Real Analysis: Interactive Notes For

University of Illinois at Chicago
Interactive Notes For

Real Analysis
Bonnie Saunders
August 21, 2015
ii
Preface
These notes are all about the Real Numbers and Calculus. We start from scratch
with definitions and a set of nine axioms. Then, using basic notions of sets and logical
reasoning, we derive what we need to know about real numbers in order to advance
through a rigorous development of the theorems of Calculus.
In Chapter 0 we review the basic ideas of mathematics and logical reasoning needed
to complete the study. Like Euclid’s Basic Notions, these are the things about sets and
logic that we hold to be self-evident and natural for gluing together formal arguments
of proof. This chapter can be covered separately at the beginning of a course or
referred to throughout on an ’as needed’ basis. It contains all the common definitions
and notation that will be used throughout the course.
Students already think about real numbers in different ways: decimal representation,
number line, fractions and solutions to equations, like square roots. They are familiar
with special real numbers, with infinite, non-repeating decimals, like π and e. All
these ways of representing real numbers will be investigated throughout this axiomatic
approach to the development of real numbers. The Axioms for Real Numbers come
in three parts:
The Field Axioms (Section 1.1) postulate basic algebraic properties of number: com-
mutative and associative properties, the existence of identities and inverses.
The Order Axioms (Section 1.2) postulate the existence of positive numbers. Con-
sequences of include the existence of integers and rational numbers.
The Completeness Axiom (Section 1.3) postulates the existence of least upper bound
for bounded sets of real numbers. Consequences of completeness include infinite
decimals are real numbers and that there are no ’gaps’ in the number line.
The completeness of the real numbers paves the way for develop the concept of
limit, Chapter 2, which in turn allows us to establish the foundational theorems of
calculus establishing function properties of continuity, differentiation and integration,
Chapters 4 and 5.
iii
iv
Goals
1. Prove the Fundamental Theorem of Calculus starting from just nine axioms that
describe the real numbers.
2. Become proficient with reading and writing the types of proofs used in the
development of Calculus, in particular proofs that use multiple quantifiers.
3. Read and repeat proofs of the important theorems of Real Analysis:
• The Nested Interval Theorem
• The Bolzano-Weierstrass Theorem
• The Intermediate Value Theorem
• The Mean Value Theorem
• The Fundamental Theorem of Calculus
4. Develop a library of the examples of functions, sequences and sets to help explain
the fundamental concepts of analysis.
v
Two Exercises to get started.

True or False 1
Which of the following statements are true? Explain your answer.
a) 0.9 > 1
b) 0.9 < 1
c) 0.9 = 1
Calculation 1
Using your
√ calculator only for addition, subtraction, multiplication and division, approx-
imate 56. Make your answer accurate to within 0.001 of the exact answer. Write a
procedure and explain why it works.
vi
Contents
0 Basic Notions 1
0.0 Getting Started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
0.1 Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
0.1.1 Common Sets . . . . . . . . . . . . . . . . . . . . . . . . . . 1
0.1.2 Set Notation . . . . . . . . . . . . . . . . . . . . . . . . . . 3
0.1.3 Operations on Sets . . . . . . . . . . . . . . . . . . . . . . . 4
0.2 Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
0.2.1 Logical Statements . . . . . . . . . . . . . . . . . . . . . . . 4
0.2.2 Quantifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
0.3 Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
0.3.1 Definitions, Notation and Examples . . . . . . . . . . . . . . 6
0.3.2 Sequences are functions Z+ → R or Z≥ → R . . . . . . . . . 7
0.4 True or False . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1 The Real Number System 11

1.0 Definitions and Basic Notions from Algebra . . . . . . . . . . . . . . 11
1.0.1 Equality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.0.2 Addition and Multiplication . . . . . . . . . . . . . . . . . . . 12
1.0.3 Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.1 The Field Axioms . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.1.0 Consequences of the Field Axioms . . . . . . . . . . . . . . . 15
1.1.1 Comments on the Field Axioms . . . . . . . . . . . . . . . . 19
1.1.2 Examples of Fields . . . . . . . . . . . . . . . . . . . . . . . 19
1.2 The Order Axioms . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.2.1 Consequences of the Order Axioms . . . . . . . . . . . . . . 20
vii
viii CONTENTS
1.2.2 Integers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Mathematical Induction. . . . . . . . . . . . . . . . . . . . . 24
The integers form a commutative ring . . . . . . . . . . . . . 25
The Well-Ordering Principle . . . . . . . . . . . . . . . . . . 25
Practice with Induction. . . . . . . . . . . . . . . . . . . . . 26
Inductive Definitions . . . . . . . . . . . . . . . . . . . . . . 27
1.2.3 Rational Numbers . . . . . . . . . . . . . . . . . . . . . . . . 27
1.2.4 Distance, absolute value and the Triangle Inequality . . . . . 29
Discussion of the number line Υ . . . . . . . . . . . . . . . 29
Absolute Value and Distance . . . . . . . . . . . . . . . . . . 30
The Triangle Inequality . . . . . . . . . . . . . . . . . . . . . 31
1.2.5 Bounded and unbounded sets . . . . . . . . . . . . . . . . . 33
1.3 The Completeness Axiom . . . . . . . . . . . . . . . . . . . . . . . . 36
1.3.1 Consequences of the Completeness Axiom . . . . . . . . . . 36
1.3.2 The Nested Interval Theorem . . . . . . . . . . . . . . . . . 37
1.3.3 Archimedes Principle . . . . . . . . . . . . . . . . . . . . . . 38
The integers are not bounded . . . . . . . . . . . . . . . . . 38
1.3.4 Optional – Nested Interval Theorem and Archimedes prove the
Completeness Axiom . . . . . . . . . . . . . . . . . . . . . . 39
1.3.5 Rational numbers are dense in R . . . . . . . . . . . . . . . . 40
1.3.6 Optional – An Alternative Definition of Interval . . . . . . . . 41
2 Limits of Sequences 43
2.1 Definition and examples . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.1.1 Sequences converging to zero. . . . . . . . . . . . . . . . . . 44
2.1.2 Sequences that converge to arbitrary limit . . . . . . . . . . . 47
Monotone sequences . . . . . . . . . . . . . . . . . . . . . . 50
Best Nested Interval Theorem . . . . . . . . . . . . . . . . . 51
Rational Approximations to Real Numbers . . . . . . . . . . . 51
2.1.3 Application: Existence of square roots . . . . . . . . . . . . . 52
2.1.4 Subsequences . . . . . . . . . . . . . . . . . . . . . . . . . . 53
2.1.5 Divergent Sequences . . . . . . . . . . . . . . . . . . . . . . 53
2.2 Limits and Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2.2.1 Limit Points and Boundary Points . . . . . . . . . . . . . . . 56
2.2.2 Open and Closed Sets . . . . . . . . . . . . . . . . . . . . . 58
2.2.3 Optional – Connected sets . . . . . . . . . . . . . . . . . . . 59
2.3 The Bolzano-Weierstrass Theorem and Cauchy Sequences . . . . . . 61
2.4 Series and Power series . . . . . . . . . . . . . . . . . . . . . . . . . 61
2.4.1 Convergence of Geometric series . . . . . . . . . . . . . . . . 62
2.4.2 Decimals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
CONTENTS ix
3 Counting 65
3.1 Finite vs Infinite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.1.1 The rational numbers . . . . . . . . . . . . . . . . . . . . . . 67
3.1.2 How many decimals are there? . . . . . . . . . . . . . . . . . 68
3.2 Cantor Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4 Functions 71
4.0 Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.1 Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.1.1 Sequential continuity . . . . . . . . . . . . . . . . . . . . . . 73
4.1.2 More Examples and Theorems . . . . . . . . . . . . . . . . . 74
An aside to discussion inverse functions . . . . . . . . . . . . 74
A library of functions . . . . . . . . . . . . . . . . . . . . . . 75
4.1.3 Uniform Continuity . . . . . . . . . . . . . . . . . . . . . . . 76
4.2 Intermediate Value Theorem . . . . . . . . . . . . . . . . . . . . . . 77
4.3 Continuous images of sets . . . . . . . . . . . . . . . . . . . . . . . 78
4.4 Optional: Connected Sets . . . . . . . . . . . . . . . . . . . . . . . 79
4.5 Existence of extrema . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.6 Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.6.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.6.2 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.6.3 Basic Theorems . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.6.4 Zero Derivative Theorem . . . . . . . . . . . . . . . . . . . . 81
5 Integration 85
5.0 Definition of Riemann Integral . . . . . . . . . . . . . . . . . . . . . 85
5.0.1 Partitions of an interval . . . . . . . . . . . . . . . . . . . . . 85
5.0.2 Definition of Riemann Integral . . . . . . . . . . . . . . . . . 86
5.0.3 Theorems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
5.1 Fundamental Theorem of Calculus . . . . . . . . . . . . . . . . . . . 90
5.1.1 Integrals as Functions . . . . . . . . . . . . . . . . . . . . . 90
5.1.2 Statement of the Theorem . . . . . . . . . . . . . . . . . . . 90
5.2 Computing integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
5.3 Application: Logarithm and Exponential Functions . . . . . . . . . . 91
5.4 Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
x CONTENTS
Chapter 0
Basic Notions
0.0 Getting Started

Biggles to secretary: Now, when I’ve got these antlers on - when I’ve got these antlers
on I am dictating and when I take them off (takes them off) I am not dictating.
– from ”Biggles Dictates a Letter,” Monty Python’s Flying Circus.
About moose antlers Υ: Many things in this book are already understood (or maybe
we just think we understand them) and we don’t want to forget them completely. At
the same time we develop the real number system with a minimal set of concepts to
guide us, we want to be able to use our intuition and ideas already mastered to guide
our way and help us understand. We do want to keep straight where we are in this
game. That’s where moose antlers come in. This is how it works: when moose antlers
are on, we can use what we already know to think about examples and proofs. When
they are off we only think about the axioms and theorems that we have proven so far.
Look for the Υ moose antlers throughout the book. At those points feel free to
use what mathematical knowledge and intuition you have to answer the questions.
Otherwise, what you have at your disposal is the nine axioms and any previous theorems
we have derived from those axioms using the basic notions of sets and logic that are
summarized in this chapter.
0.1 Sets
0.1.1 Common Sets

Υ Some of all of these sets will be familiar to you from previous mathematical expe-
riences. Throughout this book, we will be starting from scratch and defining each of
1
2 CHAPTER 0. BASIC NOTIONS
them. References are provided below. They are all listed here to establish common
notation. You may have used different notation for some of these sets and you may
have other common sets you’d like to include. Do not hesitate to make your concerns
known!
R represents the set of all real numbers. This set is the main interest and star of
this course. And as in all good books the character will be developed slowly
and carefully throughout the course. In the beginning, we assume a few things
about how the elements in this set behave under the operations of addition and
multiplication. This is quite abstract – we don’t have any idea what the elements
(which we will call numbers) of this set really are or even if such a set of things
exists in any “real” (You can decide if this pun is intended or not) sense. From
the axioms we will derive enough information to set up the familiar models for
real numbers are – principally, decimal representation and the number line. We
will also be able to conclude that any other system that satisfies the same axioms
is essentially the same as the real number system we describe.
R+ represents the set of positive real numbers. Defining characteristics of this set
will be established in Section 1.2
R≥ = R+ ∪ {0} represents the set of non-negative real numbers.
R2 = R × R is the set of ordered pairs of real numbers - also called the Cartesian
plane. In this book it is mostly used in reference to functions that map R to
R. In subsequent study of real analysis, Rn - ordered n-tuples of real numbers -
take more central roles.
N and Z+ both represent the set of positive integers. It is a subset of the real numbers
and we will later establish the characteristics of this set from the axioms of R.
Also called the set of Natural numbers. Very often the characteristics of these
sets are establish by The Peano Axioms. The real numbers are then constructed
from the integers. This is not the approach in this book. See Section 1.2.2.
N0 or Z≥ both denote the set of non-negative integers.
Z represents the set of all integers. From our Υantler-less point of view we know
nothing about this set. We will establish defining characteristics that will agree
school-based ideas of what integers are.
Q represents the set of all rational numbers. They can be defined after we clarify the
notion of devision and have defined the integers.
Q+ represents the set of all positive rational numbers.
∅ represents the empty set.

0.1. SETS 3
0.1.2 Set Notation
Υ Moose antlers are tricky in this section. The book will assume you know (antlers on
or not) about sets and functions and that you understand the set notation described
in this chapter. It’s part of the basic notions you’ll need to proceed. However, all the
examples of sets below require Υ moose antlers to understand. And you’ll want the
antlers on to come up with other examples.
We discuss three different ways to denote a set.
1. By list. This works perfectly for small finite sets, like {3, 36, 17}. It is also
used for infinite sets that can be listed. For example, N = {1, 2, 3, 4, · · · } or
{1, 6, 11, 16, · · · }. Describing a set this way requires that everyone knows what
rule is being used to generate the numbers.
Example 0.1 The set {2, 3, 5, · · · } might be the prime numbers or it might be
the Fibonacci numbers or it might be the integers of the form 2n + 1.
2. By condition.
Example 0.2 {x : x is a prime number}, read ”the set of all x such that x is a
prime number.”
3. Constructively by giving a formula that describes the elements of the set. For
example, {n2 : n ∈ Z}, the set of perfect square numbers. Note that you need
to describe the set of all possible values for each variable in the formula. Note
that an element of the set may be described more than once but this does not
change the set. That is, {n2 : n ∈ Z} = {n2 : n ∈ Z≥ }.
Example 0.3 {sin nπ

2 : n is an integer} = {0, 1, −1}
Exercise 0.1 ΥIs there a simpler description of {x 2 : x ∈ R}?
Exercise 0.2 ΥLet S be the set of all odd positive integers. Describe this set in each
of the three ways listed above.
Exercise 0.3 ΥFind numbers a, b, c so that the formula, an2 + bn + c, for n ∈ Z≥ ,

describes a set like the one indicated in Example 0.1
0.1.3 Operations on Sets
We can also construct sets from other sets. This book assumes you are familiar
with the union and intersection of a collection of sets and the complement of set
with respect to another one. We write the complement of X with respect to Y as
Y \ X = {y ∈ Y : y 6∈ X}.
0.2 Logic
0.2.1 Logical Statements
Throughout this book we will be proving theorems about real numbers. Theorems are
statements that are either true or false and are stated in the form ’If p, then q’ and
notated p =⇒ q, where p and q are also statements. p is called the hypothesis (or
antecedent) and q is the conclusion (or consequence) of the theorem. The proof of
the theorem proceeds from the assumed fact that p is true and goes through a series
of logically valid statements until one can conclude q.
Sometimes it is easier to show that the negation of the statement of the theorem is
false in order to prove the theorem true. Therefore it is good to understand how to
negate a statement. ∼ p denotes the negation of p. Please keep in mind that either
p is true or ∼ p is true, but not both. Here is a table of different related statements
and their negations. We will use these names for related statements throughout the
book.
statement negation of statement
p =⇒ q p and ∼ q
converse q =⇒ p q and ∼ p
contrapositive ∼ q =⇒ ∼ p ∼ q and p
A statement and its contrapositive always have the same truth value. A statement
and its converse may have different values.
Exercise 0.4 How do the following two statements fit into the table.
a) ∼ p or q
b) p or ∼ q
0.2. LOGIC 5
Sometimes the hypothesis of a theorem is not explicitly stated but it is always there
– if for no other purpose than to establish what sets the variables in the statement
belong to. Throughout the book you are encouraged to explicitly state the hypothesis
and the conclusion of theorems.
True or False 2
Which of the following statements is true? Explain. Modify the false statement to
make a true statement.
a) If x ∈ [2, 4], then x 2 ∈ [4, 16].
b) If x 2 ∈ [4, 16], then x ∈ [2, 4].
0.2.2 Quantifiers
Our logical statements will almost always contain variables and those variables may
be ’quantified.’ One way to think about it is that the statement is describing a set by
a condition. The quantifiers tell you ’how many’ numbers are in the set. There are
two different quantifiers:
1. FOR ALL Consider the following two statements and convince yourself that they
mean the same thing:
For all x ∈ R, x 2 ≥ 0 (0.1)

2
{x ∈ R : x ≥ 0} = R (0.2)
2. THERE EXISTS Consider the following two statements and convince yourself that
they mean the same thing:
There exists x ∈ R, x 2 = 9 (0.3)

2
{x ∈ R : x = 9} =
6 ∅ (0.4)
Exercise 0.5 Write the negation of each of the above statements.
How to prove statements that contain quantifiers is a main concern in later discussions
about real numbers.
0.3 Functions
No Υmooseantlers needed for this section. Everything is defined in terms of sets and
using basic logic!
0.3.1 Definitions, Notation and Examples

Definition A function, f , is a set of order pairs in R × R with the property that for
each first coordinate x, there is a unique second coordinate, y , such that (x, y ) is
in the function. Because the y is unique for a given x, we can unambiguously write
y = f (x), the customary functional notation. The domain of f is the set of all first
coordinates in f and the target of f is a set that contains all second coordinates in
f . In these notes the target will always be a subset R. Thus the function is the set,
{(x, f (x)) : x ∈ D}. Sometimes this set is called the graph of f , but the graph is not
a separate object from the function, as we have defined it.
Here is a long list of definitions and notation conventions that we will use throughout
the book. It is assumed they are mostly familiar to the reader:
Notation If D is the domain of a function f , we write f : D → R
Definition The following definitions apply to the function, f : D → R. These notions

should be familiar as they are fundamental to any study of mathematics.
· For any S ⊂ D, we say the image of S under f , and write f (S), to mean
{f (x) : x ∈ S}. The image of f is f (D).
· For any T ⊂ f (D), we say the pre-image of T under f , and write f −1 (T ), to

mean {x : f (x) ∈ T }.
· We say that f is 1 − 1 or f is an injection, if for all x, y ∈ D

f (x) = f (y ) =⇒ x = y ,
· If E ⊂ R is the target of f and if f (D) = E, we say that f is onto E ⊂ R, In
this case, we may also say that f is a surjection.
· We say that f is bijection, if it is both an injection and surjection. In this case,

we also say that f shows a 1 − 1 correspondence between D and E.
Example 0.4 The pre-image of [.25, 1] under the function x → x 2 is [−1, −.5] ∪ [.5, 1]
Exercise 0.6 Give an example of each of the following. Include the domain as part of
the description. Sketch the graph of the function.
a) a function that is not 1 − 1 but is onto
b) a function whose pre-image of some set T is R+
c) a function whose image is R+
0.3. FUNCTIONS 7
1
Figure 0.1: Graph of sn =
n
8
Figure 0.2: The image of sn = .
n3
0.3.2 Sequences are functions Z+ → R or Z≥ → R

Definition A function whose domain is Z+ or Z≥ is called a sequence.
Notation A sequence is most usually denoted with subscript notation rather than
standard function notation, that is we write sn rather than s(n).
Example 0.5 The graph in Figure 0.1 shows part of the graph of a sequence that maps
1
Z+ → R and is given by the formula, sn = . In addition, the first 100 numbers in
n
the image of the sequence on the y − axis.
Example 0.6 Another way to picture a sequence is to plot the image on a number
line, as shown in Figure 0.2. The downside is that the order of the sequence is not
8
explicitly given. Here the image of the sequence, sn = 3 , is shown on a horizontal
n
number line. The order of the sequence values is not shown on this picture. You need
to see the formula, as well, to understand that the values are being listed in order from
right to left. The values in the image bunch up at zero to become indistinguishable
from each other and from 0. The picture is insightful, but imprecise.
Notation Because of the ordering of the natural numbers, a sequence can be given
by listing the first few values without reference to the domain or a formula, as in
3 1 3 1
3, , , , ··· (0.5)
4 3 25 12
This type of notation can be convenient but it never tells the whole story. How does
the sequence continue past the values given? The finite sequence may suggest a
pattern but one can’t be sure without more information. We don’t know whether the
domain starts at 0 or 1, but a formula could be adjusted to fit either case. NOTE: It
would be wrong to include braces {} around the sequence because that would indicate
a set. It would be how to denote the image of the sequence.
Exercise 0.7 Find a possible formula that would generate the sequence in Example
0.5. Sketch the graph of this sequence.
0.4. TRUE OR FALSE 9
0.4 True or False
The following True or False problems explore different logical statements using quan-
tifiers in a variety of ways.
True or False 3
Which of the following statements are true? Explain. Change Z to R and redo.
a) There exists an x ∈ Z, such that x is odd.
b) For all x ∈ Z, x is even.
c) There exists an x ∈ Z, such that 2x is odd.
d) For all x ∈ Z, 2x is even.
True or False 4
(From Morgan) Which of the following statements are true? Explain.
a) For all x ∈ R, there exists a y ∈ R such that y > x 2 .
b) There exists an y ∈ R, such that for all x ∈ R, y > x 2 .
c) There exists an y ∈ R, such that for all x ∈ R, y < x 2 .
d) For all a, b, c ∈ R, there exists x ∈ R such that ax 2 + bx + c = 0.
True or False 5
Which of the following statements are true? Explain.
a) There exists a real number, x, such that x 2 = 9.
b) There exists a unique real number, x, such that x 2 = 9.
c) There exists a unique positive real number, x, such that x 2 = 9.

The following True or False problem concerns the notion of pre-image.
True or False 6
Which of the following statements true? Prove or give a counterexample. Consider

conditions of f that would make the statements True.
a) f (X ∩ Y ) = f (X) ∩ f (Y )
b) f −1 (X ∩ Y ) = f −1 (X) ∩ f −1 (Y )
c) f (f −1 (Y )) = Y.
d) f −1 (f (Y )) = Y.
Exercise 0.8 In some cases, it may be easier to determine if the negation of a statement
is true or false. If you haven’t already, write the negation of each statement in the
True or False problems.
Chapter 1
The Real Number System
We begin by supposing the existence of a set, R, whose elements we call real

numbers We suppose we know a few things about numbers. We know when two of
them are equal, section 1.0.1. We know that we can add or multiple two of them
and get an unique number, section 1.0.2. We know some conventions for how to
write expressions involving adding and multiplying, section 1.0.3. Most importantly
we postulate nine Axioms, the defining characteristics of real numbers. These include
the Field Axioms, Section 1.1; the Order Axioms, Section 1.2; and the Completeness
Axiom, Section 1.3.
We do not assume that we can represent real numbers as decimals. Nor how to
represent real numbers on a number line. However, ourΥintuition using these two
models for real numbers can guide our thinking.
1.0 Definitions and Basic Notions from Algebra
1.0.1 Equality
Basic Notion 1 EQUALITY OF REAL NUMBERS IS AN EQUIVALENCE RELATION. The following

properties apply for all real number x and y :
Reflexive x = x
Symmetric if x = y , then y = x
Transitive if x = y and y = z, then x = z
11
12 CHAPTER 1. THE REAL NUMBER SYSTEM
1.0.2 Addition and Multiplication
We assume the existence of a set of two binary operations on the ’numbers’ in this
set. Basic notions about equality apply. Both addition and multiplication produce a
unique, answer, meaning that adding a number a and multiplying by a number m are
both functions.
Basic Notion 2 UNIQUENESS OF ADDITION For all a, x, y ∈ R.
x = y =⇒ a + x = a + y .
Basic Notion 3 UNIQUENESS OF MULTIPLICATION For all m, x, y ∈ R,
x = y =⇒ m · x = m · y .
It is often useful, and some people prefer, to consider addition and multiplication as
functions. That is, for every real number a, there is a function, sa : R → R, given by
sa (x) = x + a,
and, for every real number m, there is a function, tm : R → R, given by
tm (x) = m · x.
The UNIQUENESS OF ADDITION AND MULTIPLICATION says that these functions are indeed
functions, i.e. there is only one value for each element in the domain.
The uniqueness of these operations is used in our preliminary work when doing
things like adding the same number to both sides of an equation.
1.0.3 Expressions
Binary means that the operation works on only two numbers at a time, so expres-
sions like a + b + c aren’t meaningful until we know more about what rules apply.
However, we can include parentheses in expressions and so legitimately know what to
do. The expression, a + (b + c), is meaningful: first add b to c, then add a to the
result. This use of parentheses is assumed familiar to the student of this book. (Υ
Once we know the associative and commutative rules, a + b + c is not an ambiguous
expression.)
1.0. DEFINITIONS AND BASIC NOTIONS FROM ALGEBRA 13
Basic to working with equations with variables and real numbers is being able to
’substitute’ equal expressions for each other. This is how we can build more and more
complicated, and thus interesting, expressions. This is however, not an easy concept
to formalize. Here is a stab at it:
Basic Notion 4 SUBSTITUTION If u is a real number such that u = E(x, y , z, · · · ),

where E(x, y , z, · · · ) is any legitimate expression for a real number involving other
real numbers x, y , z, · · · , then u may be interchanged for E(x, y , z, · · · ) in any other
expression without changing the value of that expression. NOTE: transitivity, from
Basic Notion 1, is an elementary example of substitution.
Υ Eventually, we want the substitution principle to apply to more sophisticated ex-

pressions like sin(a + b) or r e −3c or lim f (x), so we use the (admittedly) imprecise
x→p
language, ’legitimate expression’. Once one learns that these expressions, or oth-
ers, represent real numbers, we are free to use substitution on such expressions: If
u = sin(x), then sin2 (x) + 2 sin(x) + 1 = u 2 + 2u + 1
1.1 The Field Axioms

There is a set R of real numbers with an addition ( + ) and a multiplication ( · )
operator, that satisfy the following properties:
Axiom 1 COMMUTATIVE LAWS. For all real numbers, a and b,
FOR ADDITION a+b =b+a

FOR MULTIPLICATION a · b = b · a
Axiom 2 ASSOCIATIVE LAWS. For all real numbers, a, b and c,
FOR ADDITION a + (b + c) = (a + b) + c
FOR MULTIPLICATION a · (b · c) = (a · b) · c
Axiom 3 DISTRIBUTIVE LAW. For all real numbers, a, b and c,
a · (b + c) = a · b + a · c
Axiom 4 EXISTENCE OF IDENTITY ELEMENTS.
FOR ADDITION There is a real number, 0, such that, a + 0 = a, for all a ∈ R

FOR MULTIPLICATION There is a real number, 1, such that, 1 · a = a, for all a ∈ R
FURTHERMORE, 0 6= 1
Axiom 5 EXISTENCE OF INVERSES.
FOR ADDITION For all a ∈ R, there is an x ∈ R, such that a + x = 0.

FOR MULTIPLICATION For all a ∈ R, a 6= 0, there is an x ∈ R, such that a · x = 1.
1.1. THE FIELD AXIOMS 15
1.1.0 Consequences of the Field Axioms
In this section we state and prove many facts about real numbers. We call these
facts theorems and, occasionally, corollaries or lemmas. Theorems are not always
stated explicitly in the form p =⇒ q, but they can be stated that way. Often the
only hypothesis is that the elements being discussed are real numbers, i.e. we are
always assuming that there is a set of real numbers that satisfy the basic axioms as
well as all theorems we prove.
Exercise 1.1 Provide proofs for Theorems 1.0 - 1.15. Some of the proofs are provided.
Theorem 1.0 For all a ∈ R, 0 + a = a and 1 · a = a
Theorem 1.1 CANCELLATION LAW FOR ADDITION If a + b = a + c, then b = c.
Proof of Theorem. Let e be a real number from Axiom 5 such that a + e = 0. By

Axiom 1, it is also true that e + a = 0. Note: here we use what is sometimes called
a two column proof: the left side is a valid conclusion following from the previous
statements and the right side is the justification or warrant for that conclusion.
a+b =a+c given, the hypothesis

e + (a + b) = e + (a + c) uniqueness of addition, add e to both sides
(e + a) + b = (e + a) + c associative law for addition, Axiom 2
0+b =0+c by substitution, as stated above: e + a = 0
b=c Theorem 1.0
Theorem 1.2 The number, 0, the additive identity of Axiom 4, is unique.
Proof. Note: the strategy used to show that a number is unique is to assume there
are two numbers that satisfy the given condition and then show that they are equal.
Assume there exists another real number 00 such that a + 00 = a for all a ∈ R, then
for all a ∈ R, we have that a + 0 = a + 00 . By the CANCELLATION LAW FOR ADDITION,
Theorem 1.1, we can cancel the a’s to get, 0 = 00 . This shows that 0 is unique.
Theorem 1.3 EXISTENCE AND UNIQUENESS OF SUBTRACTION For all a, b ∈ R, there is a

unique solution, x ∈ R, to the equation a + x = b.
Notation The unique solution to the equation, a + x = b, is denoted by b − a. The

unique solution to the equation, a + x = 0, 0 − a, is written simply −a. We call it the
additive inverse of a or the negative of a.
Proof. Let e be a real number from Axiom 5 such that a + e = 0 and let x = e + b.
Now consider
a + x = a + (e + b) substituting our definition of x

= (a + e) + b associative law for addition, Axiom 2
=0+b substituting 0 for a + e
=b Theorem 1.0
To show uniqueness, assume that there is another real number, y , such that a+y = b.
Then a+x = a+y . So by the CANCELLATION LAW FOR ADDITION, Theorem 1.1, x = y .
Theorem 1.4 For all a ∈ R,

−(−a) = a.
Proof. Given any number a ∈ R,
−a + a = a + (−a) commutative law for addition, Axiom 1

a + (−a) = 0 definition of −a from Theorem 1.3.
−a + a = 0 transitive property of equality
So a is the negative of −a. The negative of −a is written, −(−a), as seen in the note
after Theorem 1.3.
Theorem 1.5 ADDITION DISTRIBUTES ACROSS SUBTRACTION For all a, b, c ∈ R,
a(b − c) = ab − ac.
Proof.
a(b − c) + ac = a((b − c) + c) distributive law, Axiom 3

a((b − c) + c) = ab definition b − c from Theorem 1.3
a(b − c) + ac = ab transitive property of equality
a(b − c) = ab − bc EXISTENCE AND UNIQUENESS OF SUBTRACTION, Theorem 1.3
Theorem 1.6 For all a ∈ R,

0 · a = a · 0 = 0.
Proof. This is a special case of Theorem 1.5, where b = c.
Theorem 1.7 CANCELLATION LAW FOR MULTIPLICATION For all a, b, c ∈ R,
ab = ac and a 6= 0 =⇒ b = c.
Proof. EFS
Theorem 1.8 The number, 1, of Axiom 4 is unique.
Proof. EFS
Theorem 1.9 EXISTENCE AND UNIQUENESS OF DIVISION For all a, b ∈ R, a 6= 0, there is a

unique solution, x ∈ R, to the equation a · x = b.
Notation The unique solution to the equation a · x = b is written ba . The unique

solution to the equation a · x = 1, 1a , is also written as a−1 . It is the multiplicative
inverse of a and also callied the reciprocal of a.
Proof. EFS
Theorem 1.10 THERE ARE NO ZERO DIVISORS
a · b = 0 =⇒ a = 0 or b = 0.
Proof. EFS
Example 1.1 Modular arithmetic is an example where there are zero divisors. Because
2 · 4 ≡ 0 mod 8 and neither is ≡ 0 mod 8 we call both 2 and 4 zero divisors mod 8.
Theorem 1.11 For all a ∈ R, a 6= 0,

1
(a−1 )−1 = a or 1 = a.
a
Proof. EFS
Theorem 1.12 For all a, b ∈ R, b 6= 0,

a
= a · b−1 .
b
Proof. EFS
Theorem 1.13 ADDITION OF FRACTIONS For all a, b, c, d ∈ R, b 6= 0 and d 6= 0,

a c ad + bc
+ =
b d bd
Proof. EFS
Theorem 1.14 MULTIPLICATION OF FRACTIONS For all a, b, c, d ∈ R, b 6= 0 and d 6= 0,

a c a·c
· =
b d b·d
Proof. EFS
Corollary 1.14.1 For all a, b ∈ R, a 6= 0 and b 6= 0,
1 1 1
= · or (a · b)−1 = a−1 · b−1
ab a b
Proof. EFS
NOTE: Sometimes corollary 1.14.1 is proven first as a lemma and is then used to
prove theorem 1.14.
Theorem 1.15 EXISTENCE AND UNIQUENESS OF SOLUTION TO LINEAR EQUATIONS For all a, b, c ∈
R, a 6= 0, there is a unique solution, x ∈ R, to the equation
a · x + b = c.
Proof. EFS
Exercise 1.2 Prove using the Axioms 1 through 5 and the Theorems 1.0 - 1.15.
For all a, b, c, d ∈ R.
a) −0 = 0; 1−1 = 1
b) −(a + b) = −a − b
c) −(a − b) = −a + b
d) (a − b) + (b − c) = a − c
a
b ad
e) c =
d bc
a −a a
f) −( ) = =
b b −b
1.1.1 Comments on the Field Axioms
Addition vs multiplication Except for the Distributive Law, Axiom 3, and the part
of Axiom 5 where 0 is excluded from having a multiplicative inverse, the axioms
are symmetric in addition and multiplication. Axiom 3 says that ’multiplication dis-
tributes across addition.’ What would happen if the opposite were true that ’addition
distributes across multiplication?’
Exercise 1.3 Write down a rule that would say that addition distributes across multi-
plication. Prove that it cannot be true if the Field Axioms are true.
Exercise 1.4 Something to contemplate: Why is it necessary to exclude a multiplicative

inverse for 0?
Associativity and commutativity Without parentheses, we do not know how to re-

solve an expression like a + b + c or a · b · c The Associative Law, Axiom 2, tells us
that the two ways to add parentheses will give you the same answer. This is worth
pondering with specific numbers (Υrequired). Consider the expression, 3 + 4 + 5. We
can resolve to either 7 + 5 or 3 + 9. We get the same thing, 12, either way.
Exercise 1.5 How many ways can you add parentheses to 2+3+4+5 to get a different
way to sum the numbers? Let cn = the number of ways to put parentheses on an
addition string with n numbers. What is cn ? (These are well-known as the Carmichael
Numbers.)
1.1.2 Examples of Fields
A set, together with a well-defined addition and multiplication, is called a Field if the
Field Axioms (Section 1.1) are all satisfied.
Example 1.2 Υ Think of some other number systems, such as

a) Q, the rational numbers.
b) R2 , the plane of ordered pairs of real numbers.
c) C, the complex numbers.
d) Z/Zn, integers mod n
e) Q[x], polynomials with rational coefficients
f) Q(x), rational functions (ratios of functions ∈ Q[x]).

Does each system have a well-defined addition and multiplication? Which satisfy
the Field Axioms?
1.2 The Order Axioms

ΥSome geometric concerns One useful way to represent real numbers is on a number
line. To do this we need rules to decide how to place the numbers. When is one
number to the right or to the left of another number? Put another way, when is
one real number ’larger’ than another? Or even, what does ’larger’ mean? The
standard approach is to first decide which ones are ’larger’ than 0, section 1.2. So
we can declare that ’b is larger than a’ whenever ’b − a is larger than 0.’ The Order
Axioms, then, are closely related to notions of distance and length, section 1.2.4.
More surprisingly, perhaps, is that the Order Axioms allow us to think about integers,
section 1.2.2. Then rational numbers can be defined, section 1.2.3. In this section we
see how these concepts are developed from axioms. There exists a subset, R+ ⊂ R,
with the following properties:
Axiom 6 If a and b are in R+ , then a + b ∈ R+ and a · b ∈ R+ .
Axiom 7 If a 6= 0, then either a ∈ R+ or −a ∈ R+ but not both.
Axiom 8 0 6∈ R+
1.2.1 Consequences of the Order Axioms

Definition We say that a real number, x, is a positive number whenever x ∈ R+ . We
say that a real number, x, is a negative number whenever −x ∈ R+ .
Notation If b − a is a positive number, we write a < b or b > a. In this case, we say

’a is less than b’ or ’b is greater than a.’
An immediate and important consequence of the order axioms is:
Theorem 1.16 1 is a positive number.
/ R+ , then −1 ∈ R+ which would mean that

Proof. By Axiom 7 if 1 ∈
(−1) · (−1) = 1 ∈ R+ .
Which shows not only that 1 ∈ R+ but also that −1 ∈
/ R+ .
Exercise 1.6 Provide proofs for Theorems 1.17 - 1.21. Some of the proofs are provided.
Theorem 1.17 LAW OF TRICHOTOMY For all a, b, c ∈ R, exactly one of the following are
true:
a = b, a < b, b < a
NOTE: In the special case, when one of a and b is zero the Law of Trichotomy says
that a real number is exclusively positive, negative, or zero.
1.2. THE ORDER AXIOMS 21
Proof. EFS
Theorem 1.18 TRANSITIVITY For all a, b, c ∈ R,
a < b and b < c =⇒ a < c
Proof. EFS
NOTE: ’Less than’ (<) forms a relation between numbers. We have just shown that
it is a transitive relation. However, it is neither symmetric nor reflective so it is not
an equivalence relation like ’=’. See Basic Notion 1
Exercise 1.7 There are three variations of TRANSITIVITY when ’≤’ replaces ’<’ in one or
the other or both of spots in the hypothesis. State each one, providing the strongest
conclusion in each case. Prove at least one of your statements. Use TRANSITIVITY ,
Theorem 1.18 rather than repeating proofs.
Exercise 1.8
Notation There are many different varieties of intervals. Write each one of the
following using set notation:
(a, b) = {x ∈ R : a < x and x < b}. Sometimes abbreviated: a < x < b
[a, b]
(a, b]
[a, b)
(a, ∞)
[a, ∞) = {x ∈ R : a ≤ x}.
(∞, b)
(∞, b]
Exercise 1.9 How would you notate the set, {x : a < x or x < b}? How does it vary
with whether or not a < b or b < a?
Theorem 1.19 ADDITION PRESERVES ORDER If a is a real number, then
x < y =⇒ a + x < a + y
Proof. EFS
Theorem 1.20 If a < b and c < d, then a + c < b + d

Proof. EFS
Multiplication preserves order only when the multiplication factor is positive:
Theorem 1.21 If m > 0, then multiplication by m preserves order, that is
x < y =⇒ m · x < m · y .
If m < 0, then multiplication by m reverses order, that is
x < y =⇒ m · x > m · y .
Proof. EFS
Exercise 1.10 Here are some more basic facts about order. You may want to prove
them in a different ’order.’
a) For all a ∈ R, if a 6= 0, then a2 > 0.
b) If a < b, then −b < −a.
c) If a > 0, then a−1 > 0.
d) If 0 < a < b, then a−1 > b−1 > 0.
e) The sum of two negative numbers is negative.
f) The product of two negative numbers is positive.
g) The product of a negative number and a positive number is negative.
h) For all a, b ∈ R, a2 + b2 = 0 ⇐⇒ a = b = 0.
True or False 7
If true, prove the statement. If false, restate to make a true fact and prove it.
a) x 2 > x.
b) If w1 , w2 > 0 and w1 + w2 = 1, then
a < b =⇒ a < w1 a + w2 b < b
c) If a < b and c < d, then ac < bd.

Definition We say a function, f : D → R, is increasing whenever, for all x, y ∈ D,
x ≤ y =⇒ f (x) ≤ f (y ).
We say a function is strictly increasing whenever, for all x, y ∈ D,
x < y =⇒ f (x) < f (y ).
We say a function is decreasing whenever, for all x, y ∈ D,
x ≤ y =⇒ f (y ) ≤ f (x).
We say a function is strictly decreasing whenever, for all x, y ∈ D,
x < y =⇒ f (y ) < f (x).
Example 1.3 The function sa (x) = x + a is increasing by Theorem 1.19.
Exercise 1.11 When is multiplication by m an increasing function? When is it decreas-

ing?
Exercise 1.12 Prove that the function, f (x) = x 2 , is increasing on [0, ∞) and decreas-
ing on (−∞, 0].
Exercise 1.13 Prove that the function, f (x) = x1 , is decreasing on (0, ∞) and decreas-
ing on (−∞, 0).
x+1
Exercise 1.14 Prove that the function, f (x) = x−1 , is decreasing on (1, ∞).
1.2.2 Integers
ΥMathematical induction. We have mentioned earlier that another way to develop

the real number system is to start with integers and the Peano Axioms then construct
the rational and irrational numbers. With our current endeavor, however, we still have
nothing to say about integers. We do know 1 and so could define integers: 2 = 1 + 1
and 3 = 2 + 1 and so on. Of course, it is the ’so on’ that leaves us with less precision
than we like for a usable definition. But we have the idea of an inductive process and,
in this section, we see how to make induction part of our development. We start with
a definition:
Definition We say that a subset of R is an Inductive Set whenever both of the following
conditions hold:
• 1∈S
• If n ∈ S, then n + 1 ∈ S
Example 1.4 The set of positive real numbers is an inductive set. 1 ∈ R+ by Theorem
1.16 . The second condition follows because the sum of two positive real numbers is
a positive real number by Axiom 6.
Exercise 1.15 What is the largest inductive set you can think of? What is the smallest?
Definition We say that a real number is a positive integer if it is contained in every

inductive set.
Notation We denote the set of all positive integers as Z+ .
Z+ is the smallest inductive set in the sense that it is contained in every other one.
Theorem 1.22 The positive integers are positive real numbers.
Proof. This is because the set of positive real numbers is an inductive set, so every
positive integer is contained in it.
Definition The negative integers are {−n : n ∈ Z+ }, the negative positive integers,
denoted by Z− . The integers are the positive integers together with the negative
integers and 0, denoted by Z = Z+ ∪ {0} ∪ Z− .
Exercise 1.16 Prove: There is no integer in the open interval (0, 1).
Mathematical Induction.
To establish the algebraic structure of the integers there is some work to do and
Mathematical Induction will be a major tool. The following theorem establishes the
legitimacy of the induction procedure.
Theorem 1.23 MATHEMATICAL INDUCTION Let S be a set of positive integers that is an

inductive set, i.e. satisfies the following two conditions:
• 1∈S
• If n ∈ S, then n + 1 ∈ S
then S = Z+ .
Proof. By definition, S is an induction set so Z+ ⊂ S. By hypothesis, S ⊂ Z+ . So

S = Z+ .
This theorem is the basis for proof by mathematical induction: To prove a fact by
mathematical induction, first restate the fact as a statement about a subset of positive
integers. For example, define a set of positive integers, S, such that n ∈ S if and only
if some property, P (n), is true. Then show that S is an inductive set by
• Showing a ’base case,’ P (1) is true (1 ∈ S)

• Showing an inductive step: If P (n) is true (n ∈ S), then P (n + 1) is true
(n + 1 ∈ S).
Finally, apply Theorem 1.23 to conclude that S is all positive integers, so P (n) is true
for all n ∈ Z+ . Υ By all means use methods you have used before and are comfortable
with, but do understand how the process fits into the grand scheme of things.
The first theorems we will prove with induction establish the algebraic structure of the
integers.
The integers form a commutative ring
Since the integers are a subset of the real numbers, they satisfy of all the field axioms,
except for the existence of multiplicative inverses, which is not ring axiom. So we only
need to show closure:
Theorem 1.24 ALGEBRAIC PROPERTIES OF INTEGERS
1. The sum of two integers is an integer.
2. The product of two integers is an integer.
3. The negative of an integer is an integer.
Outline of proof for sums. First, fix a positive integer m. Use induction to show that
{n ∈ Z+ : m + n ∈ Z} = Z+ and {n ∈ Z+ : −n + m ∈ Z} = Z+ . Finally, show that
the sum of two negative integers is an integer without another induction proof.
Exercise 1.17 Prove: The only two integers that have a multiplicative inverse are 1
and −1.
The Well-Ordering Principle
When establishing integers from axioms, MATHEMATICAL INDUCTION is sometimes used

as an axiom. Sometimes the following theorem is used instead and this theorem will
be useful for us later. In any case, it can be proved with mathematical induction.
Theorem 1.25 THE WELL-ORDERING PRINCIPLE Every non-empty set of positive integers
contains a smallest integer.
Proof. Let W be a subset of the positive integers that does not contain a smallest
element. We will show that W = ∅. Let S = {k : [1, k] ∩ W = ∅}. We will show
that S is an inductive set.
• 1 ∈ S because if it were not in S it would be in W and it would be the smallest

element in W .
• Assume n ∈ S. This means that no k ≤ n are in W . Now if n + 1 ∈ W it
would be the smallest integer in W , but W does not have a smallest element so
[1, n + 1] ∩ W = ∅. In other words, n + 1 ∈ S.
Therefore S is an inductive set of positive integers, so must be all of them. In other

words W is empty. We conclude that any non-empty set of positive integers must
contain a smallest element.
NOTE: Do not confuse the above theorem with THE WELL-ORDERING THEOREM, a theo-
rem dependent on the Axiom of Choice. The Axiom of Choice is often included in
the axioms for set theory despite certain bizarre behavior such as the Banach-Tarski
Paradox. These considerations become of more interest in the the study of Lebesgue
integration and will not come up for us in this course.
Exercise 1.18 Use the THE WELL-ORDERING PRINCIPLE to prove that there is no positive
integer M such that 2k < M for all k ∈ Z+ .
Exercise 1.19 You may have noticed that our proof of THE WELL-ORDERING PRINCIPLE,
Theorem 1.25, could be simplified by using Strong Induction. Find a good statement of
strong induction and prove it using MATHEMATICAL INDUCTION, Theorem 1.23. Proceed to
rewrite the proof of THE WELL-ORDERING PRINCIPLE, Theorem 1.25, using strong induction.
Practice with Induction.
Exercise 1.20 For practice with mathematical induction, prove the following two the-
orems. We will need both of them later in the course.
n
X 1
Theorem 1.26 For all positive integers n, k 2 = · n · (n + 1) · (2n + 1)
k=0
6
Proof. EFS
Theorem 1.27 BERNOULLI’S INEQUALITY For any positive real number, x, and for all
positive integers n, (1 + x)n ≥ 1 + n · x
Proof. EFS
Exercise 1.21 Bernoulli’s Inequality is useful for many things. Use it to prove that
1 1
< , for all n > 0.
2n n
Inductive Definitions
This way to define integers is an example of using an inductive definition. Defining a

sequence by recursion is another. The first few elements in the sequence are given and
the rest are defined in terms of the previous ones. For example we define a sequence,
sn by recursion (or inductively) by first defining s0 = 1 and then, once we have defined
sk for k < n, define sn = n + 4. Now we understand the sequence to be 1, 5, 9, 13....
A direct or, closed form, description of the sequence is simply sn = 1 + 4 · n. Later,
we will see examples of defining sequences of sets recursively when using the NESTED
INTERVAL THEOREM.
Exercise 1.22 List the first 10 numbers in the following sequence, given by a recursive
formula.
a) x0 = 1 and xn = xn−1 + 5.
b) x0 = 1, x1 = 1 and xn = xn−2 + xn−1 .
Exercise 1.23 Identify a pattern for the following sequences. Write a recursive formula
and a closed formula to describe each one.
a) 6, 18, 54, 162, · · ·
b) 2.0, 0.2, 0.02, 0.002 · · ·
Exercise 1.24 Write a recursive formula

r to capture and continue this sequence:
√ p √ √ √
q p q p
2, 2 + 2, 2 + 2 + 2, 2 + 2 + 2 + 2, · · ·
Exercise 1.25 ΥMake a formal √ inductive definition that captures the procedure you
n
established for approximating 56, accurate to within (0.1) .
1.2.3 Rational Numbers
Definition We say that a real number, r , is a rational number whenever there exist
integers n and m such that
n
r=
m
We denote the set of all rational numbers by Q.
Exercise 1.26 Prove the following two theorems to establish the structure of the
rational numbers.
Theorem 1.28 ALGEBRAIC PROPERTIES OF Q
1. 0 and 1 are rational numbers.
2. The sum of two rational numbers is a rational number.
3. The product of two rational numbers is a rational number.
4. The negative of a rational number is a rational number.
5. The multiplicative inverse a rational number is a rational number.
Proof. EFS
Theorem 1.29 Q+ = {r ∈ Q : r > 0} satisfies the Order Axioms.
Proof. EFS
Exercise 1.27 Use the previous two theorems to support the claim: the rational num-
bers form an Ordered Field.
Exercise 1.28 The following exercise about even and odd integers gives enough ammu-
nition to show that the square root of 2 cannot be rational. Think of a good definition
for an even integer and an accompanying good definition for an odd integer. From
your definitions, prove the following things (not necessarily in this order; find the order
that works well for your definitions):
a) The sum of two even integers is even; the sum of two odd integers is even; the
sum of an even integer and an odd integer is odd.
b) Zero is an even; 1 is an odd; every integer is either odd or even, but not both.
c) x is odd iff −x is odd; x is even iff −x is even.
d) x is odd iff x + 1 is even; x is even iff x + 1 is odd.
e) Every integer (positive or negative or zero) is either odd or even.
f) The product of two odd integers is odd; the product of an even integer and any
integer is even.
g) For any integer n, 2n is an even integer and 2n + 1 is an odd integer.
h) If n2 is an even integer, then n is also an even integer.

i) Every integer can be expressed as a power of 2 times an odd integer.

m
j) Every rational number can be written as n where both m and n are integers and
at least one of them is odd.
If you cannot prove these things from your definitions, you will need to change your
definitions so that you can. We can now prove the following theorem.
Theorem 1.30 There is no rational number, s, such that s 2 = 2.
Proof. Assume not, so there is a rational number s = mn and s 2 = 2. By Exercise

2
1.28 j) we can assume that one of n or m is odd. Now s 2 = 2 = mn 2 or 2m2 = n2 .
n2 is even so n is even by Exercise 1.28 h). Let k be the integer such that n = 2k.
Then 2m2 = 4k 2 and m2 = 2k 2 . Now Exercise 1.28 j) says m is even, in contradiction
to the assumption that one of m and n is odd. So there can be no rational number
whose square is 2.
NOTE: Nothing is being said about whether or not the square root of 2 actually
exists, only that, if it does exist, it cannot be a rational number. Because the rational
numbers satisfy all of our axioms so far, we know we need more to be able to say that
there is a real number, s, such that s 2 = 2. The final axiom is THE COMPLETENESS AXIOM
which provides what is needed and we will get there soon.
1.2.4 Distance, absolute value and the Triangle Inequality
Discussion of the number line Υ

It’s easy to figure out a way to mark real numbers on a line using compass and
straightedge constructions from Euclidean geometry. First, label one point on the line
0 and another 1. By convention, 1 is to the right of 0. Set the compass to the length
of the segment determined by these two points. Use this setting to mark off 2, 3 and
so on. By going the other way, mark off −1, −2, −3, · · · Rational numbers can be
included by employing similar triangles. Addition is defined by copying a segment next
to another segment. Multiplication can be defined by similar triangle constructions.
But
√ this does not get all of the real numbers because we know we can construct
2 – it is the diagonal of a square which has sides of length 1. However, there are
many other real numbers that can not be constructed. Eventually we will be able to
locate numbers by using THE COMPLETENESS AXIOM. We do not intend to development
the number line rigorously but we do use it to enhance our intuition. In particular, we
are about to introduce the idea of the distance between to real numbers.
Absolute Value and Distance
Definition The absolute value of x, written as |x|, defines a function. The value of
the function is given in parts:
(
x if x ≥ 0
|x| =
−x if x ≤ 0.
Theorem 1.31 Four obvious facts about any real number, x:
1. If x > 0, then x = |x|
2. If x < 0, then x = −|x|
3. |x| = 0 ⇐⇒ x = 0.
4. If x 6= 0, then |x| > 0.
Proof. EFS
Definition We say that the distance between a and b is |b − a|. We say that the
length of an interval is the distance between its endpoints.
It can be useful to think about absolute value in the more intuitive ideas of distance.
For example, Theorem 1.31 can be restated: The distance between any real number
and zero is positive, unless that number is zero - in which case, the distance is zero.
Theorem 1.32 The distance between two numbers is zero if and only if the two
numbers are equal.
Proof. |a − b| = 0 ⇐⇒ a − b = 0 ⇐⇒ a = b.
The following, while obvious, is useful because it can eliminate cases when proving
the THE TRIANGLE INEQUALITY, Theorem 1.2.4, which in turn is very useful throughout
analysis.
Lemma 1.33 −|x| ≤ x ≤ |x|
Proof. If x is positive, it is the right endpoint of the interval [−|x|, |x| ]. If x is negative,
it is the left endpoint. If x = 0, then −|x| = x = |x|. In any case, −|x| ≤ x ≤ |x|.
Exercise 1.29 Provide proofs for Theorems 1.34 - 1.36. State the theorems in terms
of distances when possible. Some of the proofs are provided.
Theorem 1.34 |a| < |b| ⇐⇒ −|b| < a < |b|

Proof. (Distance formulation of theorem: if a is closer to zero than b, then a is in the

interval, [−|b|, |b| ]) Using Lemma 1.33 and order facts from Exercise 1.10, we have
−|b| < −|a| ≤ a ≤ |a| < |b|
Theorem 1.35 |a| ≥ |b| ⇐⇒ a ≤ −|b| or a ≥ |b|
Proof. The statement of this theorem is the contrapositive of Theorem 1.34 and
hence true.
The Triangle Inequality
THE TRIANGLE INEQUALITY and related facts will be used repeated when we discuss limits.
Theorem 1.36 THE TRIANGLE INEQUALITY in four versions.

The first version is the most commonly used in analysis:
1. THE TRIANGLE INEQUALITY For any two real number, x & y ,
|x + y | ≤ |x| + |y |
2. DISTANCE FORM OF THE TRIANGLE INEQUALITY For any three real numbers, x, y & z,
|x − y | ≤ |x − z| + |z − y |
Distance formulation: the length of one side of a triangle is less than or equal
to the sum of the lengths of the other two sides.
3. BACKWARDS TRIANGLE INEQUALITY For any two real number, x & y ,
|x − y | ≥ |x| − |y |
Distance formulation: the distance between two numbers is greater than the
difference between the absolute values of the numbers.
4. For any two real numbers, x & y ,
|x − y | ≥ ||x| − |y ||
Distance formulation: the distance between two points is greater than or equal
to the distance between the absolute values of the numbers.
Proof. The last three versions follow from the first by judicious choice of variables.
1. The following makes use of the fact that |x| + |y | = ||x| + |y || when using
Theorem 1.34.
−|x| ≤ x ≤ |x| by Lemma 1.33

−|y | ≤ y ≤ |y | by Lemma 1.33
−(|x| + |y |) ≤ x+y ≤ |x| + |y | add the two inequalities
|x + y | ≤|x| + |y | by Theorem 1.34
2. Let x = x − z and y = z − y and apply 1. to x + y .
3. EFS
4. EFS
Exercise 1.30 Prove: If a < c < d < b then the distance between c and d is less than
the distance between a and b.
Exercise 1.31 Use THE TRIANGLE INEQUALITY, Theorem 1.36, to prove these two other
versions.
a) |x − y | ≤ |x| + |y |
b) |x + y | ≥ |x| − |y |
Exercise 1.32 Here are some more basic facts about absolute value. Prove them
directly from the definition.
a) | − x| = |x|
b) |y − x| = |x − y |
c) |x 2 | = |x|2
d) If x 2 = c, then |x|2 = c
e) |x · y | = |x| · |y |
f) |x −1 | = |x|−1
True or False 8
Which of the following statements are true? Try stating and graphing each one as a
fact about distances.
a) x < 5 =⇒ |x| < 5

b) |x| < 5 =⇒ x < 5
c) |x − 5| < 2 =⇒ 3 < x < 7
d) |1 + 3x| ≤ 1 =⇒ x ≥ − 23
e) There are no real numbers, x, such that |x − 1| = |x − 2|
f) For every x > 0, there is a y > 0 such that |2x + y | = 5
g) |a − x| < ⇐⇒ x ∈ (a − , a + )
Example 1.5 Using absolute value is often a convenient way to define intervals. Confirm
that
[3, 7] = {x : |x − 5| ≤ 2|}
In general, if a ≤ b, then
d a+b
[a, b] = {x : |x − m| ≤ }, where m = and d = b − a.
2 2
Exercise 1.33 ΥGraph the set determined by each inequality on a number line. Explain
your conclusion.
a) |2x − 4| < 5
b) |2x − 4| ≥ 5
Exercise 1.34 Prove: If x is in the interval (a, b) then the distance between x and the
midpoint of the interval is less than half the length of the interval.
Exercise 1.35 If the distance between two integers is less than 1, the integers are
equal.
1.2.5 Bounded and unbounded sets
Definition We say a set S ⊂ R is bounded whenever there exists a positive real

number, M, such that |s| ≤ M for all s ∈ S.
Definition We say a set S ⊂ R is bounded above whenever there exists a real number
M such that s ≤ M for all s ∈ S.
Definition We say a set S ⊂ R is bounded below whenever there exists a real number
m such that s ≥ m for all s ∈ S.
Both of the following theorems describe common techniques used in proofs about
boundedness.
Theorem 1.37 A set S ⊂ R is bounded if and only if it is bounded above and bounded
below.
Proof. EFS
Theorem 1.38 A S ⊂ R is bounded above if and only if the set,
−S = {x ∈ R : −x ∈ S},
is bounded below.
Proof. EFS
Theorem 1.39 The real numbers are not bounded
Proof. hint: State the negation of the definition of boundedness.
Exercise 1.36 Υ Give three examples of sets that are bounded and three examples of
sets that are not bounded. Which of your assertions can you prove to be true?
Exercise 1.37 Υ Give three examples of sets that are bounded above but not below.
Which of your assertions can you prove to be true?
Definition We say that L is a least upper bound of a set S ⊂ R whenever both of

the following conditions hold:
1. L is an upper bound of S
2. if u is an upper bound of S, then L ≤ u.
If L is a least upper bound of a set S and if L ∈ S, we call L the maximum of S.
Definition Write out a definition for greatest lower bound, G, and minimum of S.
Theorem 1.40 Any two least upper bounds for a non-empty set, S, are equal. Any
two greatest lower bounds for a non-empty set, S, are equal.
Proof. Let L1 and L2 both be least upper bounds for S. Without loss of generality,
assume that L1 ≤ L2 Since L2 is a least upper bound, it must be less than or equal
to L1 , which is a upper bound of S. Hence, L2 ≤ L1 ≤ L2 . By trichotomy, L1 = L2 .
A similar proof works for the greatest lower bounds.
Notation As usual uniqueness allows us to name the least upper bound and the
greatest lower bound of a set, should they exist. We use the abbreviation sup S
and say, supremum of S, for the least upper bound of set S. Similarly we use the
abbreviation inf S and say, infinum of S, for the greatest lower bound of S. In the
case when sup S ∈ S, we also call it the maximum of S. If inf S ∈ S, we call it the
minimum of S
Example 1.6 The greatest lower bound of the open interval, (5, 10), is 5
Proof. There are two steps in the proof of this fact:
1. 5 is a lower bound: By definition of the open interval, 5 < x for all x ∈ (5, 10).
2. 5 is greater than any other lower bound: Suppose h is a lower bound greater
than 5, so 10 > h > 5. Consider m = 5+h 2 , the average of 5 and h. We know
5 < m < h < 10. Since m ∈ (5, 10), h is not a lower bound for (5, 10). Since
any number greater than 5 is not a lower bound, 5 must be the greatest one.
Exercise 1.38 For any two real numbers, a < b, the least upper bound of the interval
(a, b) is b.
Exercise 1.39 Let N be an integer and let S = {s ∈ R : s 2 ≤ N}. Find a rational

number that is a upper bound of S. Prove your assertion.
Definition We say that a set is finite whenever there exists a 1 − 1 correspondence

between the set and the set of all positive integers less than or equal to n, for some
positive integer n. The order of the set, or the number of elements in the set, is n.
[A 1 − 1 correspondence is a bijection.]
Theorem 1.41 A finite set has a maximum and a minimum element.
Proof. HINT: Use induction on the number of elements in the set.
Example 1.7 Υ What is the inf{ n1 : n ∈ Z+}?

Example 1.8 Υ Does Z+ have an upper bound?
Definition We say a function is bounded (bounded above) (bounded below) when-
ever the image of the function is bounded (bounded above) (bounded below). The
definition includes the possibility that the function is a sequence.
Exercise 1.40 Give three examples of functions that are bounded.
Exercise 1.41 Give three examples of functions that are not bounded.
1.3 The Completeness Axiom
True or False 9
a) There exists a least positive real number.

1
b) For all positive numbers , there exists a positive integer, N, such that < .
N
The answers to the last question is ’Yes’ (see Theorem 1.48), but we can’t prove it
yet. We need another axiom:
Axiom 9 THE COMPLETENESS AXIOM A non-empty set of real numbers that is bounded
above has a least upper bound.
1.3.1 Consequences of the Completeness Axiom
Clearly, there is an analogous fact for lower bounds, but it need not be stated as part
of the axiom. Instead it can be proved from the axiom. The technique is a standard
good trick to know.
Theorem 1.42 EXISTENCE OF GREATEST LOWER BOUND A non-empty set of real numbers
that is bounded below has a greatest lower bound.
Outline of Proof: If S is a non-empty set that is bounded below, let T = {x : −x ∈ S}.

Show that T is bounded above, apply Axiom 9, and make conclusions about the
original set S. Draw number line pictures to help explain the strategy of the proof.
A word about The greek letter, ’epsilon,’ , is often used in situations where the
interesting part is numbers getting arbitrary small. What we mean by arbitrarily small
is that the inf of the set of positive ’s we are considering is 0. We use to stand in
for ’error,’ which we like to be small.
Theorem 1.43 Let S be a non-empty set of real numbers that is bounded above.
For all > 0, there exists x ∈ S such that sup S − < x.
NOTE: This is what we mean when we say that there are numbers in S get arbitrarily
close to sup S.
1.3. THE COMPLETENESS AXIOM 37
Outline of Proof. If > 0, sup S − cannot be a upper bound. Draw a numberline

picture to help explain the situation.
Exercise 1.42 State and prove the analogous theorem for the greatest lower bound of
a set that is bounded below.
Theorem 1.44 Let S be a non-empty set of real numbers that is bounded above. Let
U be the set of all upper bounds for S, that is,
U = {u ∈ R : u ≥ s for all s ∈ S},
then inf U = sup S.
Outline of Proof. Any u ∈ U is an upper bound for S, so u ≥ sup S, the least upper
bound. So sup S is a lower bound for U. Any number greater than sup S, is an upper
bound for S and hence ∈ U. So sup S is the greatest lower bound. Draw a numberline
picture to help explain the situation.
Exercise 1.43 State and prove an analogous theorem for a set of real numbers that is
bounded below.
The following theorem is a forerunner to THE NESTED INTERVAL THEOREM which we will
be using extensively fro the rest of the course.
Theorem 1.45 Given two non-empty subsets, A ⊂ R and B ⊂ R, such that every
element in A is a lower bound for B and every element in B is an upper bound for A,
there exists a real number between the two sets. That is, there exists a real number,
r , such that for all a ∈ A and b ∈ B, a ≤ r ≤ b In fact,
sup A ≤ r ≤ inf B
Outline of Proof: Apply Axiom 9 to argue that A has a least upper bound and that
B has a greatest lower bound. Show that sup A is a lower bound for B, and hence
that sup A ≤ inf B. Then r could be any number in between the two. Draw a number
line picture to illustrate the proof.
Exercise 1.44 Prove: If S ⊂ R≥ and if S is bounded below, then inf{x 2 : x ∈ S} =

(inf S)2 .
1.3.2 The Nested Interval Theorem

Theorem 1.46 THE NESTED INTERVAL THEOREM The intersection of a sequence of non-
empty, closed, nested intervals is not empty. Furthermore, if the least upper bound of
the left endpoints is equal to the greatest lower bound of the right endpoints, there is
only one point in the intersection.
NOTE: Notate the sequence by In = [an , bn ], for n > 0. Convince yourself of the
following and draw a numberline picture to help explain the theorem.
1. That each interval, In , is not empty is equivalent to saying that for all n, an ≤ bn ,
2. That each interval is closed means that the endpoints are contained in the
interval and that’s why we used the closed brackets to denote the intervals.
3. That the sequence is nested means that, for all n > 0, [an+1 , bn+1 ] ⊂ [an , bn ] or
that an ≤ an+1 and that bn+1 ≤ bn .
Proof of The Nested Interval Theorem 1.46. Let A = {an : n > 0} and let B = {bn :
n > 0}. Since the intervals are non-empty and nested, an ≤ bm for all positive integers
n and m. Theorem 1.45 applies to the sets, A and B, so there exists a real number,
r , such that an ≤ r ≤ bn for all n. Since r is in all the intervals, it is also in the
intersection. And since it is also true that sup an ≤ r ≤ inf bn , if sup an = inf bn , then
any point in the intersection must be equal to both.
1.3.3 Archimedes Principle
The integers are not bounded
Theorem 1.47 The set of positive integers is not bounded above.
Proof. We will prove this theorem by contradiction: assume the set of positive integers
is bounded. By Axiom 9 there would be a least upper bound. Let L be this least upper
bound. Then L − 1, being less than L, is not an upper bound for the positive integers.
Let N be a positive integer greater than L − 1. We have L − 1 ≤ N =⇒ L ≤ N + 1.
Since L is an upper bound for the set of positive integers and N + 1 is a positive
integer, we also have that L ≥ N + 1. Together this means that L = N + 1. But then
N + 2 is an integer great than L so L couldn’t be a upper bound.
Exercise 1.45 Use Theorem 1.47 to show that the integers are not bounded.
Theorem 1.48 ARCHIMEDES PRINCIPLE For all real numbers > 0, there exists a positive
1
integer N such that < .
N
Proof. The negation of this statement is that there exists a positive real number ,
1 1
such that for all positive integers, N, ≥ . But this says that N ≤ for all N, or
N
1
that the integers are bounded by . This is false so the Archimedes Principle must be

true.
This easy restatement of Archimedes Principle is the first of many squeeze theorems.
1
Theorem 1.49 SQUEEZE THEOREM 1 If 0 ≤ h < n for all n ∈ Z+ , then h = 0.
Proof. Assume h satisfies the hypotheses but is not 0. By ARCHIMEDES PRINCIPLE,

Theorem 1.48 , there exist n ∈ Z+ such that n1 < h.
Theorem 1.50 GENERALIZED ARCHIMEDES PRINCIPLE For all real numbers x and d > 0,
there exists a positive integer N such that x < N · d.
Proof. Use Archimedes Principle to prove this theorem. Note that Archimedes Prin-
ciple, in turn, follows from this theorem.
Exercise 1.46 Show that Archimedes Principle holds for rational numbers. That is,
prove the following theorem without using the Archimedes Principle.
Theorem 1.51 For all rational numbers, r > 0, there exists a positive integer, N, such
that 0 < N1 < r.
Proof.
Exercise 1.47 The following theorem follows from Archimedes Principle using Exercise
1.21. We use it in the next (optional) section and in later work. Prove it.
B
Theorem 1.52 For all B, h > 0, there exists an integer n > 0 such that <h
2n
Proof. EFS
1.3.4 Optional – Nested Interval Theorem and Archimedes prove the

Completeness Axiom
Theorem 1.53 The Nested Interval Theorem and Archimedes Principle imply the
Completeness Axiom
Proof. Let S be a non-empty set with an upper bound. Construct a sequence of

nested closed intervals, [an , bn ], by bisection, so that [an+1 , bn+1 ] ⊂ [an , bn ], an ∈ S,
bn an upper bound for S, and |bn − an | ≤ b02−a n
0
.
• Let a0 be a point in S (S is non-empty!) and let b0 be a upper bound for S If

a0 = b0 this is the least upper bound and we are done, so assume a0 < b0 .
• Assume [an , bn ] is defined as required. Let m be the mid-point of [an , bn ]. There
are two cases:
1. If m is an upper bound for S, let an+1 = an and bn+1 = m.
2. If m is not an upper bound for S then there exists an+1 ∈ S that is greater
than or equal to m. Note that an+1 ≤ bn because bn is an upper bound for
S. Let bn+1 = bn .
If an+1 = bn+1 this is the least upper bound and we are done, so assume an+1 <
bn+1 . Notice that [an+1 , bn+1 ] ⊂ [an , bn ] and |bn+1 −an+1 | ≤ 21 |bn −an | ≤ 21 b02−a
n
0
=
b0 −a0
2n+1 , so the new interval satisfies the requirements.
By THE NESTED INTERVAL THEOREM, all these intervals contain a common point, b. By
ARCHIMEDES PRINCIPLE, b is the only common point: For, if a is another one then 0 <
|b − a| < b02−a
n
0
< n1 for all n, this implies b = a by SQUEEZE THEOREM 1.
Claim: b is an upper bound for S. Proof: Suppose not. then there is some a ∈ S
with a > b. By ARCHIMEDES PRINCIPLE, in the form of Theorem 1.52, for some n,
|a − b| > |b02−a
n
0|
= |bn − an | so (removing parentheses) a − b > bn − an > bn − b
(an < b) or a > bn . This shows that a must be an upper bound. a must be the least
upper bound since it is in S. So a ≤ bn all n. a is in all the intervals so a = b.
Supposes b is not the least upper bound, there there exists an upper bound for S, a,
with a < b ≤ bn , for all n. Since a is an upper bound for S, an ≤ a for all n so a is in
all the intervals. a = b again.
So b is the least upper bound for S.
1.3.5 Rational numbers are dense in R
Definition We say that a subset D ⊂ R is dense in R, whenever every open interval

of R contains an element of D.
Exercise 1.48 The following sequence of theorems can be used to prove that the
rational numbers are dense in R. Prove Theorems 1.55 - 1.59.
Theorem 1.54 Any non-empty set of integers that is bounded below has a minimum
number.
Proof. EFS
Theorem 1.55 For every real number, r , there exists a unique integer, n, such that
r < n ≤ r + 1.
Proof. HINT: For r > 0, use well-ordering to find smallest positive integer greater
than or equal to r . This will be n.
Theorem 1.56 For every real number, r , there exists a unique integer, m, such that
m ≤ r < m + 1.
Proof. EFS
Theorem 1.57 If b > a + 1, then the open interval, (a, b) contains an integers.
Proof. EFS
Theorem 1.58 Q IS DENSE IN R Every non-empty open interval contains a rational

number.
Proof. EFS
Theorem 1.59 Given any real number, α, there exists a rational number arbitrarily
close to α. That is, given > 0, there exists a rational number, r , such that |α−r | <
Proof. Let r be a rational number in the interval, (α − , α + ).
1.3.6 Optional – An Alternative Definition of Interval
This section is a mini-lesson on how making good, mathematical definitions can sim-
plify understanding and proving. The problem with our current definition of interval
is that there are too many parts to it. We have to worry about open and closed
endpoints as well as unbounded intervals. A simplified definition may make it easier
to prove things about intervals. The following two theorems exploit a condition that
is simple to use. Reminder: our current definition says that an interval is one of the
following:
(a, b) or [a, b] or (a, b] or [a, b) or (a, ∞) or [a, ∞) or (∞, b) or (∞, b]
Exercise 1.49 Write all cases for the proof of the following Theorem, 1.60. The
proof is straight forward but tedious because each type of interval must be dealt with
separately.
Theorem 1.60 If S is an interval, then
a, b ∈ S and a < c < b =⇒ c ∈ S. (1.1)
Proof. EFS
What if we used the condition 1.1 from the theorem as the definition of interval?
Would we get the same sets are intervals?
Theorem 1.61 If a set S ⊂ R satisfies condition 1.1, then S is an interval.

Proof. First, suppose S is bounded. Let z be any point strictly between a = inf S
and b = sup S. There is a point, x ∈ S greater than z (otherwise z would be a upper
bound, less than b, the least upper bound) and a point y ∈ S less than z (otherwise z
would be a lower bound, greater than a, the greatest lower bound). By the condition,
z ∈ S. Now (a, b) ⊂ S ⊂ [a, b], so S is an interval by our previous definition.
Second, suppose that S is unbounded above, but not below. Let z be any point
greater than a = inf S. There is a point, x ∈ S greater than z (otherwise z would be
a upper bound for S.) and a point y ∈ S less than z (otherwise z would be a lower
bound, greater than a, the greatest lower bound). By the condition, z ∈ S. Now
(a, ∞) ⊂ S ⊂ [a, ∞) so S is an interval by our previous definition.
The other two cases are similar.
Notice that this proof depends on the completeness axiom. In the rational numbers,
there are sets that satisfy condition (1.1) but are not intervals in our original definition.
Exercise 1.50 Show that {r ∈ Q : r 2 > 2} satisfies condition (1.1), but it cannot be
written as (a, +∞) for any a ∈ Q.
Chapter 2
Limits of Sequences
Calculus Student: lim sn = 0 means the sn are getting closer and closer to zero but
n→∞
never gets there.
Instructor: ARGHHHHH!
Exercise 2.1 Think of a better response for the instructor. In particular, provide a
counterexample: find a sequence of numbers that ’are getting closer and closer to
zero’ but aren’t really getting close at all. What about the ’never gets there’ part?
Should it be necessary that sequence values are never equal to its limit?
2.1 Definition and examples
We are going to discuss what it means for a sequence to converge in three stages:
First, we define what it means for a sequence to converge to zero
Then we define what it means for sequence to converge to an arbitrary real number.
Finally, we discuss the various ways a sequence may diverge (not converge).
In between we will apply what we learn to further our understanding of real numbers
and to develop tools that are useful for proving the important theorems of Calculus.
43
44 CHAPTER 2. LIMITS OF SEQUENCES
1
Figure 2.1: sn = .
n
2
0
0 5 10 15 20
2.1.1 Sequences converging to zero.
Definition We say that the sequence sn converges to 0 whenever the following hold:
For all > 0, there exists a real number, N, such that
n > N =⇒ |sn | < .
Notation To state that sn converges to 0 we write lim sn = 0 or sn → 0.

n→∞
1
Example 2.1 lim = 0. See the graph in Figure 2.1.
n→∞ n
Proof. Given any > 0, use Archimedes Principle, Theorem 1.48, to find an N, such
that N1 < . Note that, if n > N, then n1 < N1 (Exercise 1.10 d)). Now, if n > N, we
have
1 1
|sn | = < < .
n N
In short:
n > N =⇒ |sn | < ,
1
so we have shown that lim = 0.
n→∞ n
Example 2.2 If sn = 0, for all n, then lim sn = 0

n→∞
Proof. Given any > 0, let N be any number. Then we have
n > N =⇒ |sn | = 0 < ,
because that’s true for any n.

2.1. DEFINITION AND EXAMPLES 45
Figure 2.2: Some values approach 0, but others don’t.

0.5
0.0
0 5 10 15 20
Example 2.3 Why isn’t the following a good definition?
” lim sn = 0 means
n→∞
For all > 0, there exists a real number, N, such that |sN | < .”
The problem is we want the sequence to get arbitrarily close to zero and to stay there.
Consider the sequence: (
1
, if n is odd
sn = n
0, otherwise.
There is always an odd N to be less than any but there there are many even n’s with
values far from zero. The n > N is an important part of the definition. See the graph
in Figure 2.2.
1
Exercise 2.2 Sketch the sequence and prove that lim =0
n→∞ n 2
(−1)n
Exercise 2.3 Sketch the sequence and prove that lim = 0 See Figure 2.3.
n→∞ n
1
Exercise 2.4 Sketch the sequence and prove that lim = 0.
n→∞ n(n − 1)
Exercise 2.5 Prove the following theorem:
Theorem 2.1 sn → 0 ⇐⇒ |sn | → 0 ⇐⇒ −sn → 0
Proof. EFS
The above are good exercises but problems like these will be easier to prove – that is,
no epsilons will be needed – once we have some theorems. First, a useful fact:
Theorem 2.2 If lim an = 0, then the sequence, an , is bounded. That is, there exists
n→∞
a real number, M > 0 such that |an | < M for all n.
(−1)n
Figure 2.3: Picking N for smaller and smaller for the sequence sn = n .
1
1
0 5 10 15 20
1
1
0 5 10 15 20
1
1
0 5 10 15 20
Proof. Since an → 0, there exists N ∈ Z+ such that n > N =⇒ |an | < 1. Here we
use the definition of converging to 0 with = 1. (NOTE: We could use any positive
number in place of 1.) Let B be a bound for the finite set {an : n ≤ N}. This set
is bounded by Theorem 1.41. Let M = max{B, 1} Hence any an is bounded by M
because it is either in the finite set (n ≤ N ) and bounded by B or it is bounded by 1,
because n > N.
Theorem 2.3 ALGEBRAIC PROPERTIES OF LIMITS 1
Given three sequences, lim an = 0, lim bn = 0 and a real number, c, then:

n→∞ n→∞
1. lim an + bn = 0
n→∞
2. lim c · an = 0.
n→∞
3. lim an · bn = 0.
n→∞
Proof. EFS
Exercise 2.6 Prove: If cn is bounded and an → 0, then cn · an → 0
Theorem 2.4 SQUEEZE THEOREM If an → 0 and bn → 0 and an ≤ cn ≤ bn , for all n ∈ Z+ ,

then lim cn = 0.
n→∞
Proof. SQUEEZE THEOREM Given > 0, let N be large enough so that whenever n > N,
then both |bn | < and |an | < . Now, for any n > N, if cn > 0, we have |cn | ≤ |bn | < .
or if cn < 0, then |cn | = −cn ≤ −an = |an | < . So, for all n > N we have |cn | < .
We have shown that cn → 0.
n2 + n
Exercise 2.7 Prove that lim →0
n→∞ n3
1
Exercise 2.8 Prove: For any positive integer, m, lim =0
n→∞ n m
Exercise 2.9 Use BERNOULLI’S INEQUALITY Theorem 1.27 to prove the following theorem.
Theorem 2.5 If 0 ≤ r < 1, then r n → 0
Proof.
2.1.2 Sequences that converge to arbitrary limit

Definition We say that sn converges whenever there exists a real number, s, such
that |s − sn | → 0. In this case, we say that sn converges to s, and write
lim sn = s or sn → s
n→∞
n−1 n−1 1 1
Example 2.4 lim = 1, because 1 − = 1 − (1 − ) = → 0
n→∞ n n n n
Exercise 2.10 Show that sn → 0 means the same thing for both definitions: converging
to 0 and converging to an arbitrary limit that happens to be 0.
Theorem 2.6 UNIQUENESS OF LIMIT If an → a and an → b, then a = b.
Proof. Use the triangle inequality to see that 0 ≤ |a − b| = |a − an + an − b| ≤

|a − an | + |an − b|. Apply THE SQUEEZE THEOREM (Theorem 2.3 part 2.4.): the left-most
term is the constant sequence, 0, the right-most term is the sum of two sequences
that converge to 0, so also converges to 0. Hence the middle term also converges to
0. So |a − b| = 0 =⇒ a = b.
Theorem 2.7 If lim an = a, then the sequence, an , is bounded.

n→∞
Proof. EFS Consider using Theorem 2.2.
Exercise 2.11 Prove: If an = c, for all n, then lim an = c

n→∞
Theorem 2.8 ALGEBRAIC PROPERTIES OF LIMITS 2
Given two sequences, lim an = a and lim bn = b, then:

n→∞ n→∞
1. lim an + bn = a + b
n→∞
2. lim an · bn = a · b
n→∞
1 1
3. If an , a 6= 0, then lim =
n→∞ an a
Proofs. 1. EFS
2. EFS
3. To prove this one, we start by showing that the sequence an is bounded away
from 0, meaning that there exists a positive number B, such that |an | > B,
for all n. To find such a bound, B, first note that there is N > 0 such that
|an − a| < | 2a | for all positive integers n > N. (Using = | 2a | in the definition of
limit.) For those n,
a a
|an | ≥ |a| − |an − a| > |a| − | | = | |.
2 2
Now let B = min {|an | : n ≤ N}. This set has a minimum value because it is a
finite set. B > 0 because none of the an = 0. Finally, let B = min {B, | 2a |}. So
|an | > B, for all n.
Now, given > 0, find N such |an − a| < · B · |a| then
1 1 |a − an | |an − a| · B · |a|
| − |= < < =
an a |an · a| B · |a| B · |a|
An important concern of calculus is what happens to the ratio of two

LIMITS OF RATIOS
limits when both the numerator and denominator converge to 0. If the denominator
converges to zero, but the numerator is bounded away from zero, then the ratio will
be unbounded and not converge.
Exercise 2.12 Give examples of two sequences, an → 0 and bn → 0, such that
an
a. →0
bn
an
b. → c, where c is a positive real number.
bn
an
c. does not converge.
bn
Theorem 2.9 ORDER PROPERTIES OF LIMITS
For real sequences, an , bn , cn and real numbers, a and c.
1. If an > c for all n ∈ Z+ and an → a, then a ≥ c
2. If an ≤ c ≤ bn for all n and |an − bn | → 0, then an → c and bn → c.
3. THE SQUEEZE THEOREM If an → c and bn → c and an ≤ cn ≤ bn for all n ∈ Z+ ,

then cn → c
Proof. Confirm each statement and explain how it proves the corresponding part of
the theorem
1. Suppose c − a > 0, find N such that
n > N =⇒ |a − an | < c − a
2. For all n, 0 ≤ |an − c| ≤ |bn − an |
3. 0 ≤ |c − cn | ≤ |c − an | + |an − cn | ≤ |c − an | + |an − bn |
√
5n + 3
Figure 2.4: sn = √
n + 10
2
0
0 5 10 15 20
Theorem 2.10 THE TAILEND THEOREM If an → a and if bn = an+m for some fixed,
positive integer m, then
bn → a.
Proof. A tailend of a sequence is a special case of a subsequence, see Section 2.1.4.
Exercise 2.13 Prove: If an → c and bn → c, then |an − bn | → 0
Exercise 2.14 Conjecture what the limit might be and prove your result.
3n2 + 2n + 1
sn =
n2 + 1
Exercise 2.15 Conjecture what the limit might be and prove your result.
√
5n + 3
sn = √
n + 10
Exercise 2.16 Prove that, if c 6= 0, then

a·n+b a
lim = .
n→∞ c · n + d c
Exercise 2.17 Prove: If an → a, bn → b, and an < bn for all n, then a ≤ b.
Exercise 2.18 Prove: If S is a bounded set, then there exists a sequence of points,
sn ∈ S such that sn → sup S.
Exercise 2.19 Prove: If an → a then an2 → a2
Monotone sequences
Definition We say a sequence is monotone whenever it is an increasing sequence or

a decreasing sequence.
Theorem 2.11 MONOTONE CONVERGENCE THEOREM Every bounded, monotone sequence

converges to a real number.
Proof. Let sn be a bounded, increasing sequence. Let s = sup{sn } which exists

because sn is bounded above. We claim that sn → s. Given > 0, use Theorem 1.43
to find x = sN ∈ {sn } such that s − < sN . Now if n > N, we know sn > sN because
the sequence is increasing, so
|sn − s| = s − sn < s − sN < .
We conclude that sn → s. If tn is a bounded, decreasing sequence, then sn = −tn is
bounded and increasing. Since sn → s, for some s, we known that tn = −s.
Best Nested Interval Theorem
Theorem 2.12 BEST NESTED INTERVAL THEOREM There exists one and only one real num-
ber, x, in the intersection of a sequence of non-empty, closed, nested intervals if the
lengths of the intervals converge to 0. Furthermore, the sequence of right endpoints
and the sequence of left endpoints both converge to x.
Proof. Denote the intervals by [an , bn ]. Because they are nested we know that an is
increasing and bn is decreasing so, by MONOTONE CONVERGENCE THEOREM, there are real
numbers a and b such that an → a and bn → b. Let c be any real number in the
intersection of all the intervals. Then an ≤ c ≤ bn and since |bn − an | → 0 we have
by the ORDER PROPERTIES OF LIMITS (Theorem 2.9 2.), c = a = b.
Rational Approximations to Real Numbers
We have not yet shown that there are real numbers other than rational numbers.
However, if there is one, the following method indicates that you can approximate it
by rational numbers; that is, there is a sequence of rational numbers that converge it.
The method of bisection used here is a well-used tool of analysis.
Example 2.5 Let r be any non-rational real number. Find a sequence of rational
numbers that converge to r.
Using the method of bisection. There are other ways to show the existence of such a
sequence. The advantage to this method is that it gives a way to construct approxi-
mations of the given real number.
First, find rational numbers a0 and b0 such that a0 < r < b0 . (By Theorem 1.56,
they can both be integers.) Recursively define a sequence of non-empty, closed,
nested intervals [an , bn ] such that each interval contains r and is half the length of the
preceding interval. We already have the base case, [a0 , b0 ]. Assume [an , bn ] has been
defined as required. Let m be the midpoint of [an , bn ]. Since m is a rational number,
m 6= r, so there are two cases to consider:
1. If m < r, let an+1 = m and bn+1 = bn .
2. If r < m, let an+1 = an and bn+1 = m.
in either case r ∈ [an+1 , bn+1 ] ⊂ [an , bn ]. The length of this interval is half the length
of the previous interval. The lengths of the intervals converge to 0 so, by the BEST
NESTED INTERVAL THEOREM (Theorem 2.12), r is the only real number in all the intervals
and both an → r and bn → r.
2.1.3 Application: Existence of square roots

The proof given here of the existence of square roots is by construction. There are
other ways to prove the existence of square roots √
– the advantage to this method is
that it gives a way to calculate approximations to 2.
Example 2.6 There exists a unique positive real number, s, such that s 2 = 2.
Proof by bisection. We will show that there is a nested sequence of closed intervals,
In = [an , bn ], such that an 2 ≤ 2 ≤ bn 2 and |bn − an | = 21n . By the BEST NESTED INTERVAL
2
THEOREM, there is a unique number, s, in all of the intervals. We will show that s = 2
using a squeezing argument.
We define the intervals inductively: Base case: Let a0 = 1 and b0 = 2, so
1
a0 2 = 1 ≤ 2 ≤ 4 = b0 2 , and |b0 − a0 | = 1 = .
20
Assume an and bn have been defined as desired
1
an 2 ≤ 2 ≤ bn 2 , and |bn − an | = .
2n
Proceed inductively to define the next interval, In+1 = [an+1 , bn+1 ]. Let m be the
midpoint of the interval [an , bn ]. We know that m2 6= 2 because we know the square
root can not be a rational number. (Why do we claim m is rational?) So there are
only two cases to consider:
1. If m2 < 2, let an+1 = m and bn+1 = bn .
2. If 2 < m2 , let an+1 = an and bn+1 = m.
Since m is the midpoint of the previous interval the length of In+1 is half the length
of In , so |bn+1 − an+1 | = 12 · 21n = 2n+1
1
Let s be the unique number in all the intervals. We know an → s and bn → s. So

we also know that an2 → s 2 , bn2 → s 2 . (See Exercise 2.19) Now consider the image,
under the squaring function of all those intervals. The intervals [an2 , bn2 ] are closed,
non-empty (an2 < bn2 because an < bn ); and nested (an2 is increasing because an is
increasing, and bn2 is decreasing because bn is decreasing). These all follow because
the function x → x 2 is increasing on positive numbers, see Exercise 1.12. Furthermore,
|bn2 − an2 | → 0. The BEST NESTED INTERVAL THEOREM applies and we conclude there is a
unique real number in all the intervals and that the endpoints converge to that number.
2 is all the intervals and an → s 2 , so s 2 = 2, by uniqueness of limits.
Theorem 2.13 For all a > 0, there exists a unique positive real number, s, such that
s 2 = a.
Proof. HINT: What in the proof for Example 2.6 depends on the choice a = 2?
2.1.4 Subsequences
Definition We call sequence, snk , whose values are a subset of the values of sn , a
subsequence of sn , whenever the sequence nk is strictly increasing. (We will assume
that k is indexed on Z≥ , i.e. n0 is the first value.)
Example 2.7 If nk = 2k, then the subsequence is every other element, starting at 0,
of the sequence.
Exercise 2.20 Prove: If snk is a subsequence of sn , then nk ≥ k. Note that this is true
for any increasing sequence of positive integers nk .
Theorem 2.14 If sn → s, then any subsequence of sn also converges to s.
Proof. Let snk be a subsequence of sn . Given > 0, find N so that n > N =⇒

|sn − s| < . Such N exists because sn → s. Now consider the subsequence: If k > N,
then nk > N (by Exercise 2.20), so |snk − s| < .
Example 2.8 The sequence is Example 2.21 has many subsequences, each of which
converges to one of five different numbers. For example, s5k+2 → sin 4π
5 .
2.1.5 Divergent Sequences
Definition A sequence is said to diverge if there is no real number L such that the
sequence converges to L.
To show that a sequence, sn , converges we would first conjecture a possible limit,

L, and then prove sn → L. To show that the sequence does not converge is perhaps
harder because we have to show it doesn’t converge for all possible values L. And we
would need to prove the negation of the statement, sn → L, for all values of L. Here
are both statements:
Figure 2.5: sn = sin( 2πn

5 ).
1
1
0 5 10 15 20 25 30 35 40 45
sn → L, means
For all > 0, there exists N > 0, such that

n > N =⇒ |sn − L| < .
sn 6→ L, means
There exists an > 0, such that for all N > 0

there is an n > N with |sn − L| ≥ .
Fortunately there is an easier way to show that a sequence diverges by observing

subsequence behavior, using Theorem 2.14.
Example 2.9 The sequence,

(
1, n, odd
sn =
−1, n, even
diverges, i.e. does not converge to any real number L.
Proof. The subsequence given by nk = 2k + 1 is a constant sequence: snk = 1 for all

k. This subsequence converges to 1. The subsequence given by nk = 2k converges
to −1. If the sequence converged, both subsequences would have to converge to the
same number, by Theorem 2.14. So the sequence does not converge.
Exercise 2.21 Υ Explain why the following sequence, sn = sin( 2πn)

5 , diverges. The
graph of this sequence is shown in Figure 2.5.
Example 2.10 AnotherΥexample of a divergent sequence is sn = sin(n). That this

sequence does not converge seems a correct conclusion considering the graph, shown
in Figure 2.6. Proving that goes beyond the scope of our present discussion.
Another way a sequence may fail to converge is if it is unbounded. We consider

separately the case when the limit appears to be infinite as in the sequence, sn = n.
Figure 2.6: sn = sin(n).

1
1
0 5 10 15 20 25 30 35 40 45
Definition We say that sn diverges to infinity and we write, lim sn = +∞, whenever,
n→∞
for all M > 0, there exists N > 0 such that
n > N =⇒ sn ≥ M.
Example 2.11 lim n2 = +∞: If M is any positive real number, let N = M. Then, if
n→∞
n > N, we have that n2 > n > N = M or simply that n2 > M. (EFS: Explain the
step n2 > n)
Exercise 2.22 Give a sequence that is unbounded but does not diverge to +∞
Theorem 2.15 If sn → +∞, then any subsequence of sn also diverges to +∞.
Exercise 2.23 Prove the following theorem.
Theorem 2.16 ALGEBRAIC PROPERTIES OF DIVERGENT LIMITS
1. an → +∞ and bn → b or bn → +∞ =⇒ (an + bn ) → +∞.
2. an → +∞ and bn → +∞ =⇒ (an · bn ) → +∞.

1
3. an → +∞ =⇒ →0
an
Exercise 2.24 Use BERNOULLI’S INEQUALITY Theorem 1.27, to prove the following theo-
rem.
Theorem 2.17 If r ≥ 1, then r n → +∞
Proof.
Exercise 2.25 LIMITS OF RATIOS Give examples of two sequences, an → +∞ and bn →

+∞, such that
an
a. → +∞
bn
an
b. →0
bn
an
c. → c, where c is a positive real number.
bn
Exercise 2.26 Provide a definition and theorems about diverging to −∞
a. What would it mean for the limit of a sequence to be −∞?
b. If it is not part of your definition prove: lim sn = −∞ ⇐⇒ lim −sn = +∞?

n→∞ n→∞
c. If it is not part of your definition prove: If lim sn = −∞ then,

n→∞
for all M < 0, there exists N > 0 such that
n > N =⇒ sn ≤ M.
d. Give an example of a sequence that diverges to −∞ and prove that it does.
e. State and prove statements of ALGEBRAIC PROPERTIES OF DIVERGENT LIMITS for se-
quences that diverge to −∞
f. If sn → −∞, then any subsequence of sn also diverges to −∞.
1
Exercise 2.27 Give a sequence sn → 0 where does not diverge to either +∞ or
sn
−∞.
Theorem 2.18 If p is a polynomial that is not constant, then either
lim p(n) = +∞ or lim p(n) = −∞

n→∞ n→∞
Proof. EFS
2.2 Limits and Sets
2.2.1 Limit Points and Boundary Points
editor: Handle references to numbers as points and as elements of a set. BE CON-

SISTENT. editor: consider defining neighborhoods
2.2. LIMITS AND SETS 57
We begin with a definitions of different types of special points of sets that are useful
to think about: limit points and boundary points.
Definition We say the a point, p, is a limit point of a set, S, whenever every open
interval about p contains an infinite number of points in S. In particular, it contains
a point in S that is not equal to p.
NOTE: The point p need not be in S.
Definition We say the a point, p, is a limit point of a sequence, sn , whenever every

open interval about p contains an infinite number of sn .
editor: Think a bit about the difference between these two definitions. In particular,
the infinite number in the first case refers to numbers in the set,S, whereas, in the
second case, infinite number refers to numbers in the domain of sn . That, for
sequences, the points can all be p. Whereas for sets all but at most one of the points
are not p.
Example 2.12 The limit points of the image of sn may be different than the limit points
of sn . Consider sn = (−1)n . The image of sn is {, −1, 1}, a set that has no limit points.
However, both −1 and 1 are limit points of the sequence because they each appear
an infinite number of times in the sequence.
Theorem 2.19 A real number p is a limit point of a set S, if and only if there exists
sequence of points in S \ {p} that converge to p.
Proof. =⇒ Suppose p is a limit point of S. For each n ∈ Z+ , let

1 1
sn ∈ S ∩ (p − , p + ) \ {p}.
n n
Such a point exists because p is a limit point. We claim that sn → p. For we know,
for all n,
1
0 < |p − sn | <
n
From the squeeze theorem, we conclude that |p − sn | → 0 or sn → p.
⇐=Let sn ∈ S \ {p} converge to p. Let (a, b) be an open interval containing p.
Consider any positive < min(b − p, p − a), so that p ∈ (p − e, p + ) ⊂ (a, b). By the
convergences of sn , there exists N such that n > N =⇒ |p − sn | < . These sn ’s are
an infinite number of values of the sequences that are in (a, b) \ {p}. If there were only
a finite number of numbers from S in this sequence, we would have a subsequence of
sn that converges to some other number which cannot happen. So the sn for n > N
are an infinite number of points in (a, b), as required to show that p is a limit point
of the set S.
Theorem 2.20 If s is a limit point of a sequence sn , then there exists a subsequence

of sn that converge to s.
Proof. EFS
Definition We say the a point, p, is an boundary point of a set, S, whenever every

open interval containing p contains points in both S and R \ S.
Example 2.13 Every point in a finite set is a boundary point of the set. Every point in
a finite set is a boundary point of the complement of set.
True or False 10
Which of the following statements are true? If false, modify the statement to be true.
Explain.
a) The endpoints of an interval are boundary points of the interval.
b) Every point in an interval is a boundary point of the interval.
c) Every point of an interval is a limit point of the interval.
d) The inf S is a limit point of S.
e) The inf S is a boundary point of S.
f) The maximum value of a S is a boundary point of S.
Example 2.14 Give an example of each of the following and explain.
a) A set and a point that is a boundary point but not a limit point of the set.
b) A set and a point that is a limit point but not a boundary point.
c) A set and a point that is neither a limit point nor a boundary point of the set.
d) A set and a point that is both a boundary point and a limit point of the set.
2.2.2 Open and Closed Sets
Definition We say a set is open whenever it contains none of its boundary points.
Definition We say a set is closed whenever it contains all of its boundary points.
Example 2.15 Open intervals are open sets because the only boundary points of an
interval are the endpoints and neither are contained in the open interval.
2.2. LIMITS AND SETS 59
Example 2.16 Closed intervals are closed sets be because the only boundary points of
an interval are the endpoints and both are contained in the closed interval.
Theorem 2.21 The following are equivalent
1. S is an open set
2. Every s ∈ S is contained in an open interval that is completely contained in S.
3. R \ S is closed.
Theorem 2.22 The following are equivalent
1. S is an closed set
2. S contains all of its limit points.
3. R \ S is open.
Exercise 2.28 Is R open or closed? Is ∅ open or closed?
Exercise 2.29 {x : x 2 ≤ 57} is a closed set.
True or False 11
Which of the following statements are true? If false, modify the statement to be true.
Explain.
a) An open set never contains a maximum.
b) A closed set always contains a maximum element.
2.2.3 Optional – Connected sets
Suppose that there were no real number, s, such that s 2 = 57. Consider the two sets
U = {s : s 2 < 57} and V = {s : s 2 > 57}. Then an interval like [0, 8] could be
divided into two distinct sets, U and V. There would be a ’hole’ in the numberline.
This leads to the definition of connected that says a connected set cannot be covered
by two distinct open sets. That intervals are not connected a way of understanding
connectedness and investigating sets that may have more complicated structure than
intervals. Still it is reassuring that intervals are connected/
Definition We say a set, S, is connected if it is not contained in the union of two

disjoint non-empty, open sets.
Example 2.17 Finite sets are not connected
Here we present another nice application of the Nested Interval Theorem.
Theorem 2.23 A connected set is a, possibly infinite, interval.
Proof. Hint: this will be easiest to handle with an lemma: A subset of R is an interval
whenever x, y are in the set so is every z in between.
The more interesting part is that every interval is connected.
Theorem 2.24 The interval [a, b] is connected.
Proof. Assume not. Let U and V be two disjoint open sets such that [a, b] ∈ U ∪ V.
Since both R \ U and R \ V are closed this means that [a, b] is also covered by two
distinct closed sets. We will use this and the fact that closed sets contain all of their
limit points.
Since neither U nor V are empty, pick u0 ∈ U and v0 ∈ V and assume, without loss of
generality, that u0 < v0 , so
a ≤ u0 < v0 ≤ b.
Define inductively a sequence of closed, nested intervals, [un , vn ] with un ∈ U and
vn ∈ V and length, |vn − un | = |v02−u
n
0|
. The base case, [u0 , v0 ], satisfies the conditions.
Assume [un , vn ] has been defined. Let m be the midpoint of the interval [un , vn ]. Now
m ∈ (a, b) so is in either U or V. There are two cases:
1. If m ∈ U, let un+1 = m and vn+1 = vn .
2. If m ∈ V, let un+1 = un and vn+1 = m.
In either case, then un+1 ∈ U and vn+1 ∈ V. and [un+1 , vn+1 ] ⊂ [un , vn ]. So we have a
closed interval of the required form. Since m is the midpoint, we known that
1 1 v0 − u0 v0 − u0
|vn+1 − un+1 | = |vn − un | = · n
= n+1
2 2 2 2
These intervals are nested, closed and non-empty so we can applied the Extended
Nested Interval Theorem to say there is a point, x ∈ [un , vn ], for all n, and that
un → x and vn → x.
Now x is a limit point of R \ U, a closed set, so it must be in R \ U. That is, x is not
in U. But x is also a limit point of R \ V , another closed set, so x is not in V . x is
in neither V nor U, but it is a point in [a, b], so U and V cannot cover the interval as
originally supposed. Therefore, the interval is connected.
2.3. THE BOLZANO-WEIERSTRASS THEOREM AND CAUCHY SEQUENCES61
2.3 The Bolzano-Weierstrass Theorem and Cauchy

Sequences
Theorem 2.25 THE BOLZANO-WEIERSTRASS THEOREM Every bounded sequence has a con-
verging subsequence.
Outline of proof: Name the sequence sn and let M be a bound for the sequence,
that is For all n, |sn | < M. We will construct a sequence of non-empty, closed
nested intervals whose lengths converge to 0 and a subsequence of sn such that snm is
contained in the mth interval. Necessarily, this subsequence converges to the common
point of the intervals.
Definition A sequence is called a Cauchy Sequence whenever
For all > 0, there exists a real number, N, such that
if both n, m > N =⇒ |sn − sm | < .
Theorem 2.26 A sequence converges if and only if it is a Cauchy sequence
Outline of proof: =⇒ straight forward triangle inequality considerations

⇐= Mimic the proof that a convergent sequence is bounded to show that a Cauchy
sequence is bounded. Apply B-W and finally use a triangle-inequality argument to
show that the sequence must converge to the limit of the subsequence.
2.4 Series and Power series

Series are special kinds of sequences where one keeps a running sum of a sequence,
an to create a new sequence, sn :
Definition The number

n
X
sn = ak = a0 + a1 + a2 + · · · + an−1 + an
k=0
is called the nth partial sum of the generating sequence, an . If the sequence, sn ,
converges to a point, s, we say the the series converges to s and we write
∞
X
s= ak
k=0
Definition A power series is a special kind of series where the generating sequence is
of the form an = cn · r n . If the cn ’s are constant, cn = c we call it a geometric series.
∞
X ∞
X
Theorem 2.27 ALGEBRAIC PROPERTIES OF SERIES If s = ak and t = bk , then
k=0 k=0
∞
X
1. s + t = (ak + bk ).
k=0
∞
X
2. If c ∈ R, then c · s = c · ak .
k=0
Exercise 2.30 Give examples of each of the following:
a. A series that diverges to +∞.
b. A series whose partial sums oscillate between positive and negative numbers.
n
X
Exercise 2.31 If an ≥ 0 for all n, then sn = ak is an increasing sequence.
k=0
∞
X ∞
X
Exercise 2.32 If s = ak and t = bk , write a possible formula for the terms of
k=0 k=0
a series that might be s · t. Prove that your series converges and is equal to s · t.
2.4.1 Convergence of Geometric series
Theorem 2.28 If 0 ≤ |r | < 1, then

∞
X 1
rn =
k=0
1−r
1
Proof. Hint: Mimic this trick from high school to see why 0.33333 · · · is equal to 3 –
be sure that you can justify each step:
S = 0.33333 · · · given
10 · S = 3.33333 · · · multiplication of decimals
9·S =3 Subtracting the first line from the second line
3
S= divide by 9
9
1
S= reduce fraction
3
2.4. SERIES AND POWER SERIES 63
2.4.2 Decimals
Definition A decimal representation looks like a0 .a1 a2 a3 · · · an · · · , where a0 is an

integer and an are integers between 0 and 9, inclusively. It represents the number
1
which is given by the power series where r = 10 . We write
∞
1 12 13 X 1n
a0 .a1 a2 a3 · · · an · · · = a0 + a1 · + a2 · + a3 · + ··· = an ·
10 10 10 n=0
10
Exercise 2.33 Prove: 0.999999... = 1.
Theorem 2.29 Every decimal representation is a convergent power series and hence
every decimal representation is a real number.
n
1k
X
Proof. The partial sums are bounded by a0 + 9·
≤ a0 + 1. The partial sums
k=1
10
are increasing. Every bounded increasing sequence converges.
As well, every real number has a decimal representation that converges to it, more
precisely
Theorem 2.30 Given any r ∈ R, there exists a sequence of integers an , where 0 ≤

ak ≤ 10 for all k ≥ 1 such that
∞
X 1n
an · =r
n=0
10
Proof. Use Theorem 1.55. and THE EXTENDED NESTED INTERVAL THEOREM.
Notice that nothing is said about the representation of a real number being unique.
In fact any rational number that has a representation that ends with an infinite string
of 0’s has another representation that ends in a string of 90 s. And vice versa.
Exercise 2.34 Make a flowchart of theorems and connective arguments that take us
from the Completeness Axiom to the representation of real numbers as decimals.
Chapter 3
Counting
3.1 Finite vs Infinite

We will be using the notion of a 1 − 1 correspondence of two sets. This is just a
bijection between the sets.
Definition Recall that we say that a set, S, is finite whenever there exists a integer,
N ≥ 0, and a 1 − 1 correspondence between S and {n ∈ Z+ : n ≤ N}. N is then the
number of elements in the set. We say that a set is infinite if it is not finite.
editor: define cardinality
Definition We say that a set, S, is countable whenever there exists 1 − 1 correspon-

dence between S and a subset of Z+ .
Theorem 3.1 Given two countable sets, U and V
1. Any subset of U is countable.
2. U ∪ V is countable
3. U ∩ V is countable
Proof.
True or False 12
Which of the following statements are true? Prove or give a counterexample.
a) If A0 ⊃ A1 ⊃ A2 ⊃ A3 ⊃ A4 · · · are all infinite subsets of R, then ∩∞

n=1 An is
infinite.
65
66 CHAPTER 3. COUNTING
b) If A0 ⊃ A1 ⊃ A2 ⊃ A3 ⊃ A4 · · · are all finite, non-empty subsets of R, then

∩∞n=1 An is finite and non-empty.
3.1. FINITE VS INFINITE 67
3.1.1 The rational numbers
Theorem 3.2 The rational numbers are countable.
Proof. Consider this array that contains all the positive rational numbers. By counting
along the diagonals, we can assign each rational number a positive integer to a rational
number.
111 221 531 641 11

5
1
6
1
7
1
8
1
9
1
10
1
11
1 · · ·
321 /22 732 /42 5

2
6
2
7
2
8
2
9
2
10
2
11
2 · · ·
431 823 /33 4

3
5
3
6
3
7
3
8
3
9
3
10
3
11
3 · · ·
941 /24 3
4
4
4
5
4
6
4
7
4
8
4
9
4
10
4
11
4 · · ·
1051 2
5
3
5
4
5
5
5
6
5
7
5
8
5
9
5
10
5
11
5 · · ·
1 2 3 4 5 6 7 8 9 10 11
6 6 6 6 6 6 6 6 6 6 6 · · ·
1 2 3 4 5 6 7 8 9 10 11
7 7 7 7 7 7 7 7 7 7 7 · · ·
1 2 3 4 5 6 7 8 9 10 11
8 8 8 8 8 8 8 8 8 8 8 · · ·
· · · · · · · · · · · · · ·
Example 3.1 This procedure that gives a list of all positive rational numbers is an
example of a sequence whose image is all positive rational numbers. This sequence
clearly diverges as we can find a subsequence that is unbounded by picking out the
sequence elements that are integers. We can also find subsequences that converge to
any given positive number.
Exercise 3.2 Using the sequence of rational numbers given above, describe a procedure
to find a subsequence that converges to a particular real number, r.
Example 3.2 There is a sequence whose image is the rational numbers between 0 and
1. List these numbers by first listing in order all fractions that have denominator 2,
followed by those with denominator 3, and so on. In Figure 3.1 we show this sequence.
In this case we eliminated duplicates as we listed the fractions.
Exercise 3.3 Using a similar argument show that NxN is countable.
Exercise 3.4 Using a similar argument show that the countable union of countable sets
is countable.
Figure 3.1: A sequence whose image is all rational numbers in [0,1]
3.1.2 How many decimals are there?

Theorem 3.3 The real numbers are not countable.
Proof. We will in fact show that the interval (0, 1) is not countable. The method of
proof is called ’Cantor’s diagonal argument’ after the mathematician who first used
it. Suppose we had had a correspondence between Z+ and the real numbers, using
the decimal representation of the real numbers we could list them like this:
1 0.a11 a12 a13 a14 a15 a16 a17 a18 a19 · · ·
2 0.a21 a22 a23 a24 a25 a26 a27 a28 a29 · · ·
3 0.a31 a32 a33 a34 a35 a36 a37 a38 a39 · · ·
4 0.a41 a42 a43 a44 a45 a46 a47 a48 a49 · · ·
5 0.a51 a52 a53 a54 a55 a56 a57 a58 a59 · · ·
6 0.a61 a62 a63 a64 a65 a66 a67 a68 a69 · · ·
7 0.a71 a72 a73 a74 a75 a76 a77 a78 a79 · · ·
8 0.a81 a82 a83 a84 a85 a86 a87 a88 a89 · · ·
··· ···
Make a number, 0.a1 a2 a3 a4 a5 a6 a7 a8 · · · that is not on the list by defining
ak = akk + 1 mod 10.
This number can not be on our list because it differs from the k th element on the list
in the k th decimal place.
3.2. CANTOR SETS 69
3.2 Cantor Sets

Draw a picture of the recursive procedure that defines the Cantor Set.
View the Cantor set as the intersection of a countable number of closed sets. Conclude
that the Cantor Set is closed.
The Cantor Set is not countable.
The complement of the Cantor Set has accumulated length of 1.
The Cantor Set consists of all real numbers in [0, 1] whose tertiary expansion can be
written with no 2’s
The dimension of the Cantor Set is log2 /log3
Exercises: construct other Cantor Sets.
Chapter 4
Functions
4.0 Limits
Definition Let f : D → R.
We write lim f (x) = L and say the limit as x approaches p of f (x) is L, whenever,
x→p
For all > 0, there exists δ > 0, so that for x ∈ D,

0 < |x − p| < δ =⇒ |f (x) − L| < .
To emphasize the domain, we may say: the limit as x approaches p in D.

Note: The 0 < is included because it is not necessary for the limit to be equal to
f (p). It is not even necessary for f to be defined at p. This is important later on when
we define derivatives as limits. The concern disappears when we use limits to define
continuity and derivatives.
We don’t go through the step of L = 0 as we did for limits of sequences because we
are so much better at deconstructing limit statements now. However, it is true that
Theorem 4.1 lim f (x) = L ⇐⇒ lim (f (x) − L) = 0

x→p x→p
Definition We say that f (x) converges to L sequentially as x → p whenever,
for any sequence, xn ∈ D − {p}

xn → p =⇒ f (xn ) → L
We do not consider values of p for the sequence because f (p) may not be equal to
L. Again, this won’t be a concern with continuity.
Theorem 4.2 lim f (x) = L if and only if f (x) converges to L sequentially as x → p

x→p
71
72 CHAPTER 4. FUNCTIONS
Proof. EFS
Now, either definition of limit can be used in the proof of the following theorem.
Theorem 4.3 ALGEBRAIC PROPERTIES OF LIMITS 3 Given two functions f and g defined on
a domain, D, with lim f (x) = A and lim g(x) = B, the following are true:
x→p x→p
1. lim (f + g)(x) = A + B
x→p
2. lim (f · g)(x) = A · B
x→p
1 1
3. If f (x) 6= 0 for any x ∈ D and if A 6= 0, then lim ( )(x) =
x→p f A
Proof. A good way to prove these theorems is using sequential convergence and the
corresponding theorems about sequences.
Theorem 4.4 SQUEEZE THEOREM Given two functions f and g defined on a domain,
D, with lim f (x) = lim g(x) = C, and another function, h, defined on D, with
x→p x→p
f (x) ≤ h(x) ≤ g(x) for all x ∈ D, then lim h(x) = C
x→p
Proof. Given > 0, find δ > 0 such that
|x − p| < δ =⇒ |f (x) − C| < and |f (x) − C| <
The following inequalities demonstrate that |h(x) − C| < , so lim h(x) = C

x→p
− ≤ −|f (x) − C| ≤ f (x) − C ≤ h(x) − C ≤ g(x) − C ≤ |g(x) − C| <
Exercise 4.1 Because we are taking advantage of sequential limits, we did not need to
prove the following theorem to prove Theorem 4.3 2. Prove this theorem:
Theorem 4.5 If lim f (x) = A, then f is bounded in some open interval about p.
x→p
4.1 Continuity
For the following, consider a function, f : D → R.
Definition CONTINUITY AT A POINT We say that the f is continuous at p ∈ D if
lim f (x) = f (p).

x→p
4.1. CONTINUITY 73
Example 4.1 The function x 2 is continuous at the point p = 3.
Proof. directly from the definition: We need to show that lim x 2 = 9.

x→3
Given > 0, find a δ < min{1, 7 }. Then we can see that if |x − 3| < δ we have

|x 2 − 9| = |x − 3| · |x + 3| < δ · 7 < · 7 = .
7
[Justifications: Because δ < min{1, 7 }, we know that δ is less than both 1 and 7 . if
|x − 3| < 1, then |x + 3| = |x − 3 + 6| ≤ |x − 3| + |6| < 1 + 6 = 7. This justifies the
first ’<’. The second ’<’ is justified because δ < 7 ]
So |x − 3| < δ =⇒ |x 2 − 9| < , as we needed to show.
√
Exercise 4.2 Directly from the definition of limit, show that the function x is con-
tinuous at the point p = 4.
Definition CONTINUITY ON A SUBSET OF THE DOMAIN We say that a function, f , is contin-

uous on E ⊂ D if it is continuous at all points of E.
Exercise 4.3 The function x 2 is continuous on R.

√
Exercise 4.4 The function x is continuous on R≥ .
Exercise 4.5 Give an example of a function is continuous on all but 3 points in R.
4.1.1 Sequential continuity
We could have used a sequentially definition for continuous at a point. Here it is said
in a theorem.
Theorem 4.6 A function f is continuous at a point p ∈ D if and only if
for any sequence, xn ∈ D

xn → p =⇒ f (xn ) → f (p)
Proof.
Another way to state the conclusion of this theorem is: If f continuous on E and
sn ∈ E converges to a point in E, then
f ( lim sn ) = lim f (sn )

n→∞ n→∞
Exercise 4.6 Use the sequential definition of continuity to show that the function x 2
is continuous on R.
Exercise 4.7 Give an example of a function, f , and a sequence of points, sn , is the

domain of f , such that f (sn ) converges, but sn does not.
Exercise 4.8 Prove the basic facts about continuous functions stated in the following
theorem:
Theorem 4.7 ALGEBRAIC PROPERTIES OF CONTINUITY Given two functions f and g defined
and continuous on a domain, D, the following are true:
1. f + g is continuous on D.
2. f · g is continuous on D.
1
3. If f (x) 6= 0 for any x ∈ D, is continuous on D.
f
Proof. A good way to prove these theorems is to use the corresponding theorems
about limits of functions.
Theorem 4.8 The addition functions, sb : x → x + b, and multiplication functions,

tm : x → m · x, are continuous on R.
Theorem 4.9 Polynomials are continuous.
Proof. EFS First make a flowchart of the proof. What simpler facts should you prove
first? What is the overall plan of attack.
Theorem 4.10 Rational functions are continuous wherever they are defined.
Proof. EFS Use theorem 4.9
4.1.2 More Examples and Theorems

An aside to discussion inverse functions
The composition of two functions and the inverse of a function can be defined in
the most abstract settings. We review those concepts here before considering the
continuity of compositions and inverse functions.
Definition Given two functions, f : D → E and g : E → F , the function, g◦f : D → F

is defined for each x ∈ D by
g ◦ f (x) = g(f (x)).
We write g ◦ f and say composition of g with f .

4.1. CONTINUITY 75
Theorem 4.11 If f : D → E and g : E → F are continuous on their respective

domains, then g ◦ f : D → F is continuous on D.
Proof.
Definition We say that a function, f : D → E has an inverse function, if there exists

a function, f −1 : E → D such that
for all d ∈ D, f −1 ◦ f (d) = d.
Exercise 4.10 The proof of the following theorem is fundamental to understanding

inverse functions.
Theorem 4.12 If f : D → E has an inverse f −1 : E → D, then f −1 has an inverse

and it is f , that is
for all e ∈ E, f ◦ f −1 (e) = e.
Proof.
Exercise 4.11 There are not many examples available to us now without Υ. However,
we can note that the addition function, sb : x → x + b has an inverse. Show that
sb−1 = s−b .
Also, the multiplication function, tm : x → m · x, has an inverse as long as m 6= 0.

Show that
tm−1 = tm−1 .
Exercise 4.12 Write a formula for the composition function tm ◦ sb and another for
sb ◦ tm What are the inverses of these composition functions?
Theorem 4.13 Given two functions, f : D → E and g : E → F , If both have inverses,

so does g ◦ f : D → F . In fact,
(g ◦ f )−1 = f −1 ◦ g −1
Proof. EFS
A library of functions
It is good to have a well stocked library of functions to help think about the various
properties and theorems.
1. polynomials and rational functions
2. piecewise continuous: step functions

3. piecewise differentiable: absolute value
4. characteristic function of the rationals
5. cantor function
√
6. inverse functions, like x
7. Υ exponential, logarithms and trigonometric

8. power series
4.1.3 Uniform Continuity
Later we will need a stronger type of continuity:
Definition We say that a function f is uniformly continuous on E ⊂ D whenever
For all > 0, there exists a real number, δ > 0, such that
for all x and p ∈ E, |x − p| < δ =⇒ |f (x) − f (p)| < .
This defines a stronger condition than the just continuity. Given > 0 one must find
a δ that works for all p ∈ E. If one does not want to show uniform continuity, one
can find δ that is dependent on the particular value of p.
√
Exercise 4.13 f (x) = x is uniformly continuous on [1, ∞).
√
Exercise 4.14 f (x) = x is uniformly continuous on [0, 1].
Exercise 4.15 f (x) = x 2 is not uniformly continuous on [1, ∞).
Exercise 4.16 f (x) = x 2 is uniformly continuous on [0, 1].
Uniform continuity will be needed when we talk about integrals. The following theorem
is very helpful.
Theorem 4.14 A continuous function on a closed and bounded set is uniformly con-
tinuous on that set.
Proof. Let f :K → R, where K is a closed and bounded subset of R, be continuous at

all points in K. Assume f is not uniformly continuous on K. Then there exists an > 0
and sequences xn ∈ K, pn ∈ K and δn → 0 with |xn − pn | < δn but |f (xn ) − f (pn )| ≥ .
xn is bounded and so has a converging subsequence, xnk . Now pnk is bounded and so
has a converging subsequence, pnkm . For all m, we have
|xnkm − pnkm | < δnkm .

4.2. INTERMEDIATE VALUE THEOREM 77
So the limits are same:

lim xnkm = lim pnkm .
m→∞ m→∞
By sequential continuity, limm→∞ |f (xnkm ) − f (ynkm )| → 0. So the difference cannot be

bounded below by .
Exercise 4.17 Is a linear combination of two uniformly continuous functions uniformly

continuous? Explain.
Exercise 4.18 What are the other analogous facts for continuous functions and uni-
formly continuous functions from Theorem 4.3? Are they true or not?
4.2 Intermediate Value Theorem
Theorem 4.15 THE INTERMEDIATE VALUE THEOREM A function that is continuous on a

closed, bounded interval [a, b] attains every value on the interval [f (a), f (b)].
Proof. If f (a) = f (b), we are done. So assume, without lose of generality, that
f (a) < f (b) and let c be any number in [f (a), f (b)]. We will show that there exists two
nested sequences of closed intervals, [an , bn ] and [f (an ), f (bn )] with c ∈ [f (an ), f (bn )]
such that the lengths of both go to zero.
We will then use the extensions of the nested interval theorem to show that the unique
number in all the [f (an ), f (bn )] is f (c). We define the intervals inductively:
Base case: Let a0 = a and b0 = b. c was chosen to be in [f (an ), f (bn )]. Assume that
[an , bn ] and [f (an ), f (bn )] are defined as required, proceed inductively:
Let m be the midpoint of the interval [an , bn ]. There are three cases:
1. If f (m) = c, we are done.
2. If f (m) < c < f (bn ), let an+1 = m and bn+1 = bn .
3. If f (an ) < c < f (m), let an+1 = an and bn+1 = m.
In either case of the last two cases, f (an+1 ) < c < f (bn+1 ). This new interval,
[an+1 , bn+1 ] is half the length of the previous because m in the midpoint. Because the
intervals are cut in half at each step, the lengths converge to 0 and the extension of
the Nested Interval Theorem: there is a point, p, in all of the intervals and an → p and
bn → p. By sequential continuity f (an ) → f (p) and f (bn ) → f (p). By the squeeze
theorem, f (p) = c.
Exercise 4.19 Show that there exists a real number c such that c 4 − c 2 = 3. Is c
unique?
Theorem 4.16 Let p : R → R be a polynomial of odd degree. There exists a real

number, c, such that p(c) = 0
Proof. EFS
Exercise 4.20 Use the THE INTERMEDIATE VALUE THEOREM to show that there was a time
in your life when your height in inches was equal to your weight in pounds.
Exercise 4.21 Prove: If f : [a, b] → R, g : [a, b] → R, f (a) < g(a) and g(b) < f (b),
there exists a real number, c ∈ (a, b), such that f (c) = g(c).
Theorem 4.17 If a function,f : [a, b] → R is strictly increasing on the interval, [a, b],
then there exists an inverse function, f −1 : [f (a), f (b)] → [a, b]. f −1 is increasing on
the interval [f (a), f (b)].
Proof. Since f is increasing on the interval, f (a) is the minimum value of f and f (b)
is the maximum value of f . We will define f −1 (y ) for each y ∈ [f (a), f (b)], and then
show that f −1 is indeed the inverse of f .
For y ∈ f (a), f (b)], there is an unique x ∈ [a, b] such that f (x) = y . The THE
INTERMEDIATE VALUE THEOREM (Theorem 4.15) guarantees there is at least one such x.
There cannot be two because the function is strictly increasing. So define f −1 (y ) = x.
Clearly, f −1 ◦ f (y ) = y .
True or False 13
a) A decreasing function defined on [a, b] has an inverse defined on [f (b), f (a)]?
b) An increasing function defined on (a, b) has an inverse defined on (f (a), f (b))?
c) If a function has an inverse on an interval, it must be either increasing or de-

creasing on that interval.
d) The function f : x → x 2 has an inverse.
4.3 Continuous images of sets

Subsets of R that are both closed and bounded are also called compact. There is
a more general definition for compact in other topological spaces but, for this book,
we stick to closed and bounded. Compact sets are important in topology for reasons
given in the next few theorems.
Theorem 4.18 The continuous image of a closed bounded set is closed and bounded.
4.4. OPTIONAL: CONNECTED SETS 79
Proof. Setup: Let C be a closed bounded set and let f : C → R be continuous. We

will show that f (C) is also closed and bounded.
First, f (C) is bounded: Assume the image of C is not bounded. Then there exists
a sequence of points in f (C) that converge to +∞ (or −∞). Now, the xn live back
in C and so, by THE BOLZANO-WEIERSTRASS THEOREM(Theorem 2.25), have a convergent
subsequence, say xnk → x. Because C is closed x ∈ C. Now, f is continuous on
C so f (xnk ) → f (x). But a convergent sequence is bounded so f (xnk ) cannot be a
subsequence of a sequence that converges to +∞ (or −∞). This is a contradiction,
so f (C) is bounded.
f (C) is closed: Let y be a limit point of C. So yn = f (xn ) → y . We will show that
y ∈ f (C) which will prove that f (C) is closed. Because xn ∈ C and C is bounded,
there is a subsequence xnk → x (BW again). Since C is closed, x ∈ C. Since f is
continuous on C, f (xnk ) → f (x). So y = f (x) and y ∈ f (C)
Theorem 4.19 The inverse image, under a continuous function, of an open set is
open.
True or False 14
a) The continuous image of an open set is open.
b) The continuous image of an unbounded set is unbounded.
c) The inverse image, under a continuous function, of an closed set is closed.
d) The inverse image, under a continuous function, of a bounded set is bounded.
4.4 Optional: Connected Sets
Theorem 4.20 The continuous image of a connect set is connected.
An alternative proof of the intermediate value theorem: Let f : [a, b] → R be contin-

uous. The image of [a, b] is connected so it is an interval. It is bounded, so it is
a finite interval. Finally it is closed, so it is a closed interval. So this interval must
contain all of the interval [f (a), f (b)]. Hence every number in [f (a), f (b)] is in the
image of f .
4.5 Existence of extrema

Theorem 4.21 A continuous function defined on a closed, bounded set obtains a
maximum (and a minimum value) on that set.
Proof.
Exercise 4.22 Find the maximum and minimum value of the function, f (x) = x 2 −bx +c
on the closed interval, [0, 1]. You will have to handle cases for different values of b
and c.
4.6 Derivatives
4.6.1 Definitions
Definition DERIVATIVE AT A POINT We say that a function,f : D → R is differentiable at
a point p ∈ D, whenever the following limit exists.
f (x) − f (p)
lim . (4.1)
x→p x −p
Notation We call the limit 4.1 the derivative of f at p, and write f 0 (p).
f (x) − f (p)
f 0 (p) = lim . (4.2)
x→p x −p
Definition THE DERIVATIVE FUNCTION If the limit 4.1 exists for all p in a subset, E, of
the domain of f , it defines a function which we write as f 0 and we say that f is
differentiable on E
Theorem 4.22 If f : R → R is constant on any open interval, then f 0 exists on the

interval and f 0 (x) = 0 for all x in the interval.
Exercise 4.23 Directly from the definition, find a formula for the derivative of the
following functions at a point, p, in the domain of the function.
a) x 2
√
b) x
c) x
4.6.2 Applications
editor: Ask students to bring in examples from their own discipline
4.6. DERIVATIVES 81
practical significance The difference quotient represents an average rate – especially

useful when the domain is a time variable. The derivative at a point p, f 0 (p) represents
an instantaneous speed.
geometric significance The derivative at a point p, f 0 (p), represents the slope of

the tangent line to the graph of f at the point (p, f (p)). [[insert picture]]
4.6.3 Basic Theorems
Theorem 4.23 If f : D → R is differentiable on D, then f is continuous on D.
Theorem 4.24 Given two real-valued functions f and g defined on a domain, D ⊂ R

and c ∈ R.
1. c · f + g is differentiable on D, and (c · f + g)0 (x) = c · f 0 (x) + g 0 (x)
2. f · g is differentiable on D, and (f · g)0 (x) = f 0 (x) · g(x) + f (x) · g 0 (x)
Exercise 4.24 Restate the following theorem to include a formula for the derivative of
a polynomial. Prove the theorem.
Theorem 4.25 Polynomials are differentiable.
Exercise 4.25 If f (x) = x1 , then f is differentiable on (0, +∞) and f 0 (x) = − x12
Exercise 4.26 Carefully state a theorem that describes the derivative of a composition
of two functions. Prove your theorem.
True or False 15
If f 0 (x) > 0 on an interval (a, b) ⊂ D, then f (x) is increasing on (a, b).
4.6.4 Zero Derivative Theorem
The following series of theorems aims to show the converse of Theorem 4.22: If a
function has a derivative that is zero on an open interval then the function is constant
on that interval.
Theorem 4.26 If a function, f , differentiable on (a, b) has a maximum value at x0 ∈

(a, b), then f 0 (x0 ) = 0.
Proof.
Exercise 4.27 Give an example of a function that has a maximum value but the deriva-
tive at that point is not zero. Give one where the function is not differentiable and
one where the interval is closed instead of open.
Exercise 4.28 Give an example of a differentiable function, f , and a point, p, such that
f 0 (p) = 0, but f does not have a maximum or minimum value at p.
Theorem 4.27 ROLLE’S THEOREM f :[a, b] → R, f continuous on [a, b], and f differen-
tiable on (a, b).
f (a) = f (b) =⇒ There exists x ∗ ∈ (a, b) with f 0 (x ∗ ) = 0
Definition The mean value of a function, f , over the interval [a, b] is

f (b) − f (a)
b−a
Note that this is the slope of the line connecting (a, f (a)) and (b, f (b)).
This is an example of an existence theorem. The proof relies ultimately on BW which

was constructive but unwinding that construction may be hard. In calculus, we find
typically find those points where a function has maximum by solving f 0 (x) = 0.
Exercise 4.29 Write an equation for the straight line determined by the two points,
(a, f (a)) and (b, f (b)). Use the point-slope formula for the equation of a line.
Example 4.2 average speed
Theorem 4.28 THE MEAN VALUE THEOREM f :[a, b] → R, f continuous on [a, b], and f
differentiable on (a, b).
f (b) − f (a)
There exists x ∗ ∈ (a, b) with f 0 (x ∗ ) =
b−a
Exercise 4.30 Where on the interval [−1, 1] does the derivative of f (x) = x 3 assume
mean value of the function? Sketch a graph.
Theorem 4.29 UNIQUENESS THEOREM If f 0 (x) = 0 for all x ∈ (a, b), then f is constant
on the interval.
Proof. Let z1 and z2 be any two points in the interval, (a, b). By THE MEAN VALUE
∗
THEOREM there is a point, x ∈ (z1 , z2 ), such that
f (z2 ) − f (z1 )
f 0 (x ∗ ) = .
z2 − z1
4.6. DERIVATIVES 83
But f 0 (x ∗ ) = 0 by hypothesis. This then means that f (z1 ) = f (z2 ). Since they were
any two points in (a, b), f must be constant.
Chapter 5
Integration
5.0 Definition of Riemann Integral
5.0.1 Partitions of an interval

Definition A partition of the interval [a, b] is a finite, increasing sequence of numbers,
xj , such that
a = x0 < x1 < x2 < · · · < xn−1 < xn = b.
The points divide [a, b] into n disjoint open subintervals, (xi , xi+1 ), such that the closed
n−1
[
intervals cover [a, b], i.e. [a, b] = [xi , xi+1 ]
i=0
The mesh of a partition is the length of the largest sub-interval. If P is a partition
of [a, b] then we write |P | for the mesh of P . The mesh is used as a way of ordering
partitions. This kind of ordering is called a partial ordering because there is not
trichotomy, i.e. two different partitions can have the same mesh. This ordering is
useful because we want to talk about the concept of the limit as the mesh of the
partitions goes to zero.
A tagged partition of the interval [a, b] is a partition with the additional n points,
tk∗ ∈ [xk , xk−1 ]. So we have
a = x0 ≤ t1∗ ≤ x1 ≤ t2∗ ≤ x2 ≤ t3∗ · · · tn−1

∗
≤ xn−1 ≤ tn∗ ≤ xn = b.
There are lots of partitions of an interval. The x’s may be placed with any spacing
between them and the t ∗ ’s can be any point with the interval. They are used to help
us define the integral of a function. We need to consider, at least at the beginning, all
possible partitions because we want to know that as long as we choose a reasonable
way to compute the integral we get the same answer. Our definition of integral must
make it so.
editor: stick in some pictures
85
86 CHAPTER 5. INTEGRATION
5.0.2 Definition of Riemann Integral

Definition Given any function, f : [a, b] → R, and a partition of [a, b], we define
n
X
P (f ) = f (tk∗ ) · (xk − xk−1 )
k=1
we call any such P (f ) a Riemann sum of f over [a, b].
editor: insert picture
Definition We say the a function, f : [a, b] → R is Riemann-integrable whenever there

exists I ∈ R such that
For all > 0, there exists δ > 0 so that, for any partition, P
|P | < δ =⇒ |P (f ) − I| <
We notate I by
Z b
f (x)dx
a
Named after Riemann xxxx was the first to give a precise definition of integral
Example 5.1 This is a pretty unmanageable definition but we can show that
Z b
1 · dx = b − a,
a
for let P be any tagged partition. Since the function value at each ti∗ is 1, the Riemann
sum is just the sum of the lengths of all the intervals or b − a. And so all R-sums are
the same.
With a definition that involves the existence of something over a seemingly unmanage-
able sets as the set of all tagged partitions of an interval, it is easier to find examples
of what is not than to prove something is. Here are two examples of functions that
are not integrable:
Example 5.2 The function

(
1
x, if 0 < x ≤ 1
f (x) =
0, if x = 0
is not integrable on [0, 1].

5.0. DEFINITION OF RIEMANN INTEGRAL 87
Proof that f is not integrable: Consider a sequence of partition, Pn , with n equal-

length intervals, given by points, xi = ni . Let t0∗ = n12 . For any choice of the rest of
the tk∗ ’s, the Riemann sum,
n
X 1 1
Pn (f ) = f (tk∗ ) · (xk − xk−1 ) > 1 · = n.
k=1 n2
n
Although mesh(Pn ) → 0, Pn → ∞. So f is not Riemann integrable.
Similar considerations allow us to conclude the following theorem:
Theorem 5.1 If f is integrable on [a, b], then f is bounded on [a, b].
Proof. Hint: Assume f is not bounded and, for every n > 0, find a partition, Pn such
that |Pn (f )| < n1 , but Pn (f ) > n.
Example 5.3 The function (

1, if x ∈ Q
g(x) =
0, if x ∈
/Q
is not integrable on [0, 1].
Proof that g is not integrable: Let Pn be any partition with mesh less than n1 and ti∗
all rational numbers. Let Qn be the same except the tagged numbers are all irrational.
Then Pn (g) = 1 and Qn (g) = 0. We have given to sequences of Riemann sums that
converge to different numbers, so g cannot be Riemann integrable.
In each example, we only needed to find a sequence of partitions that converged to

zero but that the corresponding Riemann sums did not converge. It would be so nice
if we could simplify the existence of the integral by looking at a single sequence of
partitions. That’s what the next theorem allows:
Theorem 5.2 SEQUENTIAL DEFINITION OF RIEMANN INTEGRAL A function f : [a, b] → R is

Riemann integrable if and only if for every sequence of partitions, Pn , with |Pn | →
0 =⇒ Pn (f ) converges.
Proof. NOTE: If all partitions converge they necessarily converge to a common limit
because we can intertwine the sequences.
Example 5.4 Consider The function

(
√1 , if x ∈ (0, 1]
x
f (x) =
0, if x = 0.
Find two sequences of tagged partitions of [0, 1], one diverging, one converging.
Now if we only knew that a function were integrable, we could pick out a convenient
sequence of partitions and life would be easier. The next theorem gives us what we
need for a large class of functions.
Theorem 5.3 Continuous functions are Riemann integrable
Proof. Let f : [a, b] → R be continuous on [a,b]. We know by Theorem XXX, that f

is uniformly continuous on [a, b].
Let Pn be a sequence of tagged partitions such that mesh(Pn )→ 0. We will show that
the corresponding Riemann sums is Cauchy and hence converges.
Given > 0, use the uniform continuity of f to find δ such that |x − y | < δ =⇒

|f (x) − f (y )| < b−a
Now, find a positive integer N such that
δ
n > N =⇒ mesh(Pn ) < .
2
Consider any two partitions, Pn and Pm , with n, m > N. Label the tags from Pn with t ∗
and the tags from Pm with s ∗ . Consider a refinement of the two partitions, Qn+m with
n + m elements, xk , the combined points in the two given partitions. The difference
between the Riemann sums will look like:
n+m
X
Pn (f ) − Pm (f ) = (f (tk∗ ) − f (sk∗ ) · (xk − xk−1 )
k=1
tk∗ ∗
where the is the t assigned to whatever Pn interval contains [xk , xk−1 ] and similarly
for sk∗ . We claim that |tk∗ − sk∗ | < mesh(Pn )+mesh(Pm ) < δ. Proof of claim: Each
interval in the refinement is contained in a unique interval from each of Pn and Pm . If
both endpoints are from Pn , then t ∗ comes from that interval so |tk∗ −sk∗ | < mesh(Pm ).
Similarly, |tk∗ − sk∗ | < mesh(Pn ) if both xk and xk+1 are in Pm . If each endpoints is from
a different partition, t ∗ and s ∗ come from two overlapping intervals – they may be as
far apart as the length of one plus the length of the other. Since the lengths are less
than respective mesh, we have |tk∗ − sk∗ | < mesh(Pn )+mesh(Pm ). Which means we
also have:

|tk∗ − sk∗ | < δ =⇒ |f (tk∗ ) − f (sk∗ )| <
b−a
Putting it all together:
n+m
X
|Pn (f ) − Pm (f )| = | (f (tk∗ ) − f (sk∗ ) · (xk − xk−1 )|
k=1
n
X
≤ |f (tk∗ ) − f (sk∗ )| · (xk − xk−1 )
k=1
n+m
X
≤· (xk − xk−1 )
k=1
≤ (b − a)
5.0. DEFINITION OF RIEMANN INTEGRAL 89
Now we are free to use any of the techniques that are normally seen in a calculus
courses: Consider partitions that are evenly spaced, tagged values that are endpoints,
or maximum values or whatever.
Z b
Example 5.5 Compute x 2 dx
a
Z b
Exercise 5.1 Compute xdx (Please do note that the answer is the area bounded
a
by the curve, the x-axis, and the vertical lines x = a and x = b.)
5.0.3 Theorems
Theorem 5.4 ALGEBRAIC PROPERTIES OF RIEMANN INTEGRALS Given two functions, f and g,
both integrable on the closed interval [a, b]. The following are true:
Z b Z b Z b
1. f (x) + g(x)dx = f (x)dx + g(x)dx
a a a
Z b Z b
2. c · f (x)dx = c · f (x)dx
a a
Z b Z c Z b
3. f (x)dx = f (x)dx + f (x)dx
a a c
Theorem 5.5 ORDER PROPERTIES OF RIEMANN INTEGRALS Given two functions, f and g,
both integrable on the closed interval [a, b]. The following are true:
1. If f (x) ≤ g(x)for all x ∈ [a, b], then

Z b Z b
f (x)dx ≤ g(x)dx)
a a
2. If M and m are respectively upper and lower bounds for f on [a, b], then
Z b
m · (b − a) ≤ f (x)dx ≤ M · (b − a)
a
Z b Z b
3. | f (x)dx| ≤ |f (x)|dx
a a
We may know what integrable functions are but it can be cumbersome if not impossible
to show the exact (or even approximate) value for a particular function, or class of
functions. That is where the theorems of the next section make life easier.
There is nothing in the definition of the Riemann Integral that prohibits b ≤ a.
Theorem 5.6 For any function, f , integrable on [a, b] ,

Z b Z a
1. f (x)dx = − f (x)dx
a b
Z a
2. f (x)dx = 0.
a
Proof.
5.1 Fundamental Theorem of Calculus
5.1.1 Integrals as Functions
Given a function, f , integrable on an interval [A, B] and a ∈ [A, B], we define a new
function F : [A, B] → R by Z x
F (x) = f (x)dx
a
Theorem 5.7 For any function, f , integrable on [A, B] and F ,
1. On any interval where f is positive, F is increasing.
2. F (a) = 0
Proof.
5.1.2 Statement of the Theorem
Theorem 5.8 FIRST FUNDAMENTAL THEOREM OF CALCULUS, AN EXISTENCE THEOREM Let f be

integrable on [a, b] and continuous on (a, b). For x ∈ [a, b] define
Z x
F (x) = f (t)dt.
a
Then the derivative of F exists for all x ∈ (a, b) and
F 0 (x) = f (x).
5.2. COMPUTING INTEGRALS 91
editor: The conclusion holds for any point,x, where f is continuous. f need not be
continuous on all of the open interval.
Proof.
Theorem 5.9 SECOND FUNDAMENTAL THEOREM OF CALCULUS, A UNIQUENESS THEOREM Let f

be continuous on an open interval, I, let F be an antiderivative for f on I, that is
F 0 (x) = f (x) for all x ∈ I. Then, for a ∈ I and each x ∈ I,
Z x
F (x) = F (a) + f (t)dt.
a
editor: Well, unique up to a constant.
5.2 Computing integrals
5.3 Application: Logarithm and Exponential Functions

Definition We define a function, log : R+ → R, using Riemann integral of x1 . That is,
Z x
1
log(x) = dt
1 t
1
NOTE: Because t is not integrable on any interval, (0, a), the log is not defined for
negative values.
Common properties Υ of logarithms can be derived from properties of integrals.
Exercise 5.2 Prove the follow theorem:
Theorem 5.10 1. log(1) = 0
2. For all x, y ∈ R+ , log(xy ) = log(x) + log(y )
5.4 Flowchart

Real Analysis: Interactive Notes For

Uploaded by

Copyright:

Available Formats

Real Analysis: Interactive Notes For

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Real Analysis: Interactive Notes For

Uploaded by

Copyright:

Available Formats

University of Illinois at Chicago

Interactive Notes For

3. Read and repeat proofs of the important theorems of Real Analysis:

• The Nested Interval Theorem

• The Bolzano-Weierstrass Theorem

• The Intermediate Value Theorem

• The Mean Value Theorem

• The Fundamental Theorem of Calculus

Two Exercises to get started.

Which of the following statements are true? Explain your answer.

1 The Real Number System 11

0.0 Getting Started

0.1.1 Common Sets

R≥ = R+ ∪ {0} represents the set of non-negative real numbers.

N0 or Z≥ both denote the set of non-negative integers.

Q+ represents the set of all positive rational numbers.

∅ represents the empty set.

0.1.2 Set Notation

We discuss three different ways to denote a set.

Example 0.3 {sin nπ

Exercise 0.1 ΥIs there a simpler description of {x 2 : x ∈ R}?

Exercise 0.3 ΥFind numbers a, b, c so that the formula, an2 + bn + c, for n ∈ Z≥ ,

0.1.3 Operations on Sets

0.2.1 Logical Statements

statement negation of statement

a) If x ∈ [2, 4], then x 2 ∈ [4, 16].

b) If x 2 ∈ [4, 16], then x ∈ [2, 4].

For all x ∈ R, x 2 ≥ 0 (0.1)

There exists x ∈ R, x 2 = 9 (0.3)

Exercise 0.5 Write the negation of each of the above statements.

0.3.1 Definitions, Notation and Examples

Notation If D is the domain of a function f , we write f : D → R

Definition The following definitions apply to the function, f : D → R. These notions

· For any T ⊂ f (D), we say the pre-image of T under f , and write f −1 (T ), to

· We say that f is 1 − 1 or f is an injection, if for all x, y ∈ D

· We say that f is bijection, if it is both an injection and surjection. In this case,

0.3.2 Sequences are functions Z+ → R or Z≥ → R

0.4 True or False

a) There exists an x ∈ Z, such that x is odd.

b) For all x ∈ Z, x is even.

c) There exists an x ∈ Z, such that 2x is odd.

d) For all x ∈ Z, 2x is even.

(From Morgan) Which of the following statements are true? Explain.

a) For all x ∈ R, there exists a y ∈ R such that y > x 2 .

b) There exists an y ∈ R, such that for all x ∈ R, y > x 2 .

c) There exists an y ∈ R, such that for all x ∈ R, y < x 2 .

d) For all a, b, c ∈ R, there exists x ∈ R such that ax 2 + bx + c = 0.

Which of the following statements are true? Explain.

a) There exists a real number, x, such that x 2 = 9.

b) There exists a unique real number, x, such that x 2 = 9.

c) There exists a unique positive real number, x, such that x 2 = 9.

The following True or False problem concerns the notion of pre-image.

Which of the following statements true? Prove or give a counterexample. Consider

The Real Number System

We begin by supposing the existence of a set, R, whose elements we call real

1.0 Definitions and Basic Notions from Algebra

Basic Notion 1 EQUALITY OF REAL NUMBERS IS AN EQUIVALENCE RELATION. The following

1.0.2 Addition and Multiplication