; 2
* 2
+
X
3 i j
t i t ; t D 0 for all j D 0; 1; 2
iD0
2
X
i ht i ; t j i D ht 3 ; t j ; i for all j D 0; 1; 2
iD0
2
X Z1 Z1
i j
i t t dt D t 3 t j dt for all j D 0; 1; 2
iD0 1 1
2
X Z1 Z1
i t iCj dt D t 3Cj dt for all j D 0; 1; 2
iD0 1 1
2
X iCjC1
1 1
t t jC4
i D for all j D 0; 1; 2
i Cj C1 1 j C4 1
iD0
2
X 1 h i 1 h i
i 1 . 1/iCjC1 D 1 . 1/jC4
i Cj C1 j C4
iD0
for all j D 0; 1; 2
8 2
< 2 0 C 0 1 C 3 2 D0
2
0 C C0 D 52
: 2 0 3 1 2 2
C 0 1 C 5 2
3 0
D0
3
0 D 0; 1 D ; 2 D 0:
5
3
Thus, PM .x/ D y0 D t.
5
3.4.3 Theorem
(Projection Theorem). Let H be a Hilbert space, M a closed subspace of H. Then
[1] H D M M ? . That is, each x 2 H can be uniquely decomposed in the form
x D y C z with y 2 M and z 2 M ?:
[2] M D M ?? .
Proof.
[1] If x 2 M , then x D x C 0, and we are done. Assume that x 62 M . Let y D PM .x/ be the unique
best approximation to x from M as advertised in Theorem 3.4.1. Then z D x PM .x/ 2 M ?, and
x D PM .x/ C .x PM .x// D y C z
is the unique representation of x as a sum of an element of M and an element of M ? .
48
[2] Since the containment M M ?? is clear, we only show that M ?? M . To that end, let x 2
M ??. Then by [1] above
x D y C z; where y 2 M and z 2 M ? :
Since M M ?? and M ?? is a subspace, z D x y 2 M ?? . But z 2 M ? implies that

z 2 M ? \ M ?? which, in turn, implies that z D 0. Thus, x D y 2 M .

3.4.5 Corollary
If M is a closed subspace of a Hilbert space H, and if M 6D H, then there exists z 2 H n f0g such that
z ? M.
Proof. Let x 2 H n M . Then by the Projection Theorem,
x D y C z; where y 2 M and z 2 M ?:
Hence z 6D 0 and z ? M .
3.4.6 Proposition
Let S be a nonempty subset of a Hilbert space H. Then
[1] S ?? D linS .
[2] S ? D f0g if and only if linS D H.
Proof.
[1] Since S ? D .linS/? by Proposition 3.3.2, we have, by the Projection Theorem, that
linS D .linS/?? D S ?? :
[2] If S ? D f0g, then by [1]

linS D S ?? D f0g? D H:
On the other hand, if H D linS, then H D S ?? by [1], and so
S ? D S ??? D H? D f0g:
3.5 Orthonormal Sets and Orthonormal Bases
In this section we extend to Hilbert spaces the finite-dimensional concept of an orthonormal basis.
3.5.1 Definition
Let .X; h; i/ be an inner product space over F. A set S D fx W 2 g of elements of X is called an
orthonormal set if
(a) hx ; x i D 0 for all 6D (i.e., S is an orthogonal set), and
(b) kx k D 1 for all 2 .
If S D fx W 2 g is an orthonormal set and x 2 X ,X then the numbers hx; x i are called the Fourier
coefficients of x with respect to S and the formal series hx; x ix the Fourier series of x .
2
49
3.5.1 Theorem
An orthonormal set S in a separable inner product space .X; h; i/ is at most countable.
Proof. If S is finite,
p then there is nothing to prove. Assume that S is infinite. Observe that if x; y 2 S,
then kx yk D 2 (since x and y are orthonormal). Let D D fyn j n 2 Ng be a pcountable dense subset
of X . Then to each x 2 S corresponds an element yn 2 D such that kx yn k < 42 . This defines a map
f W S ! N given by f .x/ D n, where n corresponds to the yn as indicated above. Now, if x and y are
distinct elements of S, then there are distinct elements yn and ym in D such that
p p
2 2
kx yn k < and ky ym k < :
4 4
Hence,
p p
p 2 2
2 D kx yk kx yn k C kyn ym k C kym yk < C kyn ym k < kyn ym k;
2 2
and so yn ym . In particular, n m. Thus, we have a one-to-one correspondence between the elements
of S and a subset of N.
3.5.2 Definition
An orthonormal set S in an inner product space .X; h; i/ is said to be complete in X if S T and T is
an orthonormal set in X , then S D T .
Simply put, a complete orthonormal set S in an inner product space is an orthonormal set that is not
properly contained in any other orthonormal set in X ; in other words, S is complete if it is a maximal
orthonormal set in X .
It is easy exercise to show that a set S is complete in an inner product .X; h; i/ if and only if S ? D f0g.
3.5.3 Examples
[1] In R3 the set S D f.1; 0; 0/; .0; 1; 0/; .0; 0; 1/g is orthonormal.
[2] In `2 , let S D fen W n 2 Ng, where en D .1n ; 2n ; : : :/ with

1 if i D j
ij D
0 otherwise:
Then S is an orthonormal set. Furthermore, for each x D .xi /1

iD1 2 `2 , hx; eni D xn for all n.
Thus
hx; eni D 0 for all n xn D 0 for all n x D 0:
That is, S ? D f0g, hence, S is complete in `2 .
3.5.2 Theorem
Let .X; h; i/ be a separable inner product space over F.
[1] (Best Fit). If fx1 ; x2; : : : ; xn g is a finite orthonormal set in X and M D linfx1 ; x2; : : : ; xn g, then for
each x 2 X there exists y0 2 M such that
kx y0 k D d.x; M /:
n
X
In fact, y0 D hx; xk ixk :
kD1
50
[2] (Bessels Inequality). Let .xn /1

nD1 be an orthonormal sequence in X . Then for any x 2 X ,
1
X
jhx; xk ij2 kxk2:
kD1
In particular, hx; xk i ! 0 as k ! 1.
Proof.
[1] For any choice of scalars 1 ; 2; : : : ; n,
2 * +
n
X n
X n
X

x k xk D x i xi ; x j xj

kD1 iD1 jD1
n
X n
X n
X
D kxk2 i hxi ; xi j hx; xj i C i i
iD1 jD1 iD1
n
X n
X n
X
D kxk2 i hx; xi i j hx; xj i C i i
iD1 jD1 iD1
n
X
D kxk2 C i i i hx; xi i i hx; xi i C hx; xi ihx; xi i
iD1
Xn
hx; xi ihx; xi i
iD1
n
X n
X
D kxk2 C .i hx; xi i/.i hx; xi i/ jhx; xi ij2
iD1 iD1
n
X n
X
D kxk2 C .i hx; xi i/.i hx; xi i/ jhx; xi ij2
iD1 iD1
n
X n
X
D kxk2 jhx; xi ij2 C ji hx; xi ij2 :
iD1 iD1
2
n
X

Therefore, x k xk is minimal if and only if k D hx; xk i for each k D 1; 2; : : : ; n.

kD1
[2] For each positive integer n, and with k D hx; xk i, the above argument shows that
2
n
X n
X

0 x k xk D kxk2 jhx; xi ij2:

kD1 iD1
Thus,
n
X
jhx; xk ij2 kxk2:
kD1
Taking the limit as n ! 1, we get

1
X
jhx; xk ij2 kxk2:
kD1
51
3.5.3 Theorem
(Riesz-Fischer Theorem). Let .xn /1 1 be an orthonormal sequence in a separable Hilbert space H and let
1
X
.cn /1
1 be a sequence of scalars. Then the series ck xk converges in H if and only if c D .cn /1
1 2 `2 . In
kD1
this case,
1 ! 12
X 1
X

ck x k D jck j2 :

kD1 kD1
1
X
Proof. Assume that the series ck xk converges to x. Then for each j ; n 2 N,
kD1
* n
+ n
X X
ck xk ; xj D ck hxk ; xj i D cj :
kD1 kD1
Using continuity of the inner product

*1 + * n +
X X
hx; xj i D ck xk ; xj D lim ck xk ; xj D lim cj D cj :
n!1 n!1
kD1 kD1
By Bessels Inequality, we have that

1
X 1
X
jck j2 D jhx; xk ij2 kxk2 < 1:
kD1 kD1
That is, c D .cn /1

1 2 `2 .
n
X
Conversely, assume that c D .cn /1
1 2 `2 . Set zn D ck xk . Then for 1 n m,
kD1
2
m m
X X
kzn 2
zm k D ck x k D jck j2 ! 0 as n ! 1:

kDnC1 kDnC1
Hence, .zn /1 1
1 is a Cauchy sequence in H. Since H is complete the sequence .zn /1 converges to some
1
X
x 2 H. Hence the series ck xk converges to some element in H.
kD1
Also, 1 2 n 2
X X n
X

c x
k k D lim c x
k k D lim jck j2 ;
n!1 n!1
kD1 kD1 kD1
whence,
1 ! 12
X 1
X
2
ck x k D jck j :

kD1 kD1
Note that Bessels Inequality says that

1
X
jhx; xk ij2 kxk2 < 1:
kD1
52
1
X
That is, .hx; xni/1
1 2 `2 . Hence, by Theorem 3.5.3, the series hx; xk ixk converges. There is however
kD1
no reason why this series should converge to x. In fact, the following example shows that this series may
not converge to x.
3.5.4 Example
Let .en / 2 `2 , where en D .1n ; 2n ; : : :/ with

1 if i D j
ij D
0 otherwise:
For each n 2 N, let fn D enC1 . Then .fn /1

nD1 is an orthonormal sequence in `2 . For any x D
.xn /1
1 2 `2 ,
1
X 1
X
hx; fk ifk D hx; ekC1 iekC1 D .0; x2; x3; : : :/ 6D .x1 ; x2; x3; : : :/ D x:
kD1 kD1
3.5.5 Definition
Let .X; h; i/ be an inner product space over F. An orthonormal set fxn g is called an orthonormal basis
for X if for each x 2 X ,
1
X
xD hx; xk ixk :
kD1
n
X
That is, the sequence of partial sums .sn /, where sn D hx; xk ixk , converges to x .
kD1
3.5.4 Theorem
Let H be a separable infinite-dimensional Hilbert space and assume that S D fxn g is an orthonormal set in
H. Then the following statements are equivalent:
[1] S is complete in H; i.e., S ? D f0g.
[2] linS D H; i.e., the linear span of S is norm-dense in H.
[3] (Fourier Series Expansion.) For any x 2 H, we have
1
X
xD hx; xi ixi :
iD1
That is, S is an orthonormal basis for H.

[4] (Parsevals Identity.) For all x; y 2 H,
1
X
hx; yi D hx; xk ihy; xk i:
kD1
[5] For any x 2 H,

1
X
kxk2 D jhx; xk ij2 :
kD1
53
Proof. [1] [2]. This equivalence was proved in Proposition 3.4.6[2].

n
X
[1] ) [3]. Let x 2 H and sn D hx; xi ixi . Then for all n > m,
iD1
n 2
X n
X
2
ksn sm k D hx; xi ixi D jhx; xi ij2 kxk2:

iDmC1 mC1
Thus, .sn / is a Cauchy sequence in H. Since H is complete, this sequence converges to some element
1
X 1
X
which we denote by hx; xi ixi . We show that x D hx; xi ixi . Indeed, for each fixed j 2 N,
iD1 iD1
* 1
+ * n
+
X X
x hx; xi ixi ; xj D x lim hx; xi ixi ; xj
n!1
iD1 iD1
* n
+
X
D lim x hx; xi ixi ; xj
n!1
iD1
n
!
X
D lim hx; xj i hx; xi ihxi ; xj i
n!1
iD1
D lim .hx; xj i hx; xj i/ D 0:
n!1
1
X
Thus, by [1], x hx; xi ixi D 0, whence
iD1
1
X
xD hx; xi ixi :
iD1
[3] ) [4]. Let x; y 2 H. Then

* n n
+
X X
hx; yi D lim hx; xi ixi ; hy; xj ixj
n!1
iD1 jD1
n X
X n
D lim hx; xi ihy; xj ihxi ; xj i
n!1
iD1 jD1
n
X 1
X
D lim hx; xi ihy; xi i D hx; xi ihy; xi i:
n!1
iD1 iD1
[4] ) [5]. Take x D y in [4].

X
[5] ) [1]. Since kxk2 D hx; xk ihx; xk i, if x ? S then hx; xk i D 0 for all k. Thus, kxk2 D 0,
k
whence x D 0. That is, S ? D f0g.
3.5.6 Examples
[1] In `2 , the set S D fen W n 2 Ng, where en D .1n ; 2n ; : : :/ with

1 if i D j
ij D
0 otherwise
is an orthonormal basis for `2 .
54

1 int
[2] In L2 ; , the set p e W n 2 Z is an orthonormal basis for the complex L2 ; .
2

1 cos nt sin nt 1
[3] The set S D p ; p ; p is an orthonormal basis for the real L2 ; .
2 nD1
Hence, if x 2 L2 ; , then by Theorem 3.5.4[3],

1
X
1 1 cos nt cos nt sin nt sin nt
x.t/ D x.t/; p p C x.t/; p p C x.t/; p p
2 2 nD1
X 1
1 1 1
D hx.t/; 1i C hx.t/; cos nti cos nt C hx.t/; sin nti sin nt
2
nD1
Z
1
D x.t/ dt
2

20 1 0 1 3
1
X Z Z
1 1
C 4@ x.t/ cos nt dt A cos nt C @ x.t/ sin nt dt A sin nt 5

nD1
1
X
D a0 C .an cos nt C bn sin nt/;
nD1
where
D E R 9
a0 D p1 x.t/; p1 D 1
x.t/ dt; >
>
2 2 2 >
>
>
>
D E >
=
R
an D p1 x.t/; p nt
cos
D 1
x.t/ cos nt dt; and n D 1; 2; : : : :
>
>
>
D E >
>
R >
>
bn D p1 x.t/; pnt
sin
D 1
x.t/ sin nt dt ;

That is, the Fourier series expansion of x is

1
X
x.t/ D a0 C .an cos nt C bn sin nt/: .3:5:6:1/
nD1
It is clear from above that for all n D 1; 2; : : :,

1 2 cos nt 2
2 ja0j2 D x.t/; p ; ja n j 2
D
x.t/; p ; and
2

sin nt 2
jbnj2 D x.t/; p :

55
By Theorem 3.5.4 [5] we have that

Z
2 2

1 2
jx.t/j dt D kxk D x.t/; p
2

1 !
X
cos nt 2 sin nt 2
C x.t/; p C x.t/; p
nD1
1
X
2
D 2 ja0j C . jan j2 C jbnj2 /
nD1
1
!
X
2 2 2
D 2ja0 j C .jan j C jbn j / :
nD1
We now apply the above results to a particular function: Let x.t/ D t. Then
Z
1
a0 D t dt D 0 since x.t/ D t is an odd function.
2

Z
1
For n D 1; 2; : : : ; an D t cos nt dt D 0 since t cos nt is an odd function,

Z Z
1 2
bn D t sin nt dt D t sin nt dt

0
2 3
Z
2
t cos nt 1
D 4 5
n C n cos nt dt
0
0
2h i 2. 1/nC1
D cos n D :
n n
Hence, by Theorem 3.5.4[3],

1
X 1
X p
2. 1/nC1 2. 1/nC1 sin nt
x.t/ D sin nt D p :
n n
nD1 nD1
It now follows that p

2. 1/nC1 sin nt
D t; p :
n
Now,
Z Z
2 3 2 3
kxk22 D 2
t dt D 2 2
t dt D t D :
3 0 3
0
Also, by Theorem 3.5.4[5],

1 1 p
X sin nt 2 X 2. 1/nC1 2 X1
4
kxk22 D t; p D D :
n n2
nD1 nD1 nD1
56
Thus,
1
X 1 2
2
D :
n 6
nD1
We can express the Fourier Series Expansion (3.5.6.1) of x 2 L2 ; in exponential form. Recall
that
e i D cos C i sin (Eulers Formula).
Therefore
e i C e i
e i e i
cos D and sin D :
2 2i
Equation (3.5.6.1) now becomes
1
X
x.t/ D a0 C .an cos nt C bn sin nt/
nD1
X1 int
e int C e int e e int
D a0 C an C bn
2 2i
nD1
X1
an i b n int an C i b n int
D a0 C e C e
2 2
nD1
X1 X1
an i b n an C i b n
D a0 C e int C e int : .3:5:6:2/
2 2
nD1 nD1
1 1
For each n D 1; 2; 3; : : :, let cn D 2
.an C i bn /. Then cn D 2
.an i bn / for each n D 1; 2; 3; : : :, and
so equation (3.5.6.2) becomes
1
X 1
X
x.t/ D a0 C cn e int C cn e int
: .3:5:6:3/
nD1 nD1
Re-index the first sum in (3.5.6.3) by letting n D k. Then

1
X 1
X
ikt int
x.t/ D a0 C c k e C cn e : .3:5:6:4/
kD 1 nD1
For n D 1; 2; 3; : : : ; define
cn D c n
and let c0 D a0 . The we can rewrite equation (3.5.6.4) as

1
X
int
x.t/ D cn e : .3:5:6:5/
1
This is the complex exponential form of the Fourier Series of x 2 L2 ; .

Note that,
Z
1
c 0 D a0 D x.t/dt
2

57
and for n D 1; 2; 3; : : : ,
2 3
Z Z
1 1 41 1
cn D .an C i bn / D x.t/ cos nt dt C i x.t/ sin nt dt 5
2 2

Z
1
D x.t/ .cos nt C i sin nt/ dt
2

Z
1
D x.t/e int dt;
2

and
Z Z Z
1 1 1
c n D cn D x.t/e int dt D x.t/e int dt D x.t/e int dt:
2 2 2

Therefore, for all n 2 Z,
Z
1
cn D x.t/e int dt:
2

Now, for n D 1; 2; 3; : : : ,
1 1
jcn j2 D cn cn D .an C i bn / .an i bn /
2 2
1 2 1
D an C bn2 D jan j2 C jbn j2 :
4 4
Therefore
1 X
1
X 1
jcn j2 D jan j2 C jbn j2 : .3:5:6:6/
4
nD1 nD1
Since for n D 1; 2; 3; : : : , c n D cn , it follows that
2
jc nj Dc n c n D cn cn D cn cn D jcn j2 :
Hence, for n D 1; 2; 3; : : : ,
1 X
1
X 1
X 1
2
jc nj D jcn j2 D jan j2 C jbn j2 : .3:5:6:7/
4
nD1 nD1 nD1
From (3.5.6.6) and (3.5.6.7), we have that

X 1
X 1
X
jcn j2 D jc nj
2
C jc0 j2 C jcn j2
n2Z nD1 nD1
1 1 X
X 1
1
D jan j2 C jbn j2 C ja0 j2 C jan j2 C jbn j2
4 4
nD1 nD1
1 X
1
D ja0 j2 C jan j2 C jbn j2
2
nD1
!
1 X 1
D 2 ja0 j2 C jan j2 C jbn j2
2
nD1
Z
1
D jx.t/j2 dt:
2

58
That is,
X Z
2 1
jcn j D jx.t/j2 dt: .3:5:6:8/
2
n2Z
Let fx1; x2 ; : : : ; xn g be a basis of an n-dimensional linear subspace M of an inner product space

.X; h; i/. We have seen in Theorem 3.5.2 that if the set fx1 ; x2 ; : : : ; xn g is orthonormal, then the orthog-
onal projection (=best approximation) of any x 2 X onto M is given by
n
X
PM .x/ D hx; xk ixk :
kD1
It is clearly easy to compute orthogonal projections from a linear subspace that has an orthonormal basis:
the coefficients in the orthogonal projection of x 2 X are just the Fourier coefficients of x. If the basis of
M is not orthogonal, it may be advantageous to find an orthonormal basis for M and express the orthogonal
projection as a linear combination of the new orthonormal basis. The process of finding an orthonormal
basis from a given (non-orthonormal) basis is known as the Gram-Schmidt Orthonormalisation Procedure.
3.5.5 Theorem
(Gram-Schmidt Orthonormalisation Procedure). If fxk g1 1 is a linearly independent set in an inner
product space .X; h; i/ then there exists an orthonormal set fek g1
1 in X such that
linfx1; x2 ; : : : ; xng D linfe1; e2 ; : : : ; en g for all n:

x1
Proof. Set e1 D . Then linfx1g D linfe1g. Next, let y2 D x2 hx2 ; e1ie1 . Then
kx1k
hy2 ; e1i D hx2 hx2; e1ie1 ; e1 i D hx2 ; e1i hx2 ; e1ihe1 ; e1i D hx2; e1 i hx2; e1 i D 0:
y2
That is, e1 ? y2 . Set e2 D . Then fe1 ; e2g is an orthonormal set with the property that linfx1 ; x2g D
ky2 k
linfe1 ; e2g. In general, for each k D 2; 3; : : :, we let
k
X1
yk D xk hxk ; ei iei :
iD1
Then for k D 2; 3; : : :
hyk ; e1i D hyk ; e2 i D hyk ; e3i D D hyk ; ek 1i D 0:

yk
Set ek D . Then fe1 ; e2 ; : : : ; ek g is an orthonormal set in X with the property that
kyk k
linfe1 ; e2; : : : ; ek g D linfx1 ; x2; : : : ; xk g.
We have made the point that `2 is a Hilbert space. In this final part of this chapter we want to show that
every separable infinite-dimensional Hilbert space looks like `2 in the sense defined below.
3.5.7 Definition
Two linear spaces X and Y over the same field F are said to be isomorphic it there is a one-to-one map T
from X onto Y such that for all x1 ; x2 2 X and all ; 2 F,
T .x1 C x2/ D T .x1/ C T .x2 /: .3:5:7:1/
59
3.5.8 Remark
Any map that satisfies condition (3.5.7.1) of Definition 3.5.7 is called a linear operator. Chapter 4
is devoted to the study of such maps. Clearly, the linear structures of the two linear spaces X and
Y are preserved under the map T .
3.5.9 Definition
Let .X; k k/ and .Y; k k/ be two normed linear spaces and T W X ! Y . Then T is called an isometry if
kT xk D kxk for all x 2 X: .3:5:9:1/
Simply put, an isometry is a map that preserves lengths.

3.5.10 Remark
It is implicit in the above definition that the norm on the left of equation (3.5.9.1) is in Y and that on
the right is in X . In order to avoid possible confusion, we should perhaps have labelled the norms
as k kX and k kY for the norms in X and Y respectively. This notation is however cumbersome
and will therefore be avoided.
Normed linear spaces that are isometrically isomorphic are essentially identical.
3.5.11 Lemma
Let M D linfxn g be a linear subspace of X . Then there exists a subsequence fxnk g of fxn g which has the
following properties:
(i) linfxnk g D M ;
(ii) fxnk g is linearly independent.
Proof. We define the subsequence inductively as follows: Let xn1 be the first nonzero element of the
sequence fxng. Therefore xn D 0 xn1 for all n < n1 . If there is an 2 F such that xn D xn1 for all
n > n1 , then we are done. Otherwise, let xn2 be the first element of the sequence fxngn>n1 that is not a
multiple of xn1 . Thus there is an 2 F such that xn D xn1 C 0xn2 for all n < n2. If xn D xn1 C xn2
for some ; 2 F and all n > n2 then we are done. Otherwise let xn3 be the first element of the sequence
fxn g which is not a linear combination of xn1 and xn2 . Then xn D xn1 C xn2 C 0xn3 for all n < n3. If
xn D xn1 C xn2 C xn3 for all n > n3 , then we are done. Otherwise let xn4 be the first element of the
sequence fxn g that is not in linfxn1 ; xn2 ; xn3 g. Continue in this fashion to obtain elements xn1 ; xn2 ; : : :.
If x 2 linfx1 ; x2; : : : ; xng, then x 2 linfxn1 ; xn2 ; : : : ; xnr g for r sufficiently large. That is linfxnk g D M .
The subsequence fxnk g is, by its construction, linearly independent.
3.5.6 Theorem
Every separable Hilbert space H has a countable orthonormal basis.
Proof. By Theorem 2.7.1 there is a set fxn j n 2 Ng such that linfxn j n 2 Ng D H. Using Lemma 3.5.11
extract from fxn j n 2 Ng a linearly independent subsequence fxnk g such that linfxng Dlinfxnk g. Apply
the Gram-Schmidt Orthonormalisation Procedure to the subsequence fxnk g to obtain an orthonormal basis
for H.
3.5.7 Theorem
Every separable infinite-dimensional Hilbert space H is isometrically isomorphic to `2 .
Proof. Let fxn j n 2 Ng be an orthonormal basis for H. Define T W H ! `2 by
T x D .hx; xni/n2N for each x 2 H:
It follows from Bessels Inequality that the right hand side is in `2 . We must show that T is a surjective
linear isometry. (One-to-oneness follows from isometry.)
60
(i) T is linear: Let x; y 2 H and 2 F. Then
T .x C y/ D .hx C y; xni/n2N D .hx; xni C hy; xni/n2N

D .hx; xni/n2N C .hy; xni/n2N D T x C T y;
and
T .x/ D .hx; xni/n2N D .hx; xni/n2N D .hx; xni/n2N :
(ii) T is surjective: Let .cn /n2N 2 `2 . By the Riesz-Fischer Theorem (Theorem 3.5.3), the series
X1
ck xk converges to some x 2 H. By continuity of the inner product, we have that for each
kD1
j 2 N, * +
n
X
hx; xj i D lim ck xk ; xj D lim cj D cj :
n!1 n!1
kD1
Hence, T x D .hx; xni/n2N D .cn /n2N .

(iii) T is an isometry: For each x 2 H,
X
kT xk22 D jhx; xnij2 D kxk2;
n2N
where the second equality follows from the fact that fxn gn2N is an orthonormal basis and Theo-
rem 3.5.4[5].
61
Chapter 4
Bounded Linear Operators and

Functionals
4.1 Introduction
An essential part of functional analysis is the study of continuous linear operators acting on linear spaces.
This is perhaps not surprising since functional analysis arose due to the need to solve differential and
integral equations, and differentiation and integration are well known linear operators. It turns out that it is
advantageous to consider this type of operators in this more abstract way. It should also be mentioned that
in physics, operator means a linear operator from one Hilbert space to another.
4.1.1 Definition
Let X and Y be linear spaces over the same field F. A linear operator from X into Y is a mapping
T W X ! Y such that
T .x1 C x2/ D T x1 C T x2 for all x1 ; x2 2 X and all ; 2 F:
Simply put, a linear operator between linear spaces is a mapping that preserves the structure of the under-
lying linear space.
We shall denote by L.X; Y / the set of all linear operators from X into Y . We shall write L.X / for L.X; X /.
4.1.2 Exercise
Let X and Y be linear spaces over the same field F. Show that if T is a linear operator from X
into Y , then T .0/ D 0.
The range of a linear operator T W X ! Y is the set

ran.T / D fy 2 Y j y D T x for some x 2 X g D T X;
and the null space or the kernel of T 2 L.X; Y / is the set
1
N .T / D ker.T / D fx 2 X W T x D 0g D T .0/:
If T 2 L.X; Y /, then ker.T / is a linear subspace of X and ran.T / is a linear subspace of Y .
An operator T 2 L.X; Y / is one-to-one (or injective) if ker.T / D f0g and onto (or surjective) if
ran.T / D Y . If T 2 L.X; Y / is one-to-one, then there exists a map T 1 W ran.T / ! dom.T / which
maps each y 2 ran.T / onto that x 2 dom.T / for which T x D y. In this case we write T 1 y D x and
the map T 1 is called the inverse of the operator T 2 L.X; Y /. An operator T W X ! Y is invertible if it
has an inverse T 1 .
It is easy exercise to show that an invertible operator can have only one inverse.
62
4.1.3 Proposition
Let X and Y be linear spaces over F. Suppose that T 2 L.X; Y / is invertible. Then
1 1 1
(a) T is also invertible and .T / D T.
1 1
(b) T T D IY and T T D IX .
1
(c) T is a linear operator.
Proof. We shall leave (a) and (b) as an easy exercise.

(c) Linearity of T 1 : Let x; y 2 Y and 2 F. Then
1 1 1 1 1 1 1
T .x C y/ D T .T T xCTT y/ D T T .T xCT y/
1 1 1 1 1
D T T .T xCT y/ D T xCT y:
Let X and Y be linear spaces over F. For all T; S 2 L.X; Y / and 2 F, define the operations of
addition and scalar multiplication as follows:
.T C S/.x/ D T x C Sx and
.T /.x/ D T x for each x 2 X:
Then L.X; Y / is a linear space over F.
The most important class of linear operators is that of bounded linear operators.
4.1.4 Definition
Let X and Y be normed linear spaces over the same field F. A linear operator T W X ! Y is said to be
bounded if there exists a constant M > 0 such that
kT xk M kxk for all x 2 X:
(It should be emphasised that the norm on the left side is in Y and that on the right side is in X .)
An operator T W X ! Y is said to be continuous at x0 2 X if given any > 0 there is a > 0 such that
kT x T x0 k < whenever kx x0 k < :
T is continuous on X if it is continuous at each point of X .
We shall denote by B.X; Y / the set of all bounded linear operators from X into Y . We shall write B.X /
for B.X; X /.
4.1.5 Definition
Let X and Y be normed linear spaces over the same field F and let T 2 B.X; Y /. The operator norm (or
simply norm) of T , denoted by kT k, is defined as
kT k D inffM W kT xk M kxk; for all x 2 X g:
Since T is bounded, kT k < 1. Furthermore,
kT xk kT kkxk for all x 2 X:

4.1.1 Theorem
Let X and Y be normed linear spaces over a field F and let T 2 B.X; Y /. Then

kT xk
kT k D sup W x 6D 0 D supfkT xk W kxk D 1g D supfkT xk W kxk 1g:
kxk
63

kT xk
Proof. Let D sup W x 6D 0 ; D supfkT xk W kxk D 1g, and D supfkT xk W kxk
kxk
kT xk
1g. We first show that kT k D . Now, for all x 2 X n f0g we have that , and therefore
kxk
kT xk kxk. By definition of kT k we have that kT k . On the other hand, for all x 2 X ,
kT xk
we have that kT xk kT kkxk. In particular, for all x 2 X n f0g, kT k, and therefore
kxk
kT xk
D sup W x 6D 0 kT k. Thus, D kT k.
kxk
Next, we show that D D . Now, for each x 2 X

kT xk x
W x 6D 0 D T W x 6D 0 fkT xk W kxk D 1g fkT xk W kxk 1g:
kxk kxk
Thus,

kT xk
D sup W x 6D 0 D supfkT xk W kxk D 1g D supfkT xk W kxk 1g:
kxk
But for all x 6D 0
kT xk
) kT xk kxk for all x such that kxk 1:
kxk
Therefore,
D supfkT xk W kxk 1g :
That is, . Hence, D D .
4.1.2 Theorem
Let X and Y be normed linear spaces over a field F. Then the function k k defined above is a norm on
B.X; Y /.
Proof. Properties N1 and N2 of a norm are easy to verify. We prove N3 and N4. Let T 2 B.X; Y / and
2 F.
N3. kT k D supfkT xk W kxk D 1g D jj supfkT xk W kxk D 1g D jjkT k.
N4. Let T; S 2 B.X; Y /. Then for each x 2 X ,
k.T C S/.x/k D kT x C Sxk kT xk C kSxk .kT k C kSk/kxk:
Thus, kT C Sk kT k C kSk.
4.1.6 Examples
[1] Let X D Fn with the uniform norm k k1 . For x D .x1 ; x2; : : : ; xn/ 2 Fn , define
T W Fn ! Fn by
0 1
n
X n
X n
X
T x D T .x1 ; x2; : : : ; xn / D @ 1j xj ; 2j xj ; : : : ; nj xj A :
jD1 jD1 jD1
It is easy to show that T is a linear operator on X . We show that T is bounded.

n n
X X
kT xk1 D sup ij xj sup jij jjxj j
1in jD1 1in jD1
n
X
sup jij j sup jxj j D M kxk1 ;
1in jD1 1jn
64
n
X
where M D sup jij j. Hence, kT k M .
1in jD1
We claim that kT k D M . We need to show that kT xk1 M kxk1 . To that end, choose
n
X Xn
an index k such that jkj j D M D sup jij j and let x be the unit vector whose j-th
jD1 1in jD1
kj
component is . Then
jkj j

n n n
X X X
kT xk1 D sup
ij xj kj xj D jkj j D M kxk1 :
1in jD1 jD1 jD1
n
X
Thus kT k D sup jij j.
1in jD1
[2] Let fxn j n 2 Ng be an orthonormal set in a Hilbert space H. For .i /1

iD1 2 `1 , define
T W H ! H by
X1
Tx D i hx; xi ixi :
iD1
Then T is a bounded linear operator on H. Linearity is an immediate consequence of the

inner product.
Boundedness:
1 2
X 1
X
2
kT xk D i hx; xi ixi D ji j2 jhx; xi ij2 kxi k2

iD1 iD1
1
X
M2 jhx; xi ij2 ; where M D sup ji j
iD1 i2N
M 2kxk2 by Bessels Inequality:
Thus, kT xk M kxk, and consequently kT k M .

We show that kT k D sup ji j. Indeed, for any > 0, there exists k such that jk j > M .
i2N
Hence,
kT k kT xk k D kk xk k D jk j > M :
Since is arbitrary, we have that kT k M .
[3] Define an operator L W `2 ! `2 by
Lx D L..x1 ; x2; x3 ; : : :// D .x2 ; x3; : : :/:
The L is a bounded linear operator on `2 .

Linearity: Easy.
Boundedness: For all x D .x1 ; x2; x3; : : :/ 2 `2 ,
1
X 1
X
kLxk22 D 2
jxi j jxi j2 D kxk22:
iD2 iD1
65
That is, L is a bounded linear operator and kLk 1. We show that kLk D 1. To that end,
consider e2 D .0; 1; 0; 0; : : :/ 2 `2 . Then
ke2k2 D 1 and Le2 D .1; 0; 0; : : :/ which implies that kLe2k2 D 1:
Thus, kLk D 1. The operator L is called the left-shift operator.

[4] Let BC0; 1/ be the linear space of all bounded continuous functions on the interval 0; 1/
with the uniform norm k k1 . Define T W BC0; 1/ ! BC0; 1/ by
Zt
1
.T x/ .t/ D x./d :
t
0
Then T is a bounded linear operator on BC0; 1/.

Linearity: For all x; y 2 BC0; 1/ and all ; 2 F,
0 t 1 0 t 1
Zt Z Z
1 1 1
.T .x C y// .t/ D .x C y/./d D @ x./d A C @ y./d A :
t t t
0 0 0
Boundedness: For each x 2 BC0; 1/,

t
Z
1
kT xk1
D sup j.T x/.t/j D sup x./d
t t t
0
0 1
Zt Zt
1 1
sup jx./jd @sup d A kxk1 D kxk1:
t t t t
0 0
[5] Let M be a closed subspace of a normed linear space X and QM W X ! X=M the qoutient
map. Then QM is bounded and kQM k D 1. Indeed, since kQM .x/k D kx C M k kxk, QM
is bounded and kQM k 1. But since QM maps the open unit ball in X onto the open unit
ball in X=M , it follows that kQM k D 1.
[6] Let X D mathcalP 0; 1 - the set of polynomials on the interval 0; 1 with the uniform norm
kxk1 D max jx.t/j. For each x 2 X , define T W X ! X by
0t 1
dx
T x D x 0 .t/ D (differentiation with respect to t/:
dt
Linearity: For x; y 2 X and all ; 2 F,
T .x C y/ D .x C y/0 .t/ D x 0 .t/ C y 0 .t/ D T x C T y:
T is not bounded: Let xn .t/ D t n ; n 2 N. Then

kT xn k
kxn k D 1; T xn D xn0 .t/ D nt n 1
; and kT xn k D D n:
kxn k
Hence T is unbounded.
4.1.3 Theorem
Let X and Y be normed linear spaces over a field F. Then T 2 L.X; Y / is bounded if and only if T maps
a bounded set into a bounded set.
66
Proof. Assume that T is bounded. That is, there exists a constant M > 0 such that kT xk M kxk for all
x 2 X . If kxk k for some constant k, then kT xk M kxk kM . That is, T maps a bounded set into
a bounded set.
Now assume that T maps a bounded set into a bounded set. Then T maps the unit ball B D fx 2 X W
kxk 1g into a bounded set. That is, there exists a constant M > 0 such that kT xk M for all x 2 B.
Therefore, for any nonzero x 2 X ,

kT xk x
D T M:
kxk kxk
Hence, kT xk M kxk. That is, T is bounded.

4.1.7 Exercise
Show that the inverse of a bounded linear operator is not necessarily bounded.
4.1.8 Proposition
1
Let T 2 B.X; Y /. Then T exists and is bounded if and only if there is a constant K > 0 such that
kT xk Kkxk for all x 2 X:
Proof. Assume that there is a constant K > 0 such that kT xk Kkxk for all x 2 X . If x 0, then
T x 0 and so T is one-to-one and hence T 1 exists. Also, given y 2 ran.T /, let y D T x for some
x 2 X . Then
1 1
kT 1 yk D kT 1 .T x/k D kxk kT xk D kyk;
K K
i.e., kT 1 yk K1 kyk for all y 2 Y . Thus T 1 is bounded.
Assume that T 1 exists and is bounded. Then for each x 2 X ,
1 1 1
kxk D kT .T x/k kT kkT xk 1k
kxk kT xk Kkxk kT xk;
kT
1
where K D kT 1k .
The following theorem asserts that continuity and boundedness are equivalent concepts for linear oper-
ators.
4.1.4 Theorem
Let X and Y be normed linear spaces over a field F and T 2 L.X; Y /. The following statements are
equivalent:
(1) T is continuous on X ;
(2) T is continuous at some point in X ;
(3) T is bounded on X .
Proof. The implication (1) ) (2) is obvious.

(2) ) (3): Assume that T is continuous at x 2 X , but T is not bounded on X . Then there is a sequence
.xn / in X such that kT xn k > nkxn k for each n 2 N. For each n 2 N, let
xn
yn D C x:
nkxnk
Then
1
kyn xk D ! 0as n ! 1I
n
67
n!1
i.e., yn ! x, but
kT xn k nkxnk
kT yn T xk D > D 1:
nkxn k nkxnk
That is, T yn 6! T x as n ! 1, contradicting (2).
(3) ) (1): Assume that T is bounded on X . Let .xn / be a sequence in X which converges to x 2 X .
Then
kT xn T xk D kT .xn x/k kT kkxn xk ! 0 as n ! 1:
Thus, T is continuous on X .
4.1.5 Theorem
Let .X; k k/ and .Y; k k/ be normed linear spaces with dim.X / < 1 and T W X ! Y be a linear
operator. Then T is continuous. That is, every linear operator on a finite-dimensional normed linear space
is automatically continuous.
Proof. Define a new norm k k0 on X by
kxk0 D kxk C kT xk for all x 2 X:
Since X is finte-dimensional, the norms k k0 and k k on X are equivalent. Hence there are constants
and such that
kxk0 kxk kxk0 for all x 2 X:
Hence,
1
kT xk kxk0 kxk D Kkxk;

where K D 1 . Therefore T is bounded.
4.1.9 Definition
Let X and Y be normed linear spaces over a field F.
(1) A sequence .Tn /1
1 in B.X; Y / is said to be uniformly operator convergent to T if
lim kTn T k D 0:
n!1
This is also referred to as convergence in the uniform topology or convergence in the operator
norm topology of B.X; Y /. In this case T is called the uniform operator limit of the sequence
.Tn /1
1 .
(2) A sequence .Tn /1

1 in B.X; Y / is said to be strongly operator convergent to T if
lim kTn x T xk D 0 for each x 2 X:

n!1
In this case T is called the strong operator limit of the sequence .Tn /1
1 .
Of course, if T is the uniform operator limit of the sequence .Tn /1

1 B.X; Y /, then T 2 B.X; Y /.
On the other hand, the strong operator limit T of a sequence .Tn /1
1 B.X; Y / need not be bounded in
general.
The following proposition asserts that uniform convergence implies strong convergence.
4.1.10 Proposition
If the sequence .Tn /1
1 in B.X; Y / is uniformly convergent to T 2 B.X; Y /, then it is strongly convergent
to T .
68
Proof. Since, for each x 2 X , kTn x T xk D k.Tn T /.x/k kTn T kkxk, if kTn T k ! 0 as
n ! 1, then
k.Tn T /.x/k ! 0 as n ! 1:
The converse of Proposition 4.1.10 does not hold.
4.1.11 Example
Consider the sequence .Tn / of operators, where for each n 2 N,
Tn W `2 ! `2 is given by
Tn .x1 ; x2; : : :/ D .0; 0; : : : ; 0; xnC1; xnC2 ; : : :/:
Let > 0 be given. Then for each x D .xi /1

iD1 2 `2 , there exists N such that
1
X
jxi j2 < 2 ; for all n N:
nC1
Hence, for all n N ,

1
X
kTn xk22 D jxi j2 < 2 :
nC1
That is, for each x 2 `2 , Tn x ! 0. Hence, Tn ! 0 strongly.

Now, since
1
X 1
X
kTn xk22 D jxi j2 jxi j2 D kxk22
nC1 1
for n 2 N and x D .xi /1 2 `2 , it follows that kTnk 1 for each n 2 N.

iD1
But kTn k 1 for all n. To see this, take x D .0; 0; : : : ; 0; xnC1; 0; : : :/ 2 `2 , where xnC1 6D 0.
Then
Tn x D x and hence kTnxk2 D jxnC1 j; and consequently; kTnk 1:
That is, .Tn / does not converge to zero in the uniform topology.
4.1.6 Theorem
Let X and Y be normed linear spaces over a field F. Then B.X; Y / is a Banach space if Y is a Banach
space.
Proof. We have shown that B.X; Y / is a normed linear space. It remains to show that it is complete if Y
is complete. To that end, let .Tn /1
1 be a Cauchy sequence in B.X; Y /. Then given any > 0 there exists a
positive integer N such that
kTn TnCr k < for all n > N;
whence,
kTn x TnCr xk kTn TnCr kkxk < kxk for all x 2 X: .4:1:6:1/
Hence, .Tn x/1
1is a Cauchy sequence in Y . Since Y is complete there exists y 2 Y such that Tn x ! y
as n ! 1. Set T x D y. We show that T 2 B.X; Y / and Tn ! T . Let x1 ; x2 2 X , and ; 2 F. Then
T .x1 C x2 / D lim Tn .x1 C x2/ D lim Tnx1 C Tn x2

n!1 n!1
D lim Tn x1 C lim Tn x2 D T x1 C T x2 :
n!1 n!1
That is, T 2 L.X; Y /. Taking the limit as r ! 1 in (4.1.6.1) we get that
k.Tn T /xk D kTn x T xk kxk for all n > N; and all x 2 X:
69
That is, Tn T is a bounded operator for all n > N . Since B.X; Y / is a linear space,
T D Tn .Tn T / 2 B.X; Y /.
Finally,
kTn T k D supfkTnx T xk W kxk 1g supfkxk W kxk 1g for all n > N:
That is, Tn ! T as n ! 1.
4.1.12 Definition
Let T W X ! Y and S W Y ! Z . We define the composition of T and S as the map ST W X ! Z
defined by
.ST /.x/ D .S T /.x/ D S.T x/:
4.1.7 Theorem
Let X , Y and Z be normed linear spaces over a field F and let T 2 B.X; Y / and S 2 B.Y; Z/. Then
ST 2 B.X; Z/ and kST k kSkkT k.
Proof. Since linearity is trivial, we only prove boundedness of ST . Let x 2 X . Then
k.ST /.x/k D kS.T x/k kSkkT xk kSkkT kkxk:
Thus, kST k kSkkT k.
Let X be a normed linear space over F. For S; T1 ; T2 2 B.X / it is easy to show that
.ST1 /T2 D S.T1 T2 /

S.T1 C T2 / D ST1 C ST2
.T1 C T2 /S D T1S C T2 S:
The operator I defined by I x D x for all x 2 X belongs to B.X /; kI k D 1, and it has the property that
I T D T I D T for all T 2 B.X /. We call I the identity operator. The set B.X / is therefore an algebra
with an identity element. In fact, B.X / is a normed algebra with an identity element. If X is a Banach
space then B.X / is a Banach algebra.
We now turn our attention to a very special and important class of bounded linear operators, namely,
bounded linear functionals.
4.1.13 Definition
Let X be a linear space over F. A linear operator f W X ! F is called a linear functional on X . Of course,
L.X; F/ denotes the set of all linear functionals on X .
Since every linear functional is a linear operator, all of the foregoing discussion on linear operators
applies equally well to linear functionals. For example, if X is a normed linear space then we say that
f 2 L.X; F/ is bounded if there exists a constant M > 0 such that jf .x/j M kxk for all x 2 X . The
norm of f is defined by
kf k D supfjf .x/j W kxk 1g:
We shall denote by X D B.X; F/ the set of all bounded (i.e., continuous) linear functionals on X . We
call X the dual of X . It follows from Theorem 4.1.6 that X is always a Banach space under the above
norm.
70
4.1.14 Examples
[1] Let X D Ca; b. For each x 2 X , define f W X ! R by
Zb
f .x/ D x.t/dt:
a
Then f is a bounded linear functional on X .

Linearity: For any x; y 2 X and any ; 2 R,
Zb Zb Zb
f .x C y/ D .x C y/.t/dt D x.t/dt C y.t/dt D f .x/ C f .y/:
a a a
Boundedness:
b
Z Zb

jf .x/j D x.t/dt jx.t/jdt max jx.t/j.b
a/ D kxk1 .b a/:
at b
a a
Hence f is bounded and kf k b a. We show that kf k D b a. Take x D 1, the constant

function 1. Then
Zb
f .1/ D dt D b a; i.e., jf .1/j D b a:
a
Hence
jf .1/j jf .x/j
b aD sup W x 6D 0 D kf k b a:
1 kxk
That is, kf k D b a.
[2] Let X D Ca; b and let t 2 .a; b/ be fixed. For each x 2 X , define t W X ! R by
t .x/ D x.t/; (i.e., t is a point evaluation at t/:
Then t is a bounded linear functional on X . Linearity of t is easy to verify.

Boundedness: For each x 2 X ,
j t .x/j D jx.t/j max jx./j D kxk1 :

a b
That is, t is a bounded linear operator and k t k 1. We show that k t k D 1. Take x D 1,

the constant 1 function. Then t .1/ D 1 and so
j t .1/j
1D sup fj t .x/j W kxk D 1g D k t k 1:
1
That is, k t k D 1.
[3] Let c1 ; c2; : : : ; cn be real numbers and let X D Ca; b. Define f W X ! R by
n
X
f .x/ D ci x.ti /; where t1 ; t2 ; : : : ; tn are in a; b:
iD1
Then f is a bounded linear functional on X .
71
Linearity: Clear.
Boundedness: For any x 2 X ,
n
X X n n
X n
X

jf .x/j D ci x.ti / jci x.ti /j D jci jjx.ti /j kxk1 jci j:

iD1 iD1 iD1 iD1
Hence, f is bounded and

n
X
kf k jci j:
iD1
[4] Let X be a linear space. The norm k k W X ! R is an example of a nonlinear functional on

X.
4.2 Examples of Dual Spaces

4.2.1 Definition
Let X and Y be normed linear spaces over the same field F. Then X and Y are said to be isomorphic to
each other, denoted by X ' Y , if there is a bijective linear operator T from X onto Y . If, in addition, T is
an isometry, i.e., kT xk D kxk for each x 2 X , then we say that T is an isometric isomorphism. In this
case, X and Y are are said to be isometrically isomorphic and we write X Y .
Two normed linear spaces which are isometrically isomorphic can be regarded as identical, the isometry
merely amounting to a relabelling of the elements.
4.2.2 Proposition
Let X and Y be normed linear spaces over the same field F and T a linear operator from X onto Y . Then
T is an isometry if and only if
(i) T is one-to-one;
(ii) T is continuous on X ;
1
(iii) T has a continuous inverse (in fact, kT k D kT k D 1);
(iv) T is distance-preserving: For all x; y 2 X , kT x T yk D kx yk.
Proof. If T satisfies (iv), then, taking y D 0, we have that kT xk D kxk for each x 2 X ; i.e., T is an
isometry.
Conversely, assume that T is an isometry. If x y, then
kT x T yk D kT .x y/k D kx yk > 0:
Hence, T x T y. This shows that T is one-to-one and distance-preserving. Since kT xk D kxk for each
x 2 X , it follows that T is bounded and kT k D 1. By Theorem 4.1.4, T is continuous on X .
Let y1 ; y2 2 Y and 1 ; 2 2 F. Then there exist x1; x2 2 X such that T xi D yi for i D 1; 2. Therefore
1 y1 C 2 y2 D 1 T x1 C 2T x2 D T .1 x1 C 2 x2 / or
1 1 1
T .1 y1 C 2 y2 / D 1 x 1 C 2 x 2 D 1 T y1 C 2T y2 :
1 1
That is, T is linear. Furthermore, for y 2 Y , let x D T y. Then,
1
kT yk D kxk D kT xk D kyk:
1 1
Therefore T is bounded and kT k D 1.
72
4.2.3 Remark
It is clear from Proposition 4.2.2 that two normed linear spaces X and Y are isometrically isomor-
phic if and only if there is a linear isometry from X onto Y .
[1] The dual of `1 is (isometrically isomorphic to) `1 ; i.e., `1 `1 .

Proof. Let y D .yn / 2 `1 and define W `1 ! `1 by
1
X
.y/.x/ D xj yj for x D .xn / 2 `1 :
jD1
Claim 1: y 2 `1 .
Linearity of y: Let x D .xn / ; z D .zn / 2 `1 and 2 F. Then
1
X 1
X 1
X
.y/.x C z/ D .xj C zj /yj D xj yj C zj yj
jD1 jD1 jD1
1
X 1
X
D xj yj C zj yj
jD1 jD1
D .y/.x/ C .y/.z/:
Boundedness of y: For any x D .xn / 2 `1 ,

1 1 1
X X X
j.y/.x/j D xj yj jxj yj j kyk1 jxj j D kyk1 kxk1 :
jD1 jD1 jD1
That is, y 2 `1 and

kyk kyk1 : .?/
Claim 2: is a surjective linear isometry.

(i) is a surjective: A basis for `1 is .en /, where en D .nm / has 1 in the n-th position and zeroes
X1
elsewhere. Let f 2 `1 and x D .xn / 2 `1 . Then x D xn en and therefore
nD1
1
X 1
X
f .x/ D xn f .en / D xn zn ;
nD1 nD1
where, for each n 2 N, zn D f .en /. We show that z D .zn / 2 `1 . Indeed, for each n 2 N
jzn j D jf .en /j kf kken k D kf k:
Hence, z D .zn / 2 `1 . Also, for any x D .xn / 2 `1 ,

1
X 1
X
.z/.x/ D xn zn D xn f .en / D f .x/:
nD1 nD1
That is, z D f and so is surjective. Furthermore,
kzk1 D sup jzn j D sup jf .en /j kf k D kzk: .??/

n2N n2N
73
(ii) is linear: Let y D .yn /; z D .zn / 2 `1 and 2 F. Then, for any x D .xn / 2 `1 ,
1
X 1
X 1
X
.y C z/.x/ D xj .yj C zj / D xj yj C xj zj
jD1 jD1 jD1
D .y/.x/ C .z/.x/ D y C z.x/:
Hence, .y C z/ D y C , which proves linearity of .

(iii) is an isometry: This follows from (?) and (??).
[2] The dual of c0 is (isometrically isomorphic to) `1 , i.e., c0 `1 .
Proof. Let y D .yn / 2 `1 and define W `1 ! c0 by
1
X
.y/.x/ D xj yj for x D .xn/ 2 c0 :
jD1
Proceeding as in Example 1 above, one shows that y is a bounded linear functional on c0 and
kyk kyk1 : .?/
Claim: is a surjective linear isometry.

(i) is a surjective: A basis for c0 is .en /, where en D .nm / has 1 in the n-th position and zeroes
X1
elsewhere. Let f 2 c0 and x D .xn / 2 c0 . Then x D xn en and therefore
nD1
1
X 1
X
f .x/ D xn f .en / D xn wn ;
nD1 nD1
where, for each n 2 N, wn D f .en /. For n; k 2 N, let

8

jwk j
< if wk 0 and k n
znk D wk

:
0 if wk D 0 or k > n;
and let
zn D .zn1 ; zn2 ; : : : ; znn ; 0; 0; : : :/:
Then zn 2 c0 and
kzn k1 D sup jznk j D 1:
k2N
Also,
1
X n
X
f .zn / D znk wk D jwk j:
kD1 kD1
Hence, for each n 2 N,

n
X
jwk j D jf .zn /j kf kkzn k kf k:
kD1
74
1
X
Since the right hand side is independent of n, it follows that jwk j kf k Hence, w D
kD1
.wn / 2 `1 . Also, for any x D .xn/ 2 c0 ,
1
X 1
X
.w/.x/ D xn wn D xn f .en / D f .x/:
nD1 nD1
That is, w D f and so is surjective. Furthermore,

1
X
kwk1 D jwk j kf k D kwk: .??/
kD1
(ii) is linear: Let y D .yn /; z D .zn / 2 `1 and 2 F. Then, for any x D .xn / 2 c0 ,
1
X 1
X 1
X
jD1 jD1 jD1
D .y/.x/ C .z/.x/ D y C z.x/:
Hence, .y C z/ D y C z, which proves linearity of .

1 1
[3] Let 1 < p < 1; p
C q
D 1. Then the dual of `p is (isometrically isomorphic to) `q , i.e., `p `q .
Proof. Let y D .yn / 2 `q and define W `q ! `p by
1
X
.y/.x/ D xj yj for x D .xn / 2 `p :
jD1
It is straightforward to show that y is linear. We show that y is bounded. By Holders Inequality,

0 1 p1 0 1 1q
1 1 1 1
X X X X
j.y/.x/j D xj yj jxj yj j @ jxj jp A @ jyj jq A D kxkp kykq :
jD1 jD1 jD1 jD1
That is, y 2 `p and

kyk kykq : .?/
Claim: is a surjective linear isometry.
(i) is a surjective: A basis for `p is .en /, where en D .nm / has 1 in the n-th position and zeroes
X1
elsewhere. Let f 2 `p and x D .xn / 2 `p . Then x D xn en and therefore
nD1
1
X 1
X
f .x/ D xn f .en / D xn wn ;
nD1 nD1
where, for each n 2 N, wn D f .en /. For n; k 2 N, let

8

jwk jq
< if k n and wk 0
znk D wk

:
0 if wk D 0 or k > n;
75
and let
zn D .zn1 ; zn2 ; : : : ; znn ; 0; 0; : : :/:
Then zn 2 `p and
1
X n
X
f .zn / D znk wk D jwk jq :
kD1 kD1
Hence, for each n 2 N,

n
X
jwk jq D jf .zn /j kf kkzn kp :
kD1
Since
1
!1=p n
!1=p
X X
p p
kzn kp D jznk j D jznk j
kD1 kD1
n
!1=p n
!1=p
X X
D jwk jp.q 1/
D jwk jq ;
kD1 kD1
it follows that , for each n 2 N,

n n n
!1=p
X X X
q q q
jwk j kf kkzn kp jwk j kf k jwk j
kD1 kD1 kD1
n
!1 1=p
X
jwk jq kf k
kD1
n
!1=q
X
q
jwk j kf k:
kD1
1
!1=q
X
q
Since the right hand side is independent of n, it follows that jwk j kf k, and so
kD1
w D .wn / 2 `q . Also, for any x D .xn / 2 `p ,
1
X 1
X
.w/.x/ D xn wn D xn f .en / D f .x/:
nD1 nD1
That is, w D f and so is surjective. Furthermore,
1
!1=q
X
q
kwkq D jwk j kf k D kwk: .??/
kD1
(ii) is linear: Let y D .yn /; z D .zn / 2 `q and 2 F. Then, for any x D .xn / 2 `p ,
1
X 1
X 1
X
jD1 jD1 jD1
D .y/.x/ C .z/.x/ D y C z.x/:
Hence, .y C z/ D y C z, which proves linearity of .
76
The following result is an immediate consequence of Theorem 4.1.5.
4.2.1 Theorem
Every linear functional on a finite-dimensional normed linear space is continuous.
4.2.4 Proposition
Let X be a normed linear space over F. If X is finite-dimensional, then X is also finite-dimensional and
dimX D dimX .
Proof. Let fx1 ; x2; ; xn g be a basis for X . For each j D 1; 2; : : : ; n, let xj be defined by xj .xk / D
jk for k D 1; 2; : : : ; n. Then each xj is a bounded linear functional on X . We show that fxj j j D
1; 2; : : : ; ng is a basis for X . Let x be an element of X and define j D x .xj / for each j D
1; 2; : : : ; n. Then for any k D 1; 2; : : : ; n,
0 1
Xn Xn
@ j xj A .xk / D j jk D k D x .xk /:
jD1 jD1
n
X
Hence x D j xj ; i.e., fxj j j D 1; 2; : : : ; ng spans X . It remains to show that fxj j j D
jD1
n
X
1; 2; : : : ; ng is linearly independent. Suppose that j xj D 0. Then, for each k D 1; 2; : : : ; n,
jD1
0 1
n
X n
X
0D@ j xj A .xk / D j jk D k :
jD1 jD1
Hence fxj j j D 1; 2; : : : ; ng is a linearly independent set.
4.3 The Dual Space of a Hilbert Space

If H is a Hilbert space then bounded linear functionals on H assume a particularly simple form.
Let .X; h; i/ be an inner product space over a field F. Choose and fix y 2 X n f0g. Define a map
fy W X ! F by fy .x/ D hx; yi. We claim that fy is a bounded (= continuous) linear functional on X .
Linearity: Let x1 ; x2 2 X and ; 2 F. Then
fy .x1 C x2 / D hx1 C x2; yi D hx1; yi C hx2 ; yi D fy .x1 / C fy .x2 /:
Boundedness: For any x 2 X ,
jfy .x/j D jhx; yij kxkkyk (by the CBS Inequality):
That is, fy is bounded and kfy k kyk. Since
jfy .y/j
fy .y/ D hy; yi D kyk2 ) D kyk;
kyk
we have that kfy k D kyk.
77
The above observation simply says that each element y in an inner product space .X; h; i/ determines
a bounded linear functional on X .
The following theorem asserts that if H is a Hilbert space then the converse of this statement is true.
That is, every bounded linear functional on a Hilbert space H is, in fact, determined by some element
y 2 H.
4.3.1 Theorem

(Riesz-Frechet Theorem). Let H be a Hilbert space over F. If f W H ! F is a bounded linear functional
on H (i.e., f 2 H ) then there exists one and only one y 2 H such that
f .x/ D hx; yi for all x 2 H:
Moreover, kf k D kyk.
Proof. Existence: If f D 0 then take y D 0. Assume that f 6D 0. Let N D fx 2 H j f .x/ D 0g, the
kernel of f . Then N is a closed proper subspace of H. By Corollary 3.4.5 there exists z 2 N ? n f0g.
Without loss of generality, kzk D 1. Put u D f .x/z f .z/x. Then
f .u/ D f .f .x/z f .z/x/ D f .x/f .z/ f .z/f .x/ D 0; i.e., u 2 N:
Thus,
0 D hu; zi D hf .x/z f .z/x; zi D f .x/hz; zi f .z/hx; zi D f .x/ f .z/hx; zi;
whence, f .x/ D f .z/hx; zi D hx; f .z/zi. Take y D f .z/z. Then f .x/ D hx; yi.
Uniqueness: Assume that f .x/ D hx; yi D hx; y0 i for each x 2 H. Then
0 D hx; yi hx; y0 i D hx; y y0 i; for all x 2 H:
In particular, take x D y y0 ,
0 D hy y0 ; y y0 i D ky y0 k2 ) y y0 D 0 ) y D y0 :
Finally, for any x 2 H,
jf .x/j D jhx; yij kxkkyk (by the CBS Inequality):
That is, kf k kyk. Since

jf .y/j
f .y/ D hy; yi D kyk2 ) D kyk;
kyk
we have that kf k D kyk.
4.3.1 Remarks
(a) The element y 2 H as advertised in Theorem 4.3.1 is called the representer of the functional
f.
(b) The conclusion of Theorem 4.3.1 may fail if .X; h; i/ is an incomplete inner product space.
4.3.2 Example
Let X be the linear space of polynomials over R with the inner product defined by
Z1
hx; yi D x.t/y.t/ dt:
0
78
For each x 2 X , let f W X ! R be defined by
f .x/ D x.0/; (i.e., f is a point evaluation at 0/:
Then f is a bounded linear functional on X . We show that there does not exist an element y 2 X
such that
f .x/ D hx; yi for all x 2 X:
Assume that such an element exists. Then for each x 2 X
Z1
f .x/ D x.0/ D x.t/y.t/ dt:
0
Since for any x 2 X the functional f maps the polynomial tx.t/ onto zero, we have that
Z1
tx.t/y.t / dt D 0 for all x 2 X:
0
In particular, with x.t/ D ty.t/ we have that

Z1
t 2 y.t/2 dt D 0I
0
whence y 0, i.e. y is the zero polynomial. Hence, for all x 2 X ,
f .x/ D hx; yi D hx; 0i D 0:
That is, f is the zero functional, a contradiction since f maps a polynomial with a nonzero con-
stant term to that constant term.
4.3.2 Theorem
Let H be a Hilbert space.
(a) If H is a real Hilbert space, then H H .
(b) If H is a complex Hilbert space, then H is isometrically embedded onto H .
Proof. For each y 2 H, define W H ! H by
y D fy ; where fy .x/ D hx; yi for each x 2 H:
Let y; z 2 H. Then, for each x 2 H,
y z hx; yi hx; zi fy fz y z:
Hence, is well defined and one-to-one. Furthermore, since
kyk D kfy k D kyk
for each y 2 H, is an isometry.

If f 2 H , then by Riesz-Frechet Theorem (Theorem 4.3.1), there is a unique yf 2 H such that f .x/ D
hx; yf i. Hence yf D f , i.e., is onto.
The inverse 1 of is given by
1
f D y; where f .x/ D hx; yi for all x 2 H:
79
Since k 1 f k D kyk D kf k for each f 2 H , 1 is bounded (in fact an isometry).

If H is real, then is linear. Indeed, for all x; y; z 2 H and all 2 R, then
..y C z// .x/ D f.yCz/ .x/ D hx; y C zi

D hx; yi C hx; zi D hx; yi C hx; zi
D .y/.x/ C .z/.x/ D .y C z/ .x/:
Hence, .y C z/ D y C z.
If H is complex, then is conjugate-linear; i.e., .y C z/ D y C z.
4.3.3 Exercise
[1] Let X and Y be linear spaces over the same field F and T 2 L.X; Y /.
(a) Show that ran.T / is a linear subspace of Y and ker.T / is a linear subspace of X .
(b) T is one-to-one if and only if ker.T / D f0g.
[2] Let X and Y be normed linear spaces over the same field F. Show that if T 2 B.X; Y / then
ker.T / is a closed linear subspace of X .
[3] Show that the mapping R W `2 ! `2 given by
Rx D R.x1 ; x2 ; x3; : : :/ D .0; x1 ; x2; x3; : : :/
is a bounded linear operator on `2 and find its norm. The operator R is called the right-shift
operator.
[4] Fix x 2 C ; . Define an operator Mx W L2 ; ! L2 ; by
Mx y D xy where .Mx y/.t/ D x.t/y.t/ for all t 2 ; :
Show that Mx is a bounded linear operator on L2 ; . The operator Mx is called a

multiplication operator. The function x is the symbol of Mx .
[5] Fix x D .x1 ; x2 ; : : :/ 2 `1 . Define an operator Mx W `2 ! `2 by
Mx y D Mx .y1 ; y2 ; : : :/ D .x1 y1 ; x2y2 ; : : :/:
Show that Mx is a bounded linear operator on `2 and kMx k D kxk1.

[6] Show that if S is a subset of a Hilbert space H that is dense in H and T1 and T2 are operators
such that T1 x D T2 x for all x 2 S, then T1 D T2 .
[7] Find the general form of a bounded linear functional on L2 ; .
[8] Find the general form of a bounded linear functional on `2 .
[9] Define f W `2 ! C by
1
X xn
f .x/ D ; where x D .x1 ; x2; : : :/ 2 `2 :
n2
nD1
2
Show that f is a bounded linear functional on `2 and that kf k D p .
3 10
80
Chapter 5
The Hahn-Banach Theorem and its

Consequences
The Hahn-Banach theorem is one of the most important results in functional analysis since it is required
for many other results and also because it encapsulates the spirit of analysis. The theorem was proved
independently by Hahn in 1927 and by Banach in 1929 although Helly proved a less general version much
earlier in 1912. Intersetingly, the complex version was proved only in 1938 by Bohnenblust and Sobczyk.
We prove the Hahn-Banach theorem using Zorns lemma which is equivalent to the axiom of choice. It
should be noted, however, that the Hahn-Banach is in fact strictly weaker than the axiom of choice. Since
the publication of the original result, there have been many versions published in different settings but that
is beyond the scope of this course.
5.1 Introduction
In this chapter the Hahn-Banach theorem is established along with a few of its many consequences. Before
doing that, we briefly discuss Zorns Lemma.
5.1.1 Definition
A binary relation 4 on a set P is a partial order if it satisfies the following properties: For all x; y; z 2 P ,
(i) 4 is reflexive: x 4 x ;
(ii) 4 is antisymmetric: if x 4 y and y 4 x , then x D y ;
(iii) 4 is transitive: if x 4 y and y 4 z , then x 4 z .
A partially ordered set is a pair .P; 4/, where P is a set 4 is a partial order on P .
5.1.2 Examples
[1] Let P D R and take 4 to be , the usual less than or equal to relation on R.
[2] Let P D P.X / the power set of a set X and take 4 to be , the usual set inclusion relation.
[3] Let P D C0; 1, the space of continuous real-valued functions on the interval 0; 1 and take
4 to be the relation given by f g if and only if f .x/ g.x/ for each x 2 0; 1.
5.1.3 Definition
Let C be a subset of a partially ordered set .P; 4/.
(i) An element u 2 P is an upper bound of C if x 4 u for every x 2 C ;
81
(ii) An element m 2 C is said to be maximal if for any element y 2 C , the relation m 4 y implies that
m D y.
5.1.4 Definition
Let .P; 4/ be a partially ordered set and x; y 2 P . We say that x and y are comparable if either x 4 y or
y 4 x . Otherwise, x and y are incomparable.
A partial order 4 is called a linear order (or a total order) if any two elements of P are comparable. In
this case we say that .P; 4/ is a linearly ordered (or totally ordered) set. A linearly ordered set is also called
a chain.
5.1.1 Theorem
(Zorns Lemma). Let .P; 4/ be a partially ordered set. If each linearly ordered subset of P has an upper
bound, then P has a maximal element.
5.1.5 Definition
Let M and N be linear subspaces of a linear space X with M N and let f be a linear functional on M .
A linear functional F on N is called an extension of f to N if F jM D f ; i.e., F.x/ D f .x/ for each
x 2 M.
5.1.6 Definition
Let X be a linear space. A function p W X ! R is called a sublinear functional provided that:
(i) p.x C y/ p.x/ C p.y/ for x; y 2 X ;
(ii) p.x/ D p.x/; 0.
Observe that any linear functional or any norm on X is a sublinear functional. Also, every positive
scalar multiple of a sublinear functional is again a sublinear functional.
5.1.7 Lemma
Let M be a proper linear subspace of a real linear space X , x0 2 X nM , and N D fmCx0 j m 2 M; 2
Rg: Suppose that p W X ! R a sublinear functional defined on X , and f a linear functional defined on M
such that f .x/ p.x/ for all x 2 M . Then f can be extended to a linear functional F defined on N such
that F.x/ p.x/ for all x 2 N .
Proof. Since x0 62 M , it is readily verified that N D M linfx0g. Therefore each x 2 N has a unique
representation of the form x D m C x0 for some unique m 2 M and 2 R. Define a functional F on N
by
F.x/ D f .m/ C c for some c 2 R:
This functional F is well defined since each x 2 N is uniquely determined. Furthermore F is linear and
F.y/ D f .y/ for all y 2 M . It remains to show that it is possible to choose a c 2 R such that for each
x 2 N,
F.x/ p.x/:
Let y1 ; y2 2 M . Since f .y/ p.y/ for all y 2 M , we have that
f .y1 / f .y2 / D f .y1 y2 / p.y1 y2 / D p.y1 C x0 y2 x0 /

p.y1 C x0 / C p. y2 x0 /
f .y2 / p. y2 x0 / p.y1 C x0 / f .y1 /:
Therefore, for fixed y1 2 M , the set of real numbers f f .y2 / p. y2 x0 / j y2 2 M g is bounded above
and hence has the least upper bound. Let
a D supf f .y2 / p. y2 x0 / j y2 2 M g:
82
Similarly, for fixed y2 2 M , the set fp.y1 C x0 / f .y1 / j y1 2 M g is bounded below. Let
b D inffp.y1 C x0/ f .y1 / j y1 2 M g:
Of course, a b. Hence there is a real number c such that a c b. Therefore
f .y/ p. y x0 / c p.y C x0 / f .y/
for each y 2 M .
Now, let x D y C x0 2 N . If D 0, then F.x/ D f .x/ p.x/. If > 0, then
y
cp C x0 f .y=/ c p.y C x0/ f .y/

f .y/ C c p.y C x0 /
F.x/ p.x/:
Finally, if < 0, then

1 1
f .y=/ p . y= x0 / c f .y/ C p.y C x0 / c

f .y/ p.y C x0/ c
f .y/ C c p.y C x0/
F.x/ p.x/:
We now state our main result. What this theorem essentially states is that there are enough bounded
(continuous) linear functionals for a rich theory and as mentioned before it is used ubiquitously thoughout
functional analysis.
5.1.2 Theorem
(Hahn-Banach Extension Theorem for real linear spaces). Let p be a sublinear functional on a real
linear space X and let M be a subspace of X . If f is a linear functional on M such that f .x/ p.x/ for
all x 2 M , then f has an extension F to X such that F.x/ p.x/ for all x 2 X .
Proof. Let F be the set of all pairs .M ; f /, where M is a subspace of X containing M , f .y/ D f .y/
for each y 2 M , i.e., f is an extension of f , and f .x/ p.x/ for all x 2 M . Clearly, F 6D ; since
.M; f / 2 F . Define a partial order on F by:
.M ; f / .M ; f / M M and f jM D f :
Let T be a totally ordered subset of F and let

[
X0 D fM j .M ; f / 2 T g:
Then X0 is a linear subspace of X since T is totally ordered. Define a functional f0 W X0 ! R by f0 .x/ D

f .x/ for all x 2 M . Then f0 is well-defined, since if x 2 M \ M , then x 2 M and x 2 M .
Therefore f0 .x/ D f .x/ and f0 .x/ D f .x/. By total ordering of T , either f extends f or vice versa.
Hence f .x/ D f .x/. It is clear that f0 is a linear extension of f . Furthermore f0 .x/ p.x/ for all
x 2 X0 and .M ; f / .X0 ; f0 / for all .M ; f / 2 T , i.e., .X0 ; f0 / is an upper bound for T . By Zorns
lemma, F has a maximal element .X1 ; F /. To complete the proof, it suffices to show that X1 D X . If
X1 6D X , then choose y 2 X n X1 . By Lemma 5.1.7, we can extend F to a linear functional F e defined
e e e
on X D X1 linfyg and extending f such that F .x/ p.x/ for all x 2 X . Thus .X ; F / 2 F and e e
.X1 ; F / .Xe; F
e/, which contradicts the maximality of .X1 ; F /.
5.1.8 Definition
A seminorm p on a (complex) linear space X is a function p W X ! R such that for all x; y 2 X and
2 C,
83
(i) p.x/ 0 and p.0/ D 0;

(ii) p.x C y/ p.x/ C p.y/, and
(iii) p.x/ D jjp.x/.
5.1.3 Theorem
(Hahn-Banach Extension Theorem for (complex) linear spaces). Let X be a real or complex linear
space, p be a seminorm on X and f a linear functional on a linear subspace M of X such that jf .x/j
p.x/ for all x 2 M . Then there is a linear functional F on X such that F jM D f and jF.x/j p.x/ for
all x 2 X .
Proof. Assume first that X is a real linear space. Then, by Theorem 5.1.2, there is an extension F of f
such that F.x/ p.x/ for all x 2 X . Since
F.x/ D F. x/ p. x/ D p.x/ for all x 2 X;
it follows that p.x/ F.x/ p.x/, or jF.x/j p.x/ for all x 2 X .
Now assume that X is a complex linear space. Then we may regard X as a real linear space by restricting
the scalar field to R. We denote the resulting real linear space by XR and the real linear subspace by MR .
Write f as f D f1 C if2, where f1 and f2 are real linear functionals given by f1 .x/ D Ref .x/ and
f2 .x/ D Imf .x/. Then f1 is a real linear functional of MR and f1 .x/ jf .x/j p.x/ for all x 2 MR .
Hence, by Theorem 5.1.2, f1 has a real linear extension F1 such that F1 .x/ p.x/ for all x 2 XR . Since,
f .i x/ D if .x/ f1 .i x/ C if2.i x/ D if1.x/ f2 .x/
f2 .x/ D f1 .i x/ and f2 .i x/ D f1 .x/;
we can write f .x/ D f1 .x/ if1.i x/: Set
F.x/ D F1 .x/ iF1 .i x/ for all x 2 X:
Then F is a real linear extension of f and, for all x; y 2 X ,
F.x C y/ D F1 .x C y/ iF1 .i x C iy/ D F1 .x/ iF1 .i x/ C F1 .y/ iF1 .iy/
D F.x/ C F.y/:
For all x 2 X ,
F.i x/ D F1 .i x/ iF1 . x/ D F1 .i x/ C iF1 .x/ D i .F1 .x/ iF1 .i x// D iF.x/:
If D a C bi for a; b 2 R, and x 2 X , then
F.x/ D F..a C bi /x/ D F.ax C bi x/ D F.ax/ C F.bi x/
D aF.x/ C bF.i x/ D aF.x/ C biF.x/ D .a C bi /F.x/
D F.x/:
Hence, F is also complex linear. Finally, for x 2 X , write F.x/ D jF.x/je i . Then, since ReF D F1 ,
i i i i i
jF.x/j D F.x/e D F.xe / D F1 .xe / p.xe / D je jp.x/ D p.x/:
Suppose that M is a subspace of a normed linear space X and f is a bounded linear functional on M .
If F is any extension of f to X , then the norm of F is at least as large as kf k because
kF k D supf jF.x/j W x 2 X; kxk 1 g supf jF.x/j W x 2 M; kxk 1 g
D supf jf .x/j W x 2 M; kxk 1 g D kf k:
84
The following consequence of the Hahn-Banach theorem states that it is always possible to find a
bounded extension of f to the whole space which has the same, i.e., smallest possible, norm.
5.1.4 Theorem
(Hahn-Banach Extension Theorem for Normed linear spaces). Let M be a linear subspace of the
normed linear space .X; k k/ and let f 2 M . Then there exists an extension x 2 X of f such that
kx k D kf k.
Proof. Define p on X by p.x/ D kf kkxk. Then p is a seminorm on X and jf .x/j p.x/ for all
x 2 M . By Theorem 5.1.3, f has an extension F to X such that jF.x/j p.x/ for all x 2 X . That is,
jF.x/j kf kkxk. This shows that F is bounded and kF k kf k. Since F must have norm at least as
large as kf k, kF k D kf k and the result follows
with x D F .
5.2 Consequences of the Hahn-Banach Extension Theorem

5.2.1 Theorem
Let M be a linear subspace of a normed linear space .X; k k/ and x 2 X such that
d D d.x; M / WD inf kx yk > 0:

y2M
Then there is an x 2 X such that

(i) kx k D 1
(ii) x .x/ D d
(iii) x .m/ D 0 for all m 2 M .
Proof. Let Y D M C linfxg WD fm C x; m 2 M; 2 Fg. Then each y in Y is uniquely expressible in

the form y D m C x for some m 2 M and some scalar . Indeed, if
y D m1 C x D m2 C x
for some m1 ; m2 2 M and some scalars and , then . /x D m1 m2 2 M .

Claim: D . If 6D , then since M is a subspace
1
xD .m1 m2 / 2 M;

a contradiction since x 62 M . Hence, D and consequently m1 D m2 .

Define f W Y ! F by
f .y/ D f .m C x/ D d:
Since the scalar is uniquely determined, f is well defined.
Claim: f is a bounded linear functional on Y .
Linearity: Let y1 D m1 C 1 x and y2 D m2 C 2x be any two elements of Y and 2 F. Then
f .y1 C y2 / D f ..m1 C m2 / C .1 C 2 /x/ D .1 C 2 /d D 1d C 2 d D f .y1 / C f .y2 /:
Boundedness: Let y D m C x 2 Y . Then

m m

kyk D km C xk D jj C x D jj x jjd D jf .y/j;

85
i.e., jf .y/j kyk for all y 2 Y . Thus, f is bounded and kf k 1.

We show next that kf k D 1. By definition of infimum, given any > 0, there is an element m 2 M such
x m
that kx m k < d C . Let z D . Then z 2 Y; kzk D 1 and
kx m k
d d
jf .z/j D > :
kx m k d C
Since is arbitrary, it follows that jf .z/j 1. Thus
1 jf .z/j kf kkzk D kf k:
Thus, kf k D 1.
It is clear that and f .m/ D 0 for all m 2 M and f .x/ D d . By Theorem 5.1.4, there is an x 2 X such
that
x .y/ D f .y/ for all y 2 Y and kx k D kf k:
Hence, kx k D 1 and x .m/ D 0 for all m 2 M and x .x/ D d .
5.2.1 Corollary
Let .X; k k/ be a normed linear space and x0 2 X n f0g. Then there exists an x 2 X , such that
x .x0 / D kx0 k and kx k D 1.
Proof. Consider M D f0g. Since x0 2 X nf0g, it follows that x0 62 M and so d D d.x0 ; M / D kx0 k > 0:
By Theorem 5.2.1, there is an x 2 X such that x .x0 / D kx0 k and kx k D 1.
The following result asserts that X is big enough to distinguish the points of X .
5.2.2 Corollary
Let .X; k k/ be a normed linear space and y; z 2 X . If y 6D z , then there exists an x 2 X , such that
x .y/ 6D x .z/.
Proof. Consider M D f0g. Since y z, it follows that y z 62 M and consequently
d D d.y z; M / D ky zk > 0:
By Theorem 5.2.1, there is an x 2 X such that x .y z/ D d > 0. Hence x .y/ 6D x .z/.
5.2.3 Corollary
For each x in a normed linear space .X; k k/,
kxk D supfjx .x/j j x 2 X ; kx k D 1g:
Proof. If x D 0, then the result holds vacuously. Assume x 2 X n f0g. For any x 2 X with kx k D 1,
jx .x/j kx kkxk D kxk:
Hence, supfjx .x/j j x 2 X ; kx k D 1g kxk.

By Corollary 5.2.1, there is a x 2 X such that kx k D 1 and x .x/ D kxk. Therefore
kxk D jx .x/j supfjx .x/j j x 2 X ; kx k D 1g;
whence kxk D supfjx .x/j j x 2 X ; kx k D 1g:
5.2.2 Theorem
If the dual X of a normed linear space .X; k k/ is separable, then X is also separable.
86
Proof. Let S D S.X / D fx 2 X j kx k D 1g. Since any subset of a separable space is separable, S
is separable. Let fxn j n 2 Ng be a countable dense subset of S. Since xn 2 S for each n 2 N, we have
that kxn k D 1. Hence, for each n 2 N there is an element xn 2 X such that kxn k D 1 and jxn .xn /j > 12 .
(Otherwise jxn .x/j 21 for all x 2 X and so kxn k 12 , a contradiction.) Let
M D lin.fxn j n 2 Ng/:
Then M is separable since M contains a countable dense subset comprising all linear combinations of the
xn s with coefficients whose real and imaginary parts are rational.
Claim: M D X . If M 6D X , then there is an element x0 2 X n M such that d D d.x0 ; M / > 0. By
Theorem 5.2.1, there is an x 2 X such that kx k D 1, i.e. x 2 S, and x .y/ D 0 for all y 2 M . In
particular, x .xn / D 0 for all n 2 N. Now, for each n 2 N,
1
< jxn .xn /j D jxn .xn / x .xn /j D j.xn x /.xn /j kxn x k:
2
But this contradicts the fact that the set fxn j n 2 Ng is dense in S. Hence M D X and, consequently, X
is separable.
The converse of Theorem 5.2.2 does not hold. That is, if X separable, it does not follow that its dual
X is also be separable. Take, for example, `1 . Its dual is (isometrically isomorphic to) `1 . The space `1
is separable whereas `1 is not. This also shows that the dual of `1 is not (isometrically isomorphic to) `1 .
5.2.4 Definition
Let M be a subset of a normed linear space X . The annihilator of M , denoted by M ?, is the set
M ? D fx 2 X j x .y/ D 0 for all y 2 M g:
It is easy to show that M ? is a closed linear subspace of X .
5.2.3 Theorem
Let M be a linear subspace of a normed linear space X . Then
X =M ? M :
Proof. Define W X =M ? ! M by
.x C M ?/.m/ D x .m/ for all x 2 X and all m 2 M:
We show that is well-defined. Let x ; y 2 X such that x C M ? D y C M ? . Then x y 2 M ?

and so x .m/ D y .m/ for all m 2 M . Thus .x C M ?/ D .y C M ?/; i.e., is well-defined.
Clearly, .x C M ? / is a linear functional on M .
We show that is linear. Let x ; y 2 X and 2 F. Then, for all m 2 M ,
..x C M ? / C .y C M ? //.m/ D .x C y C M ? /.m/ D .x C y /.m/

D x .m/ C y .m/
D .x C M ? /.m/ C .y C M ?/.m/

D .x C M ? / C .y C M ? / .m/:
Hence, ..x C M ? / C .y C M ?// D .x C M ? / C .y C M ? /.

We now show that is surjective. Let y 2 M . Then, by Theorem 5.1.4, there is an x 2 X such that
y .m/ D x .m/ for all m 2 M and ky k D kx k. Hence, for all m 2 M ,
.x C M ? /.m/ D x .m/ D y .m/:
87
Thus .x C M ? / D y . Furthermore,
kx C M ? k kx k D ky k D k.x C M ?/k:
But for any y 2 M ? , x C M ? D .x C y / C M ?. Hence, for all m 2 M ,
j.x C M ? /.m/j D j.x C y /.m/j kx C y kkmk:
That is, .x C M ? / is a bounded linear functional on M and k.x C M ? /k kx C y k for all

y 2 M ? . Thus
k.x C M ?/k inf kx C y k D kx C M ? k:
y 2M ?
It now follows that k.x C M ?/k D kx C M ? k.
5.3 Bidual of a normed linear space and Reflexivity
Let .X; k k/ be a normed linear space over F and x 2 X . Define a functional x W X ! F by
x .x / D x .x/ for all x 2 X :
It is easy to verify that x is linear and for each x 2 X ,
jx .x /j D jx .x/j kx kkxk:
That is, x is bounded and kx k kxk. By Corollary 5.2.3,
kxk D supfjx .x/j j x 2 X ; kx k D 1g D supfjx .x/j j x 2 X ; kx k D 1g D kx k:
This shows that x is a bounded linear functional on X , i.e., x 2 .X / D X and kx k D kxk. The
space X is called the second dual space or bidual space of X . It now follows that we can define a map
JX W X ! X by
JX x D x ; for x 2 X; that is, .JX x/.x / D x .x/ for x 2 X and x 2 X :
It is easy to show that JX is linear and kxk D kx k D kJX xk. That is, JX is a linear isometry of X into
its bidual X . The map JX as defined above is called the canonical or natural embedding of X into its
bidual X . This shows that we can identify X with the subspace JX X D fJX x j x 2 X g of X .
5.3.1 Definition
Let .X; k k/ be a normed linear space over F. Then X is said to be reflexive if the canonical embedding
JX W X ! X of X into its bidual X is surjective. In this case X X .
If X is reflexive, we customarily write X D X . The equality simply means that X is isometrically

isomorphic to X . Reflexivity of X basically means that each bounded linear functional on X is an
evaluation functional. Since dual spaces are complete, a reflexive normed linear space is necessarily a
Banach space. It is therefore appropriate to speak of a reflexive Banach space rather than a reflexive normed
linear space.
5.3.1 Theorem
(1) Every finite-dimensional normed linear space is reflexive.
(2) A closed linear subspace of a reflexive space is reflexive.
88
Proof. (1). If dimX < 1, then Proposition 4.2.4 implies that dimX D dimX D dimX . Since JX X is
isometrically isomorphic to X , dim.JX X / D dimX D dimX . Since JX X is a subspace of X , it must
equal to X .
(2). Let X be reflexive and M a closed linear subspace of X . Given y 2 M , it must be shown that
there exists y 2 M such that JM y.y / D y .y/ D y .y / for all y 2 M . Define a functional on
X by
.x / D y .x jM /; x 2 X :
Clearly, is linear and
j .x /j ky kkx jM k ky kkx k
so 2 X . By reflexivity of X , there exists y 2 X such that JX y D . That is, .x / D x .y/ for each
x 2 X . If y 62 M , then by Theorem 5.2.1, there exists an x0 2 X such that x0 .y/ 0 and x0 .m/ D 0
for all m 2 M . Then
0 x0 .y/ D .x0 / D y .x0 jM / D y .0/ D 0
which is absurd. Thus y 2 M and x .y/ D .x / D y .x jM /; x 2 X : By Theorem 5.1.4, every
y 2 M is of the form y D x jM for some x 2 X . Thus
.JM y/.y / D y .y/ D y .y /, y 2 M , and the proof is complete.
5.3.2 Theorem
A Banach space X is reflexive if and only if its dual X is reflexive.
Proof. Assume that X is reflexive. Let JX W X ! X and JX W X ! .X / D X be the

canonical embeddings of X and X respectively. We must show that JX is surjective. To that end, let
x 2 X D .X / and consider the following diagram:
JX x
X ! X ! F:
Define a functional x on X by x D x JX . It is obvious that x is linear since both x and JX are
linear. Also, for each x 2 X ,
jx .x/j D jx JX .x/j kx kkJX xk D kx kkxk:
i.e., x is bounded and kx k kx k. Hence x 2 X . We now show that JX .x / D x . Let
x 2 X be any element. Since JX is surjective, there is an x 2 X such that x D JX x. Hence
x .x / D x .JX x/ D x .x/ D JX x.x / D JX x .JX x/ D JX x .x /;
and therefore JX x D x . That is, JX is surjective.
Assume that X is reflexive. Then the canonical embedding JX W X ! X is surjective. If
JX X X , let x 2 X n JX X . Since JX X is a closed linear subspace of X , it follows from
Theorem 5.2.1 that there is a functional 2 X such that kk D 1, .x / D d.x ; JX X /, and
.JX x/ D 0 for all x 2 X . Since JX is surjective, there is an x 2 X such that JX x D . Hence, for
each x 2 X ,
0 D .JX x/ D JX x .JX x/ D .JX x/.x / D x .x/;
i.e., x .x/ D 0 for all x 2 X . This implies that x D 0. But then 0 D JX x D , a contradiction since
0. Hence JX X D X ; i.e., JX is surjective.
5.3.2 Exercise
Show that if X is a non-reflexive normed linear space, then the natural inclusions X X
X and X X X are all strict.
We showed earlier (Theorem 5.2.2) that if the dual space X of a normed linear space X is separable,
then X is also separable, but not conversely. However, if X is reflexive, then the converse holds.
89
5.3.3 Theorem
If X is a reflexive separable Banach space, then its dual X is also separable.
Proof. Since X is reflexive and separable, its bidual X D JX X is also separable. Hence, by Theo-
rem 5.2.2, X is separable.
5.3.3 Examples
(1) For 1 < p < 1, the sequence space `p is reflexive.
(2) The spaces c0 ; c; `1 ; and `1 are non-reflexive.
(3) Every Hilbert space H is reflexive.
5.4 The Adjoint Operator

5.4.1 Definition
Let X and Y be normed linear spaces and T 2 B.X; Y /. The Banach space adjoint (or simply adjoint)
of T , denoted by T , is the operator T W Y ! X defined by
.T y /.x/ D y .T x/ for all y 2 Y and all x 2 X:
The following diagram helps make sense of the above definition.

T
X ! Y
T
X Y :
It is straightforward to show that for any y 2 Y , T y is a linear functional on X . Furthermore, for

any y 2 Y and x 2 X
jT y .x/j D jy .T x/j ky kkT kkxk;
i.e., T y is a bounded linear functional on X and kT y k kT kky k.
5.4.2 Example
Let X D `1 D Y and define T W `1 ! `1 by
T x D T .x1 ; x2 ; x3 ; : : : / D .0; x1; x2 ; x3 ; : : : /; where x D .xn / 2 `1 ;
the right-shift operator. Then the adjoint of T is T W `1 ! `1 is given by
T y D T .y1 ; y2 ; y3 ; : : : / D .y2 ; y3 ; : : : /; where y D .yn / 2 `1 ;
the left-shift operator.
5.4.1 Theorem
Let X and Y be normed linear spaces over F and let T 2 B.X; Y /.
(a) T is a bounded linear operator on Y .
(b) The map W B.X; Y / ! B.Y ; X / defined by T D T is an isometric isomorphism of
B.X; Y / into B.Y ; X /.
90
Proof. (a) Let y1 ; y2 2 Y and 2 F. Then, for all x 2 X
T .y1 C y2 /.x/ D .y1 C y2 /.T x/ D y1 .T x/ C y2 .T x/

D T y1 .x/ C T y2 .x/ D .T y1 C T y2 /.x/:
Hence, T .y1 C y2 / D T y1 C T y2 .

Furthermore, as shown above, kT y k kT kky k. Hence, T 2 B.Y ; X / and kT k kT k.
(b) We show that kT k D kT k, whence kT k D kT k. Indeed,
!
kT k D sup kT xk D sup sup jy .T x/j (by Corollary 5.2.3)
kxkD1 kxkD1 ky kD1
!
D sup sup jy .T x/j D sup kT y k

ky kD1 kxkD1 ky kD1

D kT k:
5.5 Weak Topologies
We have made the point that a norm on a linear space X induces a metric. A metric, in turn, induces
a topology on X called the metric topology. It now follows that a norm on a linear space X induces a
topology which we shall refer to as the norm topology on X . In this section we define other topologies on a
linear space X that are weaker than the norm topology. We also investigate some of the properties of these
weak topologies.
5.5.1 Definition
Let .X; k k/ be a normed linear space and F X . The weak topology on X induced by the family F ,
denoted by .X; F /, is the weakest topology on X with respect to which each x 2 F is continuous.
5.5.2 Remark
The weak topology on X induced by the dual space X is simply referred to as the weak topology on X
and is denoted by .X; X /.
What do the basic open sets for the weak topology .X; X / look like?
Unless otherwise indicated, we shall denote by ; 1 ; 2 : : : finite subsets of X .
Let x0 2 X; and > 0 be given. Consider all sets of the form
V .x0 I I / WD fx 2 X j jx .x/ x .x0 /j < ; x 2 g
\
D fx 2 X j jx .x/ x .x0 /j < g:
x 2
5.5.3 Proposition
[1] x0 2 V .x0 I I /.
[2] Given V .x0 I 1 I 1 / and V .x0 I 2I 2 /, we have
V .x0 I 1 [ 2 I minf1 ; 2 g/ V .x0 I 1 I 1 / \ V .x0 I 2 I 2 /:
[3] If x 2 V .x0 I I /, then there is a > 0 such that V .xI I / V .x0 I I /.
Proof. (1) It is obvious that x0 2 V .x0 I I /.
91
(2) Let x 2 V .x0 I 1 [ 2 I minf1 ; 2g/. Then for each x 2 1 ,
jx .x/ x .x0 /j < minf1 ; 2 g 1 :
Hence x 2 V .x0 I 1 I 1 /. Similarly, x 2 V .x0 I 2 I 2 /. It now follows that

x 2 V .x0 I 1 I 1 / \ V .x0 I 2 I 2 / and, consequently
V .x0 I 1 [ 2 I minf1 ; 2 g/ V .x0 I 1 I 1 / \ V .x0 I 2 I 2 /:
(3) Let x 2 V .x0 I I / and D maxfjx .x/ x .x0 /j j x 2 g. Then 0 < . Choose such
that 0 < < . Then, for any y 2 V .xI I / and any x 2 , we have
jx .y/ x .x0 /j jx .y/ x .x/j C jx .x/ x .x0 /j < C < :
Recall that a collection B of subsets of a set X is a base for a topology on X if and only if
S
(i) X D fB j B 2 Bg; i.e., each x 2 X belongs to some B 2 B, and
(ii) if x 2 B1 \ B2 for some B1 and B2 in B, then there is a B3 2 B such that x 2 B3 B1 \ B2 .
5.5.1 Theorem
Let B D fV .xI I / j x 2 X; .finite/ X ; > 0g. Then B is a base for a Hausdorff topology on X .
Proof. (i) It is clear that x 2 V .xI I / for each x 2 X .

(ii) Let x 2 V .x1 I 1 I 1 / \ V .x2 I 2 I 2 /. Then x 2 V .x1 I 1 I 1 / and x 2 V .x2 I 2 I 2 /.
By Proposition 5.5.3 (3), there are 1 > 0 and 2 > 0 such that V .xI 1 I 1 / V .x1 I 1 I 1 / and
V .xI 2 I 2 / V .x2 I 2 I 2 /. By Proposition 5.5.3 (2),
V .xI 1 [ 2 I minf1; 2 g/ V .xI 1 I 1 / \ V .xI 2 I 2 / V .x1 I 1 I 1 / \ V .x2 I 2 I 2 /:
Hence, B is a base for a topology on X .

Finally, we show that the topology generated by B is Hausdorff. Let x and y be distinct elements of X .
By Corollary 5.2.2, there is an x 2 X such that x .x/ x .y/. Let 0 < < jx .x/ x .y/j. Then
V .xI x I 2 / and V .yI x I 2 / are disjoint neighbourhoods of x and y respectively.
It is easy to see that each x 2 X is continuous with respect to the topology generated by B. Indeed,
let x0 2 X , x 2 X and > 0. Since x is continuous with respect to the norm topology on X , there is a
norm neighbourhood U of x0 such jx .x/ x .x0 /j < for all x 2 U . It now follows V .x0 I x I / is a
neighbourhood of x0 in the topology generated by B and jx .x/ x .x0 /j < for all x 2 V .x0 I x I /.
One shows quite easily that the topology generated by
B D fV .xI I / j x 2 X; .finite/ X ; > 0g
is precisely .X; X /, the weak topology on X induced by X . Therefore, a set G is open in the topology
.X; X / if and only if for each x 2 G there is a finite set
D fx ; x1 ; x2 ; : : : ; xn g X and an > 0 such that V .xI I / G.
It now follows that a normed linear space X carries two natural topologies: the norm topology induced
by the norm on X and the weak topology induced by its dual space X . Topological concepts that are asso-
ciated with the weak topology are usually preceded by the word weak; for example, weak compactness,
weak closure, etc. Those topological concepts pertaining to the topology generated by the norm on X are
usually preceded by the word norm, e.g. norm-closure or by the word strong, e.g. strongly open set.
5.5.4 Lemma
Let fx ; x1 ; x2 ; : : : ; xn g X . Then
92
Tn
(1) x 2 linfx1 ; x2 ; : : : ; xn g if and only if iD1 ker.xi / ker.x /.
If fx1; x2 ; : : : ; xn g is a linearly independent set, then for any set of scalars fc1; c2 ; : : : ; cn g,
(2) T
n
iD1 fx 2 X j xi .x/ D ci g ;.
n
X
Proof. (1) If x 2 linfx1; x2 ; : : : ; xn g, then x D i xi for some scalars 1 ; 2 ; : : : ; n . Let
iD1
n
\ n
X
x2 ker.xi /. Then xi .x/ D 0 for each i D 1; 2; : : : ; n. Hence, i xi .x/ D 0 and consequently,
iD1 iD1
n
\
x .x/ D 0; i.e., x 2 ker.x /. Therefore ker.xi / ker.x /.
iD1
n
\
Conversely, assume that ker.xi / ker.x /. We use induction on n. Let us first show that if
iD1
ker.x1 / D ker.x /, then x D x1 for some nonzero scalar . Let K D ker.x1 / and z 2 X n K. Then,
proceeding as in Theorem 5.2.1, each x 2 X is uniquely expressible as x D y C z, where y 2 K and
2 F. Hence, since x .y/ D 0 D x1 .y/,

x .z/ x .z/ x .z/
x .x/ D x .z/ D x1 .z/ D
x
1 .z/ D x1 .x/ D x1.x/;
x1 .z/ x1 .z/ x1 .z/
x .z/
where D .
x1 .z/
Assume that the result is true for n 1. For each i D 1; 2; : : : ; n, xi is not a linear combination of
\
the xj s for j D 1; 2; : : : ; n and i j . Hence, ker.xj / is not contained in ker.xi /. Therefore there
ji
\
is an xi 2 ker.xj / such that xi .xi / D 1. Let i D x .xi / for each i D 1; 2; : : : ; n. Let x 2 X and
ji
n
X
yDx xi .x/xi . Then, for each j D 1; 2; : : : ; n,
iD1
n
X
xj .y/ D xj .x/ xi .x/xj .xi / D xj .x/ xj .x/ D 0:
iD1
n
\
Thus, y 2 ker.xi /. By the assumption, y 2 ker.x /. Therefore
iD1
n
X n
X n
X
0 D x .y/ D x .x/ xi .x/x .xi / D x .x/ i xi .x/ x .x/ D i xi .x/;
iD1 iD1 iD1
n
X
whence x D i xi .
iD1 T
(2) Let Hi D fx 2 X j xi .x/ D ci g for each i D 1; 2; : : : ; n. We want to show that niD1 Hi ;.
The proof is by induction on n. If n D 1, then, since x1 0, it follows that H1 ;. Assume true for n D k
T Tk
and let H D kC1
iD1 Hi . By the linear independence of fx1 ; x2 ; : : : ; xkC1 g,

iD1 ker.xi / 6 ker.xkC1 /.
Tk
Tk
Hence, there is an x0 2 iD1 ker.xi / such that xkC1 .x0 / 6D 0. Take any x 2 iD1 ker.xi / and set

xkC1 .x/
y D x C x0, where D ckC1 . Then xi .y/ D xi .x/ D ci for each i D 1; 2; : : : ; k and
xkC1 .x0 /
93

xkC1 .y/ D ckC1 . That is, y 2 H .
5.5.2 Theorem
Let denote the norm topology on X . Then
(a) .X; X / .
(b) .X; X / D if and only if X is finite-dimensional. Thus, if X is infinite-dimensional, then the
weak topology .X; X / is strictly weaker than the norm topology.
Proof. (a) The topology .X; X / is the weakest topology on X making each x 2 X continuous. Hence,
.X; X / is weaker than the norm topology .
(b) Assume that .X; X / D and let x 2 X . Then, since x is continuous when X is equipped with
the norm topology and, by the hypothesis, it is continuous in the weak topology .X; X /, it is continous
at 0. Therefore there is a finite set D fx1 ; x2 ; : : : ; xn g X and an > 0 such that jx .x/j < 1 for
\n
all x 2 V .0I I /. Let z 2 ker.xi /. Then xi .z/ D 0 and so jxi .z/j < for each i D 1; 2; : : : ; n.
iD1
n
\ n
\ n
\
That is, z 2 V .0I I /. If x 2 ker.xi /, then mx 2 ker.xi / for each m 2 ZC since ker.xi / is
iD1 iD1 iD1
a linear subspace of X . It now follows that mx 2 V .0I I / for each m 2 ZC . This, in turn, implies that
1
1 > jx .mx/j D mjx .x/j jx .x/j < :
m
n
\
Since m is arbitrary, x .x/ D 0; i.e., x 2 ker.x /. Hence ker.xi / ker.x /. By Lemma 5.5.4,
iD1
n
X
x 2 X is expressible as x D i xi for some scalars 1 ; 2 ; : : : ; n . Hence X is spanned by
iD1
the set fx1 ; x2 ; : : : ; xn g. This shows that X is finite-dimensional. By Proposition 4.2.4, X is also
finite-dimensional.
Conversely, assume that X is finite-dimensional. Let fx1; x2 ; : : : ; xn g be a basis for X such that
kxk k D 1 for each k D 1; 2; : : : ; n. Let U X be open in the norm topology of X . We want to show that
U is open in the weak topology of X . Let x0 2 U . Then there is an r > 0 such that B.x0 ; r / U . For any
Xn
x 2 X, x D k xk . Define xi W X ! F by xi .x/ D i for each i D 1; 2; : : : ; n. Since the i s are
kD1
uniquely determined, xi is well-defined. One shows quite easily that xi 2 X for each i D 1; 2; : : : ; n.
r
Let D fx1 ; x2 ; : : : ; xn g and D . Then, for any x 2 V .x0 I I /, we have jxi .x/ xi .x0 /j <
n
for each i D 1; 2; : : : ; n. Hence, if x 2 V .x0 I I /, then
n
X X n

kx x0k D xk .x x0 /xk jxk.x x0 /j < n D r:

kD1 kD1
That is, x 2 B.x0 ; r / U . It now follows that for each x 2 U , there is a V .xI I / such that
V .xI I / U . Hence U is open in the weak topology .X; X /. Thus, .X; X / D .

The following result asserts that the weak topology and the norm topology yield exactly the same
continuous linear functionals. That is, the linear functionals on X that are continuous with respect to
94
the topology .X; X / are those that are in X . Therefore weakening the topology does not affect the dual
space of X .
5.5.3 Theorem
Let X be a normed linear space. Then the dual of X under the topology .X; X / is X ; i.e., .X; .X; X // D
X .
Proof. By definition of the topology .X; X /, it is clear X is a subset of the dual of X under the topology
.X; X /; i.e., X .X; .X; X // .
Let f 2 .X; .X; X // . Then, proceeding as in Theorem 5.5.2, there is a finite set
n
X
fx1 ; x2 ; : : : ; xn g X and scalars 1; 2 ; : : : ; n such that f D i xi . Therefore f 2 X .
iD1
5.5.4 Theorem
Let K be a convex subset of a normed linear space X . Then the closure of K relative to the weak topology
.X ;X /
.X; X / is the same as the norm-closure of K , i.e., K D K.
.X ;X / .X ;X /
Proof. Since K is closed and K K and K is the smallest closed set containing K, it
.X ;X /
follows that K K .
.X ;X /
It remains to show that K K. Let x0 2 X n K. Then, by Hahn-Banachs Theorem, there is
an x 2 X and real numbers c1 and c2 such that

Re x .x0 / c1 < c2 Re x .x/ for all x 2 K:
Consider V D V .x0 I x I c2 c1 / D fx 2 X j jx .x/ x .x0 /j < c2 c1 g: Then V is a weak neighbour-

.X ;X / .X ;X /
hood of x0 and V \ K D ;. Thus x0 62 K , and consequently K K.

Since the topology .X; X / is weaker than the norm topology, every weakly closed set in X is closed.
However, for convex sets we have the following.
5.5.5 Corollary
A convex subset K of a normed linear space X is closed if and only it is weakly closed.
We now turn our attention to the dual space X of a normed linear space X . X carries three natural
topologies: the norm topology, the weak topology .X ; X / induced by X and the weak* topology
.X ; X / induced by X .
Let JX be the canonical embedding of X into its bidual X . Then X JX X X . A typical basic
open set in the topology .X ; JX X / on X induced by JX X is
V .x I I / WD fy 2 X j j.JX x/.x / .JX x/.y /j < ; > 0; x 2 .finite/ X g

D fy 2 X j jx .x/ y .x/j < ; > 0; x 2 .finite/ X g:
It now follows that the weak* topology .X ; X / on X is precisely the weak topology on X induced by
JX X . That is, .X ; X / D .X ; JX X / the weak topology on X induced by elements of X acting as
continous linear functionals on X .
Let us observe, in passing, that X has a weak* topology .X ; X / induced by X . Since X
JX X X , the weak topology .X; X / on X turns out to be the relative topology on X induced by
.X ; X /.
5.5.5 Theorem
Let denote the norm topology on X . Then
95
(a) .X ; X / .X ; X / .

(b) .X ; X / D if and only if X is finite-dimensional.
(c) .X ; X / D .X ; X / if and only if X is reflexive. Thus, if X is non-reflexive, then the weak*
topology .X ; X / is strictly weaker than the weak topology .X ; X /.
Proof. (a) Since JX X X , it follows that .X ; X / D .X ; JX X / .X ; X /: The containment
.X ; X / follows from the fact that .X ; X / is the weakest topology on X making each
x 2 X continuous and each x 2 X is continuous with respect to .
(b) An argument similar to that used in Theorem 5.5.2(b) shows that .X ; X / D if and only
if X is finite-dimensional. But by Proposition 4.2.4, X is finite-dimensional if and only if X is finite-
dimensional.
(c) X is reflexive if and only if JX X X if and only if .X ; JX X / .X ; X / if and only if
.X ; X / .X ; X /.
5.5.6 Theorem
Let X be a normed linear space. Then the dual of X under the weak* topology .X ; X / is X ; i.e.,
.X ; .X ; X // D X .
Proof. Exercise.
Y
Observe that X FX D F and that the weak* topology .X ; X / on X is the relative topology
X Y

on X induced by the product topology on F.
X
5.5.7 Theorem
(Banach-Alaoglu-Bourbaki Theorem). Let X be a normed linear space over F. Then the closed unit ball
in X is weak* compact; i.e., the set
B D B.X / D fx 2 X j kx k 1g
is compact for the topology .X ; X /.
Proof. For each x 2 X , let Dx D f 2 F j jj kxkg. Then, for each x 2 X , Dx is a closed interval in R
or a closed disk in C according to whether
Q F D R or F D C. Equipped with the standard topology, Dx is
compact for each x 2 X . Let D D fDx j x 2 X g. By Tychonoffs Theorem, D is compact.
The points of D are just functions f on X such that f .x/ 2 Dx for each x 2 X . If x 2 B.X /, then
jx .x/j kx kkxk kxk for each x 2 X:

Q
Hence x .x/ 2 Dx for each x 2 B.X / and x 2 X . That is, B.X / fDx j x 2 X g. We observe
that the topology that D induces on B.X / is precisely the weak* topology on B.X /. It remains to show
that B.X / is a closed subset of D. To this end, let fx g be a net in B.X / and x ! x 2 D. Then
x .x/ ! x .x/ for all x 2 X , and for all x; y in X and ; in F,
x .x C y/ D lim x .x C y/ D lim x .x/ C x.y/ D x .x/ C x .x/:

Thus x is linear. Since

jx .x/j D lim jx .x/j kxk

for all x 2 X , x is bounded and kx k 1. That is, x 2 B.X /. Therefore B.X / is closed in D and

hence compact.
96
5.5.8 Theorem
(Helly). Let X be a normed linear space over Fand x 2 X . Then, for any finite-dimensional subspace
of X and any > 0, there is an x0 2 X such that
(i) .JX x0 /.x / D x .x / x .x0 / D x .x / for each x 2 , and
(ii) kx0k kx k C .
Proof. Let fx1 ; x2 ; : : : ; xn g be a basis for . Then (i) is equivalent to
(i) xi .x0 / D x .xi / for each i D 1; 2; : : : ; n.
n
\
Let Hi D fx 2 X j xi .x/ D x .xi /g for each i D 1; 2; : : : ; n and H D Hi . Then, by Lemma 5.27,
iD1
H ;. Choose any x0 2 H such that kx0k < d.0; H / C . Obviously, x0 satisfies (i), hence (i). To
complete the proof, it suffices
Tnto show that d.0; H / k x j k. Fix an h 2 H and set h0 D h x0 and

H0 D H x0. Then H0 D iD1 ker.xi / and d.0; H / D d. x0; H0 / D d.x0 ; H0 /. By the Hahn-Banach
Theorem and Lemma 5.5.4, it follows that
d.0; H / D maxfx .x0 / j x 2 H0? ; kx k 1g

n
Xn X

D maxf i xi .x0 / j i xi 1g

iD1 iD1

Xn X n

D maxf i x0 .xi / j i xi 1g

iD1 iD1
!
X n X n

D maxfx0 i x i j i xi 1g

iD1 iD1

maxfx0 .x / j x 2 ; kx k 1g D k x k:

5.5.9 Theorem
(Goldstine). Let X be a normed linear space and JX the canonical embedding of X into X . Let
B D fx 2 X j kxk 1g and B D fx 2 X j kx k 1g. Then JX B is dense in B relative to the
weak* topology .X ; X / on X . That is,
.X ;X /
JX B D B :
Proof. We must show that for each x 2 B , each finite subset D fx1 ; x2 ; : : : ; xn g X and each
> 0, there is an x 2 B such that JX x 2 V .x I I /; i.e.,
jJX x.xi / x .xi /j < for each i D 1; 2; : : : ; n:
Let x 2 B . If kx k < 1, then, with D 1 kx k, we have, by Theorem 5.5.8, that there is an

x 2 X such that .JX x/.xi / D x .xi / for each i D 1; 2; : : : ; n and kxk < kx k C D 1; i.e., x 2 B.
Hence, x 2 B and 0 D jJX x.xi / x .xi /j < for each i D 1; 2; : : : ; n.
If kx k D 1, let r D max kxik and y D 1 2r
x . Then ky k < 1, and so by the first
1in
part, there is an x 2 B such that .JX x/.xi / D y .xi / for each i D 1; 2; : : : ; n. Furthermore, for each
i D 1; 2; : : : ; n,

jJX x.xi / x .xi /j D jy .xi / x .xi /j <
2r 2
97
5.5.6 Corollary
Let X be a normed linear space over F and let JX be the canonical embedding of X into X . Then JX X
is dense in X relative to the weak* topology .X ; X / on X . That is,
.X ;X /
JX X D X :
Proof. Let x 2 X n f0g. Then
x .X ;X / .X ;X /

2 B D JX B JX X :
kx k
.X ;X / .X ;X /
Since JX X is a linear subspace of X , it now follows that x 2 JX X . Hence
.X ;X / .X ;X /
X JX X . Of course, since JX X X , we have that JX X X , and conse-
.X ;X /
quently JX X DX :
5.5.10 Theorem
Let X be a normed linear space over F and B D fx 2 X j kxk 1g. Then X is reflexive if and only if B
is weakly compact.
Proof. Assume that X is reflexive and let JX be the canonical embedding of X into X . Equip B (respec-
tively, B ) with the weak (respectively, weak*) topology and consider the map f W B ! B defined by
f .JX x/ D x. Now, B is weak* compact by Banach-Alaoglu-Bourbaki Theorem and f .B / D B. To
prove weak compactness of B, it suffices to show that f is continuous. To that end, let .JX x / be a net in
B that converges to JX x in the topology .X ; X / on X . Then, for each x 2 X , we have that

x f .JX x / D x .x / D .JX x /.x / ! JX x.x / D x .x/ D x .f .JX x//:
Thus, f .JX x / ! f .JX x/ in the weak topology on B.

Conversely, assume that B is weakly compact. Equip B (respectively, B ) with the weak (respec-
tively, weak*) topology. It follows that JX is continuous. Hence, JX B is weak* compact in B . But
.X ;X /
JX B D B : Hence JX X D X and so X is reflexive.
98
Chapter 6
Baires Category Theorem and its

Applications
6.1 Introduction
Recall that a subset S of a metric space .X; d / is dense in X if S D X ; i.e., for each x 2 X and each
> 0, there is an element y 2 S such that d.x; y/ < , or equivalently, S \ B.x; / ;.
6.1.1 Theorem
Let .X; d / be a complete metric space. If .Gn / is a sequence of nonempty, open and dense subsets of X
\
then G D Gn is dense in X .
n2N
Proof. Let x 2 X and > 0. Since G1 is dense in X , there is a point x1 in the open set G1 \ B.x; /.

Let r1 be a number such that 0 < r1 < and
2
B.x1 ; r1 / G1 \ B.x; /:
Since G2 is dense in X , there is a point x2 in the open set G2 \ B.x1 ; r1 /. Let r2 be a number such that

0 < r2 < 2 and
2
B.x2 ; r2 / G2 \ B.x1 ; r1 /:
Since G3 is dense in X , there is a point x3 in the open set G3 \ B.x2 ; r2 /. Let r3 be a number such that

0 < r3 < 3 and
2
B.x3 ; r3 / G3 \ B.x2 ; r2 /:
Continuing in this fashion, we obtain a sequence .xn / in X and a sequence .rn / of radii such that for each
n D 1; 2; 3; : : :,

0 < rn < ; B.xnC1 ; rnC1 / GnC1 \ B.xn ; rn / and B.x1 ; r1 / G1 \ B.x; /:
2n
It is clear that
B.xnC1 ; rnC1 / B.xn ; rn / B.xn 1 ; rn 1 / B.x1 ; r1 / B.x; /:
Let N 2 N. If k > N and ` > N , then both xk and x` lie in B.xN ; rN /. By the triangle inequality
2
d.xk ; x`/ d.xk ; xN / C d.xN ; x` / < 2rN < D N :
2N 2 1
99
Hence, .xn / is a Cauchy sequence in X . Since X is complete, there is a y 2 X such that xn ! y as

n ! 1. Since xk lies in the closed
\ set B.xn ; rn / if k > n, it follows that y lies in each B.xn ; rn /. Hence
y lies in each Gn . That is, G D Gn 6D ;. It is also clear that y 2 B.x; /.
n2N
6.1.1 Definition
A subset S of metric space .X; d / is said to be nowhere dense in X if the set X n S is dense in X ; i.e.,
X n S D X.
6.1.2 Proposition
A subset S of a metric space .X; d / is nowhere dense in X if and only if the closure S of S contains no
interior points.
Proof. Assume that S is nowhere dense in X and that .S / ;. Then there is an > 0 and an x 2 S such
that B.x; / S . But then X n S X n B.x; /. Since X n B.x; / is closed, X n B.x; / D X n B.x; /.
Therefore
X n S X n B.x; / X;
where the second containment is proper. This is a contradiction. Hence, .S / D ;.
Conversely, assume that .S / D ;. Then, for each x 2 S and each > 0,
B.x; / \ X n S ;:
This means that each x 2 S is a limit point of the set X n S. That is, S X n S. Thus,
X D S [ .X n S / X n S [ X n S D X n S X:
Hence X D X n S and so S is nowhere dense in X .
6.1.3 Example
Each finite subset of R is nowhere dense in R.
6.1.4 Definition
A subset S of a metric space .X; d / is said to be
(a) of first category or meagre in X if S can be written as a countable union of sets which are nowhere
dense in X . Such sets are also called thin.
(b) of second category or nonmeagre in X if it is not of first category in X . Such sets are also called
fat or thick.
It is clear that a subset of a set of first category is itself a set of first category. Also, a countable union of
sets of first category is again a set of first category.
6.1.5 Example
The set Q of rationals is of first category in R.
6.1.2 Theorem
(Baires Category Theorem). A complete metric space .X; d / is of second category in itself.
[X is of first category. Then there is a sequence .Gn / of

Proof. Assume that [sets which are nowhere dense in
X such that X D Gn . Replacing each Gn by its closure, we get X D Gn . The sets Gn are closed and
n n
100
nowhere dense in X . It follows that the \

sets Un D X n Gn are open and dense in X . Since X is complete,
it follows, by Theorem 6.1.1, that U D Un is dense in X and therefore nonempty since X is nonempty.
[ n
However X D Gn implies that
n
\ \ [
; Un D .X n Gn / D X n Gn D ;;
n n n
which is absurd.
6.2 Uniform Boundedness Principle

We have made the point that if X and Y are normed linear spaces, then B.X; Y / is a normed linear space.
6.2.1 Definition
A subset F of B.X; Y / is said to be
(a) norm (or uniformly) bounded if
supfkT k j T 2 F g < 1:
(b) pointwise bounded on X if

supfkT xk j T 2 F g < 1
for each x 2 X .
Clearly, a norm bounded set is pointwise bounded on X . Uniform Boundedness Principle (or Banach-
Steinhaus Theorem) says that if X is a Banach space, then the converse also holds.
6.2.1 Theorem
(Uniform Boundedness Principle). Let X be a Banach space, Y a normed linear space and let F be
subset of B.X; Y / such that supfkT xk j T 2 F g < 1 for each x 2 X . Then supfkT k j T 2 F g < 1:
Proof. For each k 2 N, let

Ak D fx 2 X j kT xk k for all T 2 F g:
Since T is continuous, Ak is closed. Indeed, let x 2 Ak . Then there is a sequence .xn / Ak such that
lim xn D x. Since xn 2 Ak for each n, kT xn k k for all T 2 F . Hence
n!1
kT xk kT x T xn k C kT xn k kT kkxn xk C k ! k as n ! 1:
That is, kT xk k and consequently x 2 Ak .

[1
By the hypothesis, X D Ak . By Baires Category Theorem, there is an index k0 such that .Ak0 / ;.
kD1
That is, there is an x0 2 Ak0 and an > 0 such that B.x0 ; / Ak0 D Ak0 .

Let x 2 X n f0g and set z D x0 C x, where D . Then kz x0 k D kxk D < . Hence
2kxk 2
z 2 B.x0 ; / Ak0 and, consequently, kT zk k0 for all T 2 F . It now follows that
1 1 2k0 4k0
kT xk D kT z T x0 k .kT zk C kT x0 k/ D kxk:

4k0
Hence kT k for all T 2 F .

101
It is essential that X be complete in Theorem 6.2.1. Consider the subset `0 `1 of finitely nonzero
sequences in `1 . The set `0 is dense but not closed in `1 . For each n 2 N, let Tn x D nxn, where
x D .xn / 2 `0 . For each x 2 `0 , Tn x D 0 for sufficiently large n. Clearly, .Tn / is pointwise bounded on
`0 . On the other hand, for .en / 2 `0 , ken k D 1 and kTn k Tn en D n for all n 2 N. Thus .Tn / is not norm
bounded.
6.2.2 Corollary
Let S be a subset of a normed linear space .X; k k/ such that the set fx .x/ j x 2 Sg is bounded for each
x 2 X . Then the set S is bounded.
Proof. Let JX be the canonical embedding of X into X . By the hypothesis, the set fJX x.x / j x 2 Sg
is bounded for each x 2 X . Since X is a Banach space, it follows from the Uniform Boundedness
Principle that the set fJX x j x 2 Sg is bounded. Since kJX xk D kxk, the set S is also bounded.
Let X and Y be normed linear spaces. We remarked earlier that the strong operator limit T of a
sequence .Tn / B.X; Y / need not be bounded. However, if X is complete, then T is also bounded. This
is a consequence of the Uniform Boundedness Principle.
6.2.3 Corollary
Let .Tn / be a sequence of bounded linear operators from a Banach space X into a normed linear space Y .
If T is the strong operator limit of the sequence .Tn /, then T 2 B.X; Y /.
Proof. The proof of linearity of T is straightforward.

We show that T is bounded. Since for each x 2 X , Tn x ! T x as n ! 1, the sequence .Tn x/ is
bounded for each x 2 X . By the Uniform Boundedness Principle, we have that the sequence .kTn k/ is
bounded. That is, there is a constant M > 0 such that kTn k M for all n 2 N. Therefore
kTn xk kTn kkxk M kxk for all n 2 N:
By continuity of the norm,
kT xk kT x Tn xk C kTnxk kT x Tn xk C M kxk ! M kxk as n ! 1:
Hence, kT xk M kxk for each x 2 X , i.e., T 2 B.X; Y /.
6.3 The Open Mapping Theorem

6.3.1 Definition
Let X and Y be normed linear spaces over the same field F and let T W X ! Y . Then we say that T is an
open mapping if T U is open in Y whenever U is open in X .
6.3.2 Lemma
Let X and Y be Banach spaces over the field F and let T be a bounded linear operator from X onto Y .
Then there is a constant r > 0 such that
BY .0; 2r / WD fy 2 Y j kyk < 2r g TBX .0; 1/:

1
[
Proof. It is easy to see that X D nBX .0; 1/. Indeed, if x 2 X , then there is an n 2 N such that
nD1
kxk < n. Hence, x 2 nBX .0; 1/. Since T is surjective,
1
! 1 1
[ [ [
Y D TX D T nBX .0; 1/ D nTBX .0; 1/ D nTBX .0; 1/:
nD1 nD1 nD1
102
By Baires Category Theorem, there is a positive integer n0 such that .n0 TBX .0; 1// ;. This implies
that .TBX .0; 1// ;. Hence, there is a constant r > 0 and an element y0 2 Y such that BY .y0 ; 4r /
TBX .0; 1/. Since y0 2 TBX .0; 1/, it follows, by symmetry, that y0 2 TBX .0; 1/. Therefore
BY .0; 4r / D BY .y0 ; 4r / y0 TBX .0; 1/ C TBX .0; 1/:
Since TBX .0; 1/ is a convex set, TBX .0; 1/ C TBX .0; 1/ D 2TBX .0; 1/. Hence, BY .0; 4r / 2TBX .0; 1/
and, consequently, BY .0; 2r / TBX .0; 1/.
6.3.3 Lemma
Let X and Y be Banach spaces over the field F and let T be a bounded linear operator from X onto Y .
Then there is a constant r > 0 such that
BY .0; r / WD fy 2 Y j kyk < r g TBX .0; 1/:
Proof. By Lemma 6.3.2, there is a constant r > 0 such that BY .0; 2r / TBX .0; 1/. Let y 2 BY .0; r /,
i.e., y 2 Y and kyk < r . Then, with D r2 , there is an element z1 2 X such that
1 r
kz1 k < and ky T z1 k < :
2 2
r
Since y T z1 2 Y and ky T z1 k < 2 < r , it follows that y T z1 2 BY .0; r /. Therefore there is an
element z2 2 X such that
1 r
kz2 k < and k.y T z1 / T z2 k < :
22 22
1
In general, having chosen elements zk 2 X , 1 k n, such that kzk k < and
2k
r
ky .T z1 C T z2 C C T zn /k < ;
2n
1
pick znC1 2 X such that kznC1 k < and
2nC1
r
ky T .z1 C z2 C C zn C znC1 /k D ky .T z1 C T z2 C C T zn C T znC1 /k <
2nC1
1
X
Claim: The series zk converges to a point x 2 BX .0; 1/ and T x D y.
kD1
1
X
Proof of Claim: Since X is complete, it suffices to show that kzk k < 1. But this is obviously true
kD1
since
1
X 1
X 1
kzk k < D 1:
2k
kD1 kD1
1
X
Hence, the series zk converges to some x 2 X with kxk < 1, i.e., x 2 BX .0; 1/. Since
kD1
!
n
X r

lim y T zk D lim n D 0;
n!1 n!1 2
kD1
continuity of T implies that !

n
X
T x D lim T zk D y:
n!1
kD1
That is, T x D y.
103
6.3.1 Theorem
(Open Mapping Theorem). Let X and Y be Banach spaces and suppose that T 2 B.X; Y /. If T maps
X onto Y , then T is an open mapping.
Proof. Let U be an open set in X . We need to show that T U is open in Y . Let y 2 T U . Since
T is surjective, there is an x 2 U such that T x D y. Since U is open, there is an > 0 such that
BX .x; / D x C BX .0; / U . But then y C TBX .0; / T U . By Lemma 6.3.3, there is a constant
r > 0 such that BY .0; r / TBX .0; 1/: Hence BY .0; r / TBX .0; /. Therefore
B.y; r / D y C BY .0; r / y C TBX .0; / T U:
Hence T U is open in Y .
6.3.4 Corollary
1
(Banachs Theorem). Let X and Y be Banach spaces and assume T 2 B.X; Y / is bijective. Then T is
a bounded linear operator from Y onto X , i.e., T 1 2 B.Y; X /.
Proof. We have shown that T 1 is linear. It remains to show that T 1 is bounded. By Theorem 4.1.4, it
suffices to show that T 1 is continuous on Y . To that end, let U be an open set in X . By Theorem 6.3.1,
.T 1 / 1 .U / D T U is open in Y . Hence T 1 is continuous on Y .
6.4 Closed Graph Theorem

6.4.1 Definition
Let X and Y be linear spaces over a field F and T W X ! Y . The graph of T, denoted by G.T /, is the
subset of X Y given by
G.T / D f.x; T x/ j x 2 X g:
Since T is linear, G.T / is a linear subspace of X Y . Let k kX and k kY be norms on X and Y

respectively. Then, for x 2 X and y 2 Y , k.x; y/k WD kxkX C kykY defines a norm on X Y . If X and
Y are Banach spaces, then so is X Y .
6.4.2 Definition
Let X and Y be normed linear spaces over F. A linear operator T W X ! Y is closed if its graph G.T / is
a closed linear subspace of X Y .
6.4.1 Theorem
(Closed Graph Theorem). Let X and Y be Banach spaces and T W X ! Y a closed linear operator. Then
T is bounded.
Proof. Since X Y , with the norm defined above, is a Banach space, and by the hypothesis G.T / is closed,
it follows that G.T / is also a Banach space. Consider the map P W G.T / ! X given by P .x; T x/ D x.
Then P is linear and bijective. It is also bounded since
kP .x; T x/k D kxk kxk C kT xk D k.x; T x/k:
That is, P is bounded and kP k 1. By Banachs Theorem (Corollary 6.3.4), it follows that P 1 W X !
G.T / given by P 1 x D .x; T x/ for x 2 X , is also bounded. Hence k.x; T x/k D kP 1 xk kP 1 kkxk:
Therefore
k.x; T x/k D kxk C kT xk kP 1 kkxk kT xk kP 1kkxk:
1
That is, T is bounded and kT k kP k.
104

Functional Analysis Master PDF

Uploaded by

Functional Analysis Master PDF

Uploaded by

F UNCTIONAL A NALYSIS N OTES

Mr. Andrew Pinchuck

Department of Mathematics (Pure & Applied)

2 Normed Linear Spaces 13

4 Bounded Linear Operators and Functionals 62

5 The Hahn-Banach Theorem and its Consequences 81

6 Baires Category Theorem and its Applications 99

 G. B ACHMAN , L. NARICI , Functional Analysis, Academic Press, N.Y. 1966.

I have also found Wikipedia to be quite useful as a general reference.

sup jxi j < 1:

Then `p is a linear space over F.

jxi C yi jp  2 maxfjxi j; jyi jgp  2p maxfjxi jp ; jyi jp g  2p .jxi jp C jyi jp / ;

Thus, x C y 2 `p . Also, if x D .xn / 2 `p and 2 F, then

1.2 Subsets of a linear space

1.3 Subspaces and Convex Sets

Clearly, a subset M of a linear space X is a linear subspace if and only if M C M  M and M  M

only has the trivial solution 1 D 2 D D n D 0. Otherwise, the set fx1 ; x2 ; : : : ; xn g is

ej D .0; 0; : : : ; 1; 0; : : : ; 0/; where 1 occurs in the j -th position.

is a linear combination of elements of S, with nonnegative coefficients, such that

That is, x C .1 /y 2 C and C is convex. Clearly S  C . Hence co.S/  C .

1.4 Quotient Space

It is easy to verify that  defines an equivalence relation on X .

Clearly, if X D M , then X=M D f0g and so codim.X / D 0.

1.5 Direct Sums and Projections

The operator P W X ! X is called an algebraic projection if P is linear (i.e., P .x Cy/ D P x CPy

Conversely, if P W X ! X is an algebraic projection, then X D M N , where M D P .X / and

Proof. Linearity of P: Let x D m1 C n1 and y D m2 C n2, where m1 ; m2 2 M and n1 ; n2 2 N . For

Conversely, let x 2 X and set m D P x and n D .I P /x. Then x D m C n, where m 2 M and

1.6 The Holder and Minkowski Inequalities

If p 6D 2, then consider the function f W 0; 1/ ! R given by

for each 2 0; 1/.

jxk j jyk j 1 jxk jp 1 jyk jq

[3] Prove Proposition 1.3.4

Normed Linear Spaces

kx C yk2 D jx C yj2 D .x C y/.x C y/ D .x C y/.x C y/ D xx C yx C xy C yy

Taking the positive square root both sides yields N4.

N2. For any x 2 X ,

N3. For any x 2 X and any  2 F,

N4. For any x; y 2 X ,

Since this is true for all t 2 a; b, we have that

kx C yk1 D sup jx.t/ C y.t/j  kxk1 C kyk1 :

[4] Let X D Ca; b. For each x 2 X , define

kxk1 D sup jx.t/j

Then .X; k k1 / and .X; k k2 / are normed linear spaces.

[5] Let X D `p ; 1  p < 1. For each x D .xi /1

Then .X; k kp / is a normed linear space.

[6] Let X D `1 ; c or c0 . For each x D .xi /1

kxk D kxk1 D sup jxi j:

Then X is a normed linear space.

where A is the conjugate transpose of the matrix A.

B.a; r / D fx 2 X j kx ak < r g .Open ball with centre a and radius r /I

k.x; y/k1 D 1 k.x; y/k2 < 1 k.x; y/k1  1

kxk  kxk0  kxk; for all x 2 X:

Equivalence of k k1 and k k1 : Let x D .x1 ; x2 ; : : : ; xn / 2 X . For each k D 1; 2; : : : ; n,

Hence, kxk1  kxk1  nkxk1.

We now show that k k2 is equivalent to k k1 . Let x D .x1 ; x2 ; : : : ; xn / 2 X . For each

Also, for each k D 1; 2; : : : ; n,

Open and Closed Sets

It is easy to show that S is closed if and only if S D S.

2.2 Quotient Norm and Quotient Map

If y 2 x, then y x 2 M and hence y D x C m for some m 2 M . Hence

N3. For any x; y 2 X and  2 F n f0g,

N4. Let x; y 2 X . Then

2.3 Completeness of Normed Linear Spaces

kxn xk <  for all n  N:

If this is the case, we shall write

Convergence in the norm is called norm convergence or strong convergence.

Proof. (1). For any a 2 A,

kxnk  maxf kx1k; kx2k; kx3k; : : : ; kxn1 1 k; 1 C kxn1 kg

The classical sequence space `p is complete.

For each fixed index i , we have

jxn .i / xm .i /j <  for all n; m  N:

That is, for each fixed index i , .xn .i //1

Now letting k ! 1, then for all n  N ,

which means that xn x 2 `p . Since xn 2 `p , we have that x D .x xn / C xn 2 `p . It also

This sequence .xi /1

G. B ACHMAN , L. NARICI , Functional Analysis, Academic Press, N.Y. 1966.

jxi C yi jp 2 maxfjxi j; jyi jgp 2p maxfjxi jp ; jyi jp g 2p .jxi jp C jyi jp / ;

Clearly, a subset M of a linear space X is a linear subspace if and only if M C M M and M M

That is, x C .1 /y 2 C and C is convex. Clearly S C . Hence co.S/ C .

It is easy to verify that defines an equivalence relation on X .

N3. For any x 2 X and any 2 F,

kx C yk1 D sup jx.t/ C y.t/j kxk1 C kyk1 :

[5] Let X D `p ; 1 p < 1. For each x D .xi /1

k.x; y/k1 D 1 k.x; y/k2 < 1 k.x; y/k1 1

kxk kxk0 kxk; for all x 2 X:

Hence, kxk1 kxk1 nkxk1.

N3. For any x; y 2 X and 2 F n f0g,

kxn xk < for all n N:

kxnk maxf kx1k; kx2k; kx3k; : : : ; kxn1 1 k; 1 C kxn1 kg

jxn .i / xm .i /j < for all n; m N:

Now letting k ! 1, then for all n N ,

kx1 x3k and kx2 x3 k :

f .1 ; 2; : : : ; n / D infff .1 ; 2 ; : : : ; n / j .1 ; 2 ; : : : ; n / 2 Ag:

Let m D f .1 ; 2 ; : : : ; n /. Since f 0, it follows that m 0. If m D 0, then

Since the set fx1 ; x2 ; : : : ; xn g is linearly independent, j D 0 for all j D 1; 2; : : : ; n. This is a

kzk z` k < for all k; ` > N:

It now follows that jkj j M m

ky zk > 1 for all y 2 M:

d kx mk < d C d D d.1 C /:

Since Q is dense in R, for each i 2 R, we can find a i 2 Q such that

hA; Bi D .B A/; where B denotes conjugate transpose of matrix B:

0 hx y; x yi D hx; xi hy; xi hx; yi C hy; yi

jhx; yij D jhy; yij D jjjhy; yij D jjhy; yi

Proof. Let x; y 2 X and 2 F. Then

jhx; yij kxkkyk:

Thus, hx; yi D hx; yi for all 2 R and all x; y 2 X .

The set M X is called orthogonal if it consists of non-zero pairwise orthogonal elements.