A FIRST COURSE IN

FUNCTIONAL
ANALYSIS
A FIRST COURSE IN
FUNCTIONAL
ANALYSIS

ORR MOSHE SHALIT

Technion - Israel Institute of Technology
Haifa, Israel

Boca Raton London New York

CRC Press is an imprint of the

Taylor & Francis Group, an informa business
A CHAPMAN & HALL BOOK
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2017 by Taylor & Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

ISBN-13: 978-1-4987-7151-0 (hbk)

ISBN-13: 978-0-367-65813-7 (pbk)

This book contains information obtained from authentic and highly regarded sources. Reasonable efforts
have been made to publish reliable data and information, but the author and publisher cannot assume
responsibility for the validity of all materials or the consequences of their use. The authors and publishers
have attempted to trace the copyright holders of all material reproduced in this publication and apologize to
copyright holders if permission to publish in this form has not been obtained. If any copyright material has
not been acknowledged please write and let us know so we may rectify in any future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmit-
ted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented,
including photocopying, microfilming, and recording, or in any information storage or retrieval system,
without written permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.copyright.
com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood
Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and
registration for a variety of users. For organizations that have been granted a photocopy license by the CCC,
a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used
only for identification and explanation without intent to infringe.

Library of Congress Cataloging‑in‑Publication Data

Names: Shalit, Orr Moshe.

Title: A first course in functional analysis / Orr Moshe Shalit.
Description: Boca Raton : CRC Press, [2016] | Includes bibliographical references and
index.
Identifiers: LCCN 2016045930| ISBN 9781498771610 (hardback : alk. paper) | ISBN
9781315367132 (ebook) | ISBN 9781498771627 (ebook) | ISBN 9781498771641 (ebook)
| ISBN 9781315319933 (ebook)
Subjects: LCSH: Functional analysis‑‑Textbooks.
Classification: LCC QA320 .S45927 2016 | DDC 515/.7‑‑dc23
LC record available at https://lccn.loc.gov/2016045930

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com

and the CRC Press Web site at

http://www.crcpress.com
To my mother and my father, Malka and Meir Shalit
Contents

Preface xi

1 Introduction and the Stone-Weierstrass theorem 1

1.1 Background and motivation . . . . . . . . . . . . . . . . . . . 1

1.2 The Weierstrass approximation theorem . . . . . . . . . . . . 3
1.3 The Stone-Weierstrass theorem . . . . . . . . . . . . . . . . . 5
1.4 The Stone-Weierstrass theorem over the complex numbers . 8
1.5 Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . 9
1.6 Additional exercises . . . . . . . . . . . . . . . . . . . . . . . 10

2 Hilbert spaces 13

2.1 Background and motivation . . . . . . . . . . . . . . . . . . . 13

2.2 The basic definitions . . . . . . . . . . . . . . . . . . . . . . . 14
2.3 Completion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4 The space of Lebesgue square integrable functions . . . . . . 21
2.5 Additional exercises . . . . . . . . . . . . . . . . . . . . . . . 25

3 Orthogonality, projections, and bases 29

3.1 Orthogonality . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.2 Orthogonal projection and orthogonal decomposition . . . . 30
3.3 Orthonormal bases . . . . . . . . . . . . . . . . . . . . . . . . 33
3.4 Dimension and isomorphism . . . . . . . . . . . . . . . . . . 38
3.5 The Gram-Schmidt process . . . . . . . . . . . . . . . . . . . 40
3.6 Additional exercises . . . . . . . . . . . . . . . . . . . . . . . 40

4 Fourier series 45

4.1 Fourier series in L2 . . . . . . . . . . . . . . . . . . . . . . . 45

4.2 Pointwise convergence of Fourier series (Dirichlet’s theorem) 50
4.3 Fejér’s theorem . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.4 *Proof of Dirichlet’s theorem . . . . . . . . . . . . . . . . . . 55
4.5 Additional exercises . . . . . . . . . . . . . . . . . . . . . . . 55

vii
viii Contents

5 Bounded linear operators on Hilbert space 63

5.1 Bounded operators . . . . . . . . . . . . . . . . . . . . . . . . 63

5.2 Linear functionals and the Riesz representation theorem . . . 66
5.3 *The Dirichlet problem for the Laplace operator . . . . . . . 67
5.4 The adjoint of a bounded operator . . . . . . . . . . . . . . . 72
5.5 Special classes of operators . . . . . . . . . . . . . . . . . . . 74
5.6 Matrix representation of operators . . . . . . . . . . . . . . . 75
5.7 Additional exercises . . . . . . . . . . . . . . . . . . . . . . . 77

6 Hilbert function spaces 85

6.1 Hilbert function spaces . . . . . . . . . . . . . . . . . . . . . 85

6.2 *The Bergman space . . . . . . . . . . . . . . . . . . . . . . 89
6.3 *Additional topics in Hilbert function space theory . . . . . 93
6.4 Additional exercises . . . . . . . . . . . . . . . . . . . . . . . 102

7 Banach spaces 105

7.1 Banach spaces . . . . . . . . . . . . . . . . . . . . . . . . . . 105

7.2 Bounded operators . . . . . . . . . . . . . . . . . . . . . . . . 110
7.3 The dual space . . . . . . . . . . . . . . . . . . . . . . . . . . 112
7.4 *Topological vector spaces . . . . . . . . . . . . . . . . . . . 115
7.5 Additional exercises . . . . . . . . . . . . . . . . . . . . . . . 116

8 The algebra of bounded operators on a Banach space 121

8.1 The algebra of bounded operators . . . . . . . . . . . . . . . 121

8.2 An application to ergodic theory . . . . . . . . . . . . . . . . 122
8.3 Invertible operators and inverses . . . . . . . . . . . . . . . . 126
8.4 *Isomorphisms . . . . . . . . . . . . . . . . . . . . . . . . . . 131
8.5 Additional exercises . . . . . . . . . . . . . . . . . . . . . . . 133

9 Compact operators 139

9.1 Compact operators . . . . . . . . . . . . . . . . . . . . . . . 139

9.2 The spectrum of a compact operator . . . . . . . . . . . . . 141
9.3 Additional exercises . . . . . . . . . . . . . . . . . . . . . . . 144

10 Compact operators on Hilbert space 147

10.1 Finite rank operators on Hilbert space . . . . . . . . . . . . . 147

10.2 The spectral theorem for compact self-adjoint operators . . . 148
10.3 The spectral theorem for compact normal operators . . . . . 152
10.4 The functional calculus for compact normal operators . . . . 154
10.5 Additional exercises . . . . . . . . . . . . . . . . . . . . . . . 157
 Contents ix

11 Applications of the theory of compact operators 163

11.1 Integral equations . . . . . . . . . . . . . . . . . . . . . . . . 163

11.2 *Functional equations . . . . . . . . . . . . . . . . . . . . . . 173
11.3 Additional exercises . . . . . . . . . . . . . . . . . . . . . . . 176

12 The Fourier transform 179

12.1 The spaces Lp (R), p ∈ [1, ∞) . . . . . . . . . . . . . . . . . . 179

12.2 The Fourier transform on L1 (R) . . . . . . . . . . . . . . . . 184
12.3 The Fourier transform on L2 (R) . . . . . . . . . . . . . . . . 191
12.4 *Shannon’s sampling theorem . . . . . . . . . . . . . . . . . 195
12.5 *The multivariate Fourier transforms . . . . . . . . . . . . . 196
12.6 Additional exercises . . . . . . . . . . . . . . . . . . . . . . . 199

13 *The Hahn-Banach theorems 203

13.1 The Hahn-Banach theorems . . . . . . . . . . . . . . . . . . 203

13.2 The dual space, the double dual, and duality . . . . . . . . . 211
13.3 Quotient spaces . . . . . . . . . . . . . . . . . . . . . . . . . 213
13.4 Additional excercises . . . . . . . . . . . . . . . . . . . . . . 216

Appendix A Metric and topological spaces 219

A.1 Metric spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

A.2 Completeness . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
A.3 Topological spaces . . . . . . . . . . . . . . . . . . . . . . . . 227
A.4 The Arzelà-Ascoli theorem . . . . . . . . . . . . . . . . . . . 233

Bibliography 235

Index 237
Preface

In a nutshell
The purpose of this book is to serve as the accompanying text for a first
course in functional analysis, taken typically by second- and third-year under-
graduate students majoring in mathematics.
As I prepared for my first time teaching such a course, I found nothing
among the countless excellent textbooks in functional analysis available that
perfectly suited my needs. I ended up writing my own lecture notes, which
evolved into this book (an earlier version appeared on my blog [31]).
The main goals of the course this book is designed to serve are to in-
troduce the student to key notions in functional analysis (complete normed
spaces, bounded operators, compact operators), alongside significant applica-
tions, with a special emphasis on the Hilbert space setting. The emphasis on
Hilbert spaces allows for a rapid development of several topics: Fourier series
and the Fourier transform, as well as the spectral theorem for compact normal
operators on a Hilbert space. I did not try to give a comprehensive treatment
of the subject, the opposite is true. I did my best to arrange the material in
a coherent and effective way, leaving large portions of the theory for a later
course. The students who finish this course will be ready (and hopefully, ea-
ger) for further study in functional analysis and operator theory, and will have
at their disposal a set of tools and a state of mind that may come in handy
in any mathematical endeavor they embark on.
The text is written for a reader who is either an undergraduate student,
or the instructor in a particular kind of undergraduate course on functional
analysis. The background required from the undergraduate student taking this
course is minimal: basic linear algebra, calculus up to Riemann integration,
and some acquaintance with topological and metric spaces (in fact, the basics
of metric spaces will suffice; and all the required material in topology/metric
spaces is collected in the appendix).
Some “mathematical maturity” is also assumed. This means that the read-
ers are expected to be able to fill in some details here and there, not freak out
when bumping into a slight abuse of notation, and so forth.

More details on the contents and on some choices made

This book is tailor-made to accompany the course Introduction to Func-
tional Analysis given at the Technion — Israel Institute of Technology. The

xi
xii Preface

official syllabus of the course is roughly: basic notions of Hilbert spaces and
Banach spaces, bounded operators, Fourier series and the Fourier transform,
the Stone-Weierstrass theorem, the spectral theorem for compact normal op-
erators on a Hilbert space, and some applications. A key objective, not less
important than the particular theorems taught, is to convey some underlying
principles of modern analysis.
The design was influenced mainly by the official syllabus, but I also took
into account the relative place of the course within the curriculum. The back-
ground that I could assume (mentioned above) did not include courses on
Lebesgue integration or complex analysis. Another thing to keep in mind was
that besides this course, there was no other course in the mathematics un-
dergraduate curriculum giving a rigorous treatment of Fourier series or the
Fourier transform. I therefore had to give these topics a respectable place in
class. Finally, I also wanted to keep in mind that students who will continue
on to graduate studies in analysis will take the department’s graduate course
on functional analysis, in which the Hahn-Banach theorems and the conse-
quences of Baire’s theorem are treated thoroughly. This allowed me to omit
these classical topics with a clean conscience, and use my limited time for
a deeper study in the context of Hilbert spaces (weak convergence, inverse
mapping theorem, spectral theorem for compact normal operators), including
some significant applications (PDEs, Hilbert function spaces, Pick interpo-
lation, the mean ergodic theorem, integral equations, functional equations,
Fourier series and the Fourier transform).
An experienced and alert reader might have recognized the inherent pitfall
in the plan: how can one give a serious treatment of L2 spaces, and in particular
the theory of Fourier series and the Fourier transform, without using the
Lebesgue integral? This is a problem which many instructors of introductory
functional analysis face, and there are several solutions which can be adopted.
In some departments, the problem is eliminated altogether, either by mak-
ing a course on Lebesgue integration a prerequisite to a course on functional
analysis, or by keeping the introductory course on functional analysis free of
Lp spaces, with the main examples of Banach spaces being sequence spaces
or spaces of continuous functions. I personally do not like either of these easy
solutions. A more pragmatic solution is to use the Lebesgue integral as much
as is needed, and to compensate for the students’ background by either giving
a crash course on Lebesgue integration or by waving one’s hands where the
going gets tough.
I chose a different approach: hit the problem head on using the tools avail-
able in basic functional analysis. I define the space L2 [a, b] to be the completion
of the space of piecewise continuous functions on [a, b] equipped with the norm
Rb
kf k2 = ( a |f (t)|2 dt)1/2 , which is defined in terms of the familiar Riemann in-
tegral. We can then use the Hilbert space framework to derive analytic results,
such as convergence of Fourier series of elements in L2 [a, b], and in particular
we can get results on Fourier series for honest functions, such as L2 conver-
 Preface xiii

gence for piecewise continuous functions, or uniform convergence for periodic

and C 1 functions.
Working in this fashion may seem clumsy when one is already used to
working with the Lebesgue integral, but, for many applications to analysis
it suffices. Moreover, it shows some of the advantages of taking a functional
analytic point of view. I did not invent the approach of defining Lp spaces as
completions of certain space of nice functions, but I think that this book is
unique in the extent to which the author really adheres to this approach: once
the spaces are defined this way, we never look back, and everything is done
with no measure theory.
To illustrate, in Section 8.2.2 we prove the mean ergodic theorem. A mea-
sure preserving composition operator on L2 [0, 1] is defined first on the dense
subspace of continuous functions, and then extended by continuity to the com-
pletion. The mean ergodic theorem is proved by Hilbert space methods, as a
nice application of some basic operator theory. The statement (see Theorem
8.2.5) in itself is significant and interesting even for piecewise continuous func-
tions — one does not need to know the traditional definition of L2 in order
to appreciate it.
Needless to say, this approach was taken because of pedagogical con-
straints, and I encourage all my students to take a course on measure the-
ory if they are serious about mathematics, especially if they are interested in
functional analysis. The disadvantages of the approach we take to L2 spaces
are highlighted whenever we stare them in the face; for example, in Section
5.3, where we obtain the existence of weak solutions to PDEs in the plane,
but fall short of showing that weak solutions are (in some cases) solutions in
the classical sense.
The choice of topics and their order was also influenced by my personal
teaching philosophy. For example, Hilbert spaces and operators on them are
studied before Banach spaces and operators on them. The reasons for this are
(a) I wanted to get to significant applications to analysis quickly, and (b)
I do not think that there is a point in introducing greater generality before
one can prove significant results in that generality. This is surely not the most
efficient way to present the material, but there are plenty of other books giving
elegant and efficient presentations, and I had no intention — nor any hope —
of outdoing them.

How to use this book

A realistic plan for teaching this course in the format given at the Technion
(13 weeks, three hours of lectures and one hour of exercises every week) is
to use the material in this book, in the order it appears, from Chapter 1
up to Chapter 12, skipping Chapters 6 and 11. In such a course, there is
often time to include a section or two from Chapters 6 or 11, as additional
illustrative applications of the theory. Going through the chapters in the order
xiv Preface

they appear, skipping chapters or sections that are marked by an asterisk,

gives more or less the version of the course that I taught.
In an undergraduate program where there is a serious course on harmonic
analysis, one may prefer to skip most of the parts on Fourier analysis (except
L2 convergence of Fourier series), and use the rest of the book as a basis for
the course, either giving more time for the applications, or by teaching the
material in Chapter 13 on the Hahn-Banach theorems. I view the chapter on
the Hahn-Banach theorems as the first chapter in further studies in functional
analysis. In the course that I taught, this topic was given as supplementary
reading to highly motivated and capable students.
There are exercises spread throughout the text, which the students are
expected to work out. These exercises play an integral part in the development
of the material. Additional exercises appear at the end of every chapter. I
recommend for the student, as well as the teacher, to read the additional
exercises, because some of them contain interesting material that is good to
know (e.g., Gibbs phenomenon, von Neumann’s inequality, Hilbert-Schmidt
operators). The teaching assistant will also find among the exercises some
material better suited for tutorials (e.g., the solution of the heat equation, or
the diagonalization of the Fourier transform). There is no solutions manual,
but I invite any instructor who uses this book to teach a course, to contact
me if there is an exercise that they cannot solve. With time I may gradually
compile a collection of solutions to the most difficult problems.
Some of the questions are original, most of them are not. Having been a
student and a teacher in functional and harmonic analysis for several years,
I have already seen many similar problems appearing in many places, and
some problems are so natural to ask that it does not seem appropriate to
try to trace who deserves credit for “inventing” them. I only give reference to
questions that I deliberately “borrowed” in the process of preparing this book.
The same goes for the body of the material: most of it is standard, and I see
no need to cite every mathematician involved; however, if a certain reference
influenced my exposition, credit is given.
The appendix contains all the material from metric and topological spaces
that is used in this book. Every once in while a serious student — typically
majoring in physics or electrical engineering — comes and asks if he or she can
take this course without having taken a course on metric spaces. The answer
is: yes, if you work through the appendix, there should be no problem.

Additional reading and alternative texts

There are countless good introductory texts on functional analysis and
operator theory, and the bibliography contains a healthy sample. As a student
and later as a teacher of functional analysis, I especially enjoyed and was
influenced by the books by Gohberg and Goldberg [12], Devito [6], Kadison
and Ringrose [16], Douglas [8], Riesz and Sz.-Nagy [26], Rudin [27], Arveson
[3], Reed and Simon [24], and Lax [18]. These are all recommended, but only
 Preface xv

the first two are appropriate for a beginner. As a service to the reader, let me
mention three more recent elementary introductions to functional analysis,
by MacCluer [19], Hasse [13] and Eidelman, Milman and Tsolomitis [9]. Each
one of these looks like an excellent choice for a textbook to accompany a first
course.
I want to acknowledge that while working on the book I also
made extensive use of the Web (mostly Wikipedia, but also MathOver-
flow/StackExchange) as a handy reference, to make sure I got things right, e.g.,
verify that I am using commonly accepted terminology, find optimal phrasing
of a problem, etc.

Acknowledgments
This book could not have been written without the support, encourage-
ment and good advice of my beloved wife, Nohar. Together with Nohar, I feel
exceptionally lucky and thankful for our dear children: Anna, Tama, Gev, Em,
Shem, Asher and Sarah.
I owe thanks to many people for reading first drafts of these notes and
giving me feedback. Among them are Alon Gonen, Shlomi Gover, Ameer Kas-
sis, Amichai Lampert, Eliahu Levy, Daniel Markiewicz, Simeon Reich, Eli
Shamovich, Yotam Shapira, and Baruch Solel. I am sorry that I do not remem-
ber the names of all the students who pointed out a mistake here or there, but
I do wish to thank them all. Shlomi Gover and Guy Salomon also contributed
a number of exercises. A special thank you goes to Michael Cwikel, Benjamin
Passer, Daniel Reem and Guy Salomon, who have read large portions of the
notes, found mistakes, and gave me numerous and detailed suggestions on how
to improve the presentation.
I bet that after all the corrections by friends and students, there are still
some errors here and there. Dear reader: if you find a mistake, please let
me know about it! I will maintain a page on my personal website (currently
http://oshalit.net.technion.ac.il) in which I will collect corrections.
I am grateful to Sarfraz Khan from CRC Press for contacting me and
inviting me to write a book. I wish to thank Sarfraz, together with Michele
Dimont the project editor, for being so helpful and kind throughout. I also
owe many thanks to Samar Haddad the proofreader, whose meticulous work
greatly improved the text.
My love for the subject and my point of view on it were strongly shaped by
my teachers, and in particular by Boris Paneah (my Master’s thesis advisor)
and Baruch Solel (my Ph.D. thesis advisor). If this book is any good, then
these men deserve much credit.
My parents, Malka and Meir Shalit, have raised me to be a man of books.
This one, my first, is dedicated to them.

Haifa, 2017
xvi Preface

Notes for second printing, 2021

This revised printing incorporates corrections and improvements which
have occurred to me after publication or have been suggested by various read-
ers and friends. Special thanks go to Daniel Reem and Ross Pinsky, who pro-
vided extensive feedback. I also wish to thank Callum Fraser for his guidance
in the production of the second printing.

Rosh Pina, 2021

Chapter 1
Introduction and the
Stone-Weierstrass theorem

1.1 Background and motivation

Welcome to A First Course in Functional Analysis. What is functional
analysis? A short answer is that functional analysis is a language and frame-
work in which to formulate and study problems in analysis from an abstract
point of view. In this introductory chapter, we will try to give a more detailed
explanation of what is functional analysis. Along the way, we will prove the
first significant theorem in this book: the Stone-Weierstrass theorem.
Functional analysis was born around 1900, when the mathematical climate
was becoming suitable for an abstract and unified treatment of analytical
objects. Consider, for example, the following integral equation:
Z b
f (x) + k(t, x)f (t)dt = g(x). (1.1)
a

This equation was considered by I. Fredholm in 1903, and, like many important
mathematical problems of the time, it arose from mathematical physics. In
the equation above, the functions g ∈ CR ([a, b]) and k ∈ CR ([a, b] × [a, b]) are
given continuous functions, and f is an unknown function (here and below
CR (X) denotes the space of continuous, real-valued functions on a topological
space X). Fixing the function k, there are three basic questions one can ask
about such equations:
1. Solvability. Does there exist a continuous solution f to this equation
given g? For what g does a solution exist?
2. Uniqueness. Is the solution unique (when it exists)? Given a particular
solution to the equation, can we describe the space of all solutions, or
at least can we tell how “big” it is?
3. Method of solution. What is the solution? In other words, given g
can we write down a formula for the solution f , or at least describe a
method of obtaining f approximately?
Questions of a similar type are dealt with in a course in linear algebra,

1
2 A First Course in Functional Analysis

when considering a system of linear equations Ax = b, where x and b are

vectors in Rn , and A is an n × n matrix. There are many nontrivial things
to say regarding the solvability of the equation Ax = b which do not require
knowing the specific matrix A, for example: if the equation Ax = 0 has a
unique solution (namely, x = 0), then Ax = b has a unique solution for any
b ∈ Rn . In the same vein, one is interested not only in answering the above
three questions for the integral equation (1.1) given a particular k; it is also of
interest to understand the unifying characteristics of equations of this type.
One may develop an ad hoc theory of integral equations, but it is most
enlightening for us to put the above rather concrete equation in an abstract
and general framework. Historically as well as methodologically, this is the
starting point of functional analysis.
Let us try to force the above analytic problem into the framework of linear
algebra. First, we note that CR ([a, b]) is a real vector space. Next we notice,
following Fredholm, that the above equation can be written as (I + K)f = g,
where I : CR ([a, b]) → CR ([a, b]) is the identity operator and K : CR ([a, b]) →
CR ([a, b]) is the so-called integral operator
Z b
Kf (x) = k(t, x)f (t)dt.
a

Finally, we observe that the operator K, and hence I +K, is a linear operator.
Thus our problem is “just” that of solving a linear equation in the vector space
CR ([a, b]). However, these observations do not bring us much closer to being
able to answer the above questions; they only suggest that it might be fruitful
to study linear operators on infinite dimensional vector spaces.
Linear operators on infinite dimensional spaces turn out to be too large a
class to be treated in a meaningful manner. We will concentrate on the study
of continuous or compact operators acting on complete normed vector spaces.
In this setting significant results can be obtained; for example, in Chapter 9
we will see that the equation (I + K)f = g has a unique solution for any g,
if and only if the equation (I + K)f = 0 has a unique solution. Methods of
finding the solution will also be developed.
Note that, unlike in the case of a linear system of equations Ax = b, the
equation (1.1) makes sense within many different vector spaces. Would it be
easier to solve the problem if we assumed that all functions are differentiable?
The equation makes sense for integrable f — maybe we should consider I + K
as a linear operator on the larger space of all integrable functions, or maybe
on the space of square integrable functions? Do the basic properties of the
operator I + K change if we think of it as an operator on the bigger space?
We will see that considering the integral equation as an operator equation
in the space of square integrable functions on [a, b] does not change some
characteristic features of the problem, while on the other hand it facilitates
actually solving the problem.
The space of square integrable functions on an interval is an example of a
Hilbert space. Hilbert spaces are the infinite dimensional spaces that are closest
 Introduction and the Stone-Weierstrass theorem 3

to finite dimensional spaces, and they are the most tractable. What makes
them so tractable are the fact that they have an inner product, and the fact
that they are complete metric spaces. In this book, a special emphasis is put
on Hilbert spaces, and in this setting integral equations are best understood.
We will begin our study of Hilbert spaces in the next chapter.
The goal of this book is to present a set of tools, ideas and results that
can be used to understand linear operators on infinite dimensional spaces.
But before we can develop a theory of linear operators on infinite dimensional
spaces, we must study the spaces themselves. We will have to wait until much
later in the book, before we can prove significant results on the solvability
of integral equations; in Chapter 11 we will complete our treatment of this
subject. For now, we part from equation (1.1), and we take a closer look at
the space CR ([a, b]).

1.2 The Weierstrass approximation theorem

In linear algebra one learns that every finite dimensional vector space has a
basis. This fact is incredibly useful, both conceptually and also from a practical
point of view; just consider how easy it is for us to think about Rn having at
hand the standard basis e1 = (1, 0, . . . , 0), . . . , en = (0, . . . , 0, 1). In particular,
the existence of bases is crucial to our understanding of linear maps and of
the solvability of linear equations of the form Ax = b.
In analysis we encounter many vector spaces that are infinite dimensional.
For example, the space CR ([a, b]) is an infinite dimensional real vector space.
Our experience in linear algebra suggests that it would be useful to have some
kind of basis for studying this space, and in particular it should be helpful for
considering linear equations as above.
A standard application of Zorn’s lemma shows that every vector space V
has a basis — a set {ui }i∈I ⊂ V such that for all nonzero v ∈ V , there is
a unique choice of finitely many distinct indices i1 , . . . , ik ∈ I and nonzero
scalars c1 , . . . , ck satisfying

v = c 1 u i1 + . . . + c k u ik .

In the context of analysis, such a basis is called a Hamel basis, to distinguish

it from other notions of basis which are used. For most infinite dimensional
vector spaces a Hamel basis is almost useless, because of the following fact.
Exercise 1.2.1. A Hamel basis for CR ([a, b]) must be uncountable (for a hint
see Exercise 7.5.16).
The fact that Hamel bases for CR ([a, b]) must be uncountable suggests that
we may want to relax our notion of basis. By the exercise, we have no hope to
4 A First Course in Functional Analysis

find a countable basis. However, we can hope for the next best thing, which
is a linearly independent sequence of functions {fn }∞
n=0 ⊂ CR ([a, b]) which in
some sense span the space. A natural guess would be that the sequence of
monomials fn (t) = tn are as close to being a basis for CR ([a, b]) as one can
reasonably expect. This is true, as Weierstrass proved in 1885.
Theorem 1.2.2 (Weierstrass’s approximation theorem). Let f : [a, b] → R
be a continuous function. For every  > 0, there exists a polynomial p with
real coefficients, such that for all t ∈ [a, b],

|p(t) − f (t)| < .

Since not every continuous function is a polynomial, we cannot hope to

obtain every continuous function as a linear combination of monomials. Weier-
strass’s approximation theorem assures us that we can at least use linear com-
binations of monomials to approximate every continuous function to any given
precision. Thus the monomials can be said to “generate” CR ([a, b]).
Are there any other sequences of nice functions that generate CR ([a, b])? To
give another example, we need a couple of definitions. A Z-periodic function
is a function f : R → R such that f (x + n) = f (x) for all x ∈ R and all n ∈ Z.
A trigonometric polynomial is a function q of the form
N
X
q(x) = an cos(2πnx) + bn sin(2πnx).
n=0

Later in this book, when we will study Fourier series, we will require the
following theorem (see also Exercise 1.6.7). Like the previous theorem, this
one is also due to Weierstrass.
Theorem 1.2.3 (Trigonometric approximation theorem). Let f : R → R be a
continuous Z-periodic function. For every  > 0, there exists a trigonometric
polynomial q, such that for all t ∈ R,

|q(t) − f (t)| < .

It turns out that the most elegant way to obtain the above two theorems
is to consider a more general problem. We will obtain both Theorem 1.2.2
and Theorem 1.2.3 as simple consequences of the Stone-Weierstrass theorem,
which is a broad generalization of these theorems. Our goal in the remainder
of this chapter is to present and prove the Stone-Weierstrass theorem. Besides
obtaining useful results for later purposes in the course, this also serves to
highlight the spirit of functional analysis. (We will return to the problem of
finding a good notion of basis for infinite dimensional spaces in Chapter 3.)
 Introduction and the Stone-Weierstrass theorem 5

1.3 The Stone-Weierstrass theorem

Let X be a compact Hausdorff topological space (the appendix contains all
the material in topological and metric spaces that is required for this book).
We will let CR (X) denote the space of continuous, real-valued functions on
X; likewise, C(X) denotes the space of continuous, complex-valued functions
on X. On both of these spaces we define the supremum norm of a function
f to be
kf k∞ = sup |f (x)|.
x∈X

The quantity d(f, g) = kf − gk∞ defines a metric on CR (X) (and also on

C(X)) and is considered to be the distance between the two functions f and
g. This distance makes both CR (X) and C(X) into complete metric spaces.
Both these spaces are vector spaces over the appropriate field, with the usual
operations of pointwise addition of functions and scalar multiplication. In fact,
if f, g ∈ CR (X), then the pointwise product f g is also in CR (X), and together
with the vector space operations, this gives CR (X) the structure of an algebra.
Definition 1.3.1. If A is a subspace of CR (X) or of C(X), then it is said to
be a subalgebra if for all f, g ∈ A, f g is also in A.
Definition 1.3.2. A subalgebra A ⊆ CR (X) is said to separate points if
for every pair of distinct points x, y ∈ X there exists some f ∈ A such that
f (x) 6= f (y).
Theorem 1.3.3 (Stone-Weierstrass theorem (real version)). Let A be a closed
subalgebra of CR (X) which contains the constant functions and separates
points. Then A = CR (X).
One obtains Theorem 1.2.2 immediately by letting X = [a, b] and taking A
to be the closure of the algebra of polynomials with respect to the supremum
norm, noting that the norm closure of an algebra is an algebra.
Exercise 1.3.4. Let A ⊆ CR (X) be a subalgebra, and let A be its closure.
Then A is also a subalgebra.
Another convenient way of stating the Stone-Weierstrass theorem is given
in the following exercise.
Exercise 1.3.5. Prove that Theorem 1.3.3 is equivalent to the following state-
ment: If A is a subalgebra of CR (X) which contains the constant functions and
separates points, then A = CR (X).
To obtain the trigonometric approximation theorem (Theorem 1.2.3), one
first notes that, due to standard trigonometric identities, the trigonometric
polynomials form an algebra. Next, one needs to realize that the continuous,
6 A First Course in Functional Analysis

real-valued, Z-periodic functions on R can be identified with the continuous,

real-valued functions on the torus T = {z ∈ C | |z| = 1}. Indeed, the mapping
Φ : CR (T) → CR (R) given by

Φ(f )(t) = f e2πit

, t∈R

maps CR (T) onto the algebra CR,per (R) of continuous Z-periodic functions.
Moreover, Φ maps the constant function to the constant function, it is lin-
ear, and it respects multiplication: Φ(f g) = Φ(f )Φ(g). Therefore, the same
holds for the inverse Φ−1 : CR,per (R) → CR (T). Finally, supt∈R |Φ(f )(t)| =
sup|z|=1 |f (z)| for all f ∈ CR (T).
Now put X = T, and take A ⊆ CR (T) to be the inverse image of all
trigonometric polynomials under Φ (if we identify CR (T) and CR,per (R), then
under this identification A is simply the algebra of all trigonometric polyno-
mials, but we have chosen to make this identification explicit with the use of
the map Φ). By the previous paragraph, in order to prove Theorem 1.2.3 it
suffices to show that A is dense in CR (X). It is elementary to check that A
contains the constants and separates points on X = T. Applying the version of
the Stone-Weierstrass theorem given in Exercise 1.3.5, we find that A is dense
in CR (T), therefore Φ(A) is dense in CR,per (R). That concludes the proof of
the trigonometric approximation theorem.
Exercise 1.3.6. Fill in the details in the proof of Theorem 1.2.3. In particular,
prove that A is an algebra that separates points, and prove that Φ is surjective.

Proof of the Stone-Weierstrass theorem

Let A ⊆ CR (X) be as in the statement of Theorem 1.3.3. We isolate a few
lemmas before reaching the main argument of the proof.
Lemma 1.3.7. On every interval [−L, L], the absolute value function is uni-
formly approximable by polynomials. That is, for every  > 0, there exists a
polynomial p such that
|p(t) − |t|| <  (1.2)
for all t ∈ [−L, L].
Proof. It suffices to prove the lemma for the interval [−1, 1] (why?). To this
end, consider the function h(x) = (1 − x)1/2 . It is a standard (but nontrivial)
exercise in first-year analysis to show that the Taylor series of h about the point
0 converges uniformly in the closed interval [−1, 1]. Truncating the series at
some high power, we find, given  > 0, a polynomial q such that |q(x)−h(x)| <
 for all x ∈ [−1, 1]. Now since |t| = h(1 − t2 ), the polynomial p(t) = q(1 − t2 )
satisfies (1.2).
Exercise 1.3.8. Fill in the details of the above proof; in particular prove the
uniform convergence of the Maclaurin series of h on [−1, 1]. (Hint: use the
integral form of the remainder for the Taylor polynomial approximation.)
 Introduction and the Stone-Weierstrass theorem 7

For every function f , we let |f | denote the function |f | : x 7→ |f (x)|.

Lemma 1.3.9. If f ∈ A, then the function |f | is also in A.

Proof. Let  > 0 be given. We will find a function g ∈ A such that kg−|f |k∞ <
. Since A is closed and since  is arbitrary, this will show that |f | ∈ A.
Let I = [−kf k∞ , kf k∞ ]. By the previous lemma there exists a polynomial
p such that supt∈I |p(t) − |t|| < . Put g = p ◦ f . Since A is an algebra and p
is a polynomial, g ∈ A. Thus

kg − |f |k∞ = sup |p(f (x)) − |f (x)|| ≤ sup |p(t) − |t|| < ,

x∈X t∈I

as required.
For any two functions f, g, we let f ∧ g and f ∨ g denote the functions
f ∧ g : x 7→ min{f (x), g(x)} and f ∨ g : x 7→ max{f (x), g(x)}.
Lemma 1.3.10. If f, g ∈ A, then the functions f ∧ g and f ∨ g are also in A.
Proof. This follows immediately from Lemma 1.3.9 together with the formulas
min{a, b} = a+b−|a−b|
2 and max{a, b} = a+b+|a−b|
2 , which hold true for all real
a and b.
Lemma 1.3.11. For every pair of distinct points x, y ∈ X, and every a, b ∈ R,
there exists a function g ∈ A such that g(x) = a and g(y) = b.
Proof. Exercise.

Completion of the proof of the Stone-Weierstrass theorem. Let f ∈

CR (X). We must show that f ∈ A. It suffices, for a fixed  > 0, to find h ∈ A
such that kf − hk∞ < .
We start by choosing, for every x, y ∈ X, a function fxy ∈ A such that
fxy (x) = f (x) and fxy (y) = f (y). This is possible thanks to Lemma 1.3.11.
Next we produce, for every x ∈ X, a function gx ∈ A such that gx (x) =
f (x) and gx (y) < f (y) +  for all y ∈ X. This is done as follows. For every
y ∈ X, let Uy be an open neighborhood of y in which fxy < f + . The
compactness of X ensures that there are finitely many of these neighborhoods,
say Uy1 , . . . , Uym , that cover X. Then gx = fxy1 ∧ . . . ∧ fxym does the job (gx
is in A, thanks to Lemma 1.3.10).
Finally, we find h ∈ A such that |h(x) − f (x)| <  for all x ∈ X. For every
x ∈ X let Vx be an open neighborhood of x where gx > f − . Again we find
a finite cover Vx1 , . . . , Vxn and then define h = gx1 ∨ . . . ∨ gxn . This function
lies between f +  and f − , so it satisfies |h(x) − f (x)| <  for all x ∈ X, and
the proof is complete.
Exercise 1.3.12. Did we use the assumption that X is Hausdorff? Explain.
8 A First Course in Functional Analysis

1.4 The Stone-Weierstrass theorem over the complex

numbers
Often, one finds it more convenient to study or to use the algebra C(X)
of continuous complex-valued functions on X. Maybe the reader has not en-
countered this algebra of functions before, but its structure as a vector space
is very close to that of CR (X).

Exercise 1.4.1. Prove that C(X) = {u + iv | u, v ∈ CR (X)}.

If f = u + iv where u, v ∈ CR (X), then we denote Ref = u and Imf = v.
Thus, for every f ∈ C(X), f = Ref + iImf .
It turns out that it is harder for a subalgebra of C(X) to be dense in
C(X) than it is for a subalgebra of CR (X) to be dense in CR (X). Consider
the following example.
Example 1.4.2. Let D denote the open unit disc in C, and let D denote its
closure. Let A(D) denote the disc algebra, which is defined to be the closure
of complex polynomials in C(D), that is
( N
)k·k∞
X
A(D) = z 7→ ak z k : N ∈ N, a0 , . . . , aN ∈ C .
k=0

Certainly, A(D) is a complex algebra which contains the constants and sepa-
rates points. By a theorem in complex analysis, the uniform limit of analytic
functions is analytic. Thus, every element of A(D) is analytic in D, so this
algebra is quite far from being the entire algebra C(D).

It is worth stressing that in the above example we mean polynomials in

one complex variable z. We do not mean polynomials in the two variables x
and y, where z = x + iy = Rez + iImz.
Exercise 1.4.3. Is the space C[x, y] of two variable complex polynomials
dense in C(D)?

To make the Stone-Weierstrass theorem work in the complex-valued case,

one needs to add one additional assumption.
Definition 1.4.4. A subspace S ⊆ C(X) is said to be self-adjoint if for
every f ∈ S, the complex conjugate of f (i.e., the function f : x 7→ f (x)) is
also in S.
Theorem 1.4.5 (Stone-Weierstrass theorem (complex version)). Let A be a
closed and self-adjoint subalgebra of C(X) which contains the constant func-
tions and separates points. Then A = C(X).
 Introduction and the Stone-Weierstrass theorem 9

Proof. Consider the real vector space ReA = {Ref : f ∈ A}. Since Ref = f +f 2
and A is self-adjoint, it follows that ReA ⊆ A. Because A is a subalgebra of
C(X), ReA is a subalgebra of CR (X). From closedness of A it follows that
ReA is closed, too.
From the assumption that A is a subspace that separates points, it follows
that ReA also separates points. Indeed, given x, y ∈ X, let f ∈ A such that
f (x) 6= f (y). Then, either Ref (x) 6= Ref (y), or Imf (x) 6= Imf (y). But Imf =
Re(−if ) ∈ ReA, so ReA separates points.
Thus, ReA is a closed, real subalgebra of CR (X) that contains the constants
and separates points. By the (real) Stone-Weierstrass theorem, ReA = CR (X).
It follows that every real-valued continuous function on X is in A. Symmetri-
cally, every imaginary valued continuous function on X is in A. By Exercise
1.4.1 we conclude that C(X) = A.

1.5 Concluding remarks

Functional analysis originated from an interest in solving analytical prob-
lems such as the integral equation (1.1). Equations of this kind can be
rephrased as problems about operators acting on infinite dimensional vector
spaces. If one wishes to understand operators on infinite dimensional vector
spaces, the first thing to do is to study the spaces themselves. In this chapter
we took a look at the space CR ([a, b]), and proved Weierstrass’s approximation
theorem, which was a byproduct of the Stone-Weierstrass theorem. Besides
obtaining important theorems to be used subsequently, these results should
give a flavor of how life in infinite dimensional vector spaces is different from
what one is used to in finite dimensional spaces.
The Stone-Weierstrass theorem and the way that we have applied it serve
as an example of functional analysis at work. We had a concrete approxi-
mation problem — approximating continuous functions by polynomials or by
trigonometric polynomials — which was solved by considering a vastly more
general approximation problem. Considering a more general problem serves
two purposes. First, after we have proved the result, we have a ready-to-use
tool that will be applicable in many situations. Second, by generalizing the
problem we strip away the irrelevant details (for example, the particular na-
ture of the functions we are trying to approximate with or the nature of the
space on which they live) and we are left with the essence of the problem.
To prove the theorem it was convenient to employ the language of abstract
analysis, namely, to introduce a norm and to consider the problem inside an
algebra which is also a metric space. It was convenient to consider a closed
subalgebra A, even though there was no closed subalgebra in the original
problems, and even though this closed subalgebra turned out to be the whole
space of continuous functions.