Set Theory Lecture Notes

Foundations of Mathematics I
Set Theory
(only a draft)
Ali Nesin
Mathematics Department
Istanbul Bilgi University
Kuştepe Şişli
Istanbul Turkey
anesin@bilgi.edu.tr
February 12, 2004

2
Contents
I Naive Set Theory 7

1 Basic Concepts and Examples 11
1.1 Sets, Subsets and Emptyset . . . . . . . . . . . . . . . . . . . . . 11
1.2 Notes on Formalism . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.3 Number Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.4 Subsets Defined By a Property . . . . . . . . . . . . . . . . . . . 20
1.5 Sets of Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.6 Parametrized Sets . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2 Operations with Sets 25

2.1 Difference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2 Intersection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3 Union . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.4 Cartesian Product of Two Sets . . . . . . . . . . . . . . . . . . . 29
3 Functions 31
3.1 Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.2 More On Functions . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.3 Binary Operations . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.4 Operations with Functions . . . . . . . . . . . . . . . . . . . . . . 37
3.5 Injections, Surjections, Bijections . . . . . . . . . . . . . . . . . . 38
3.5.1 Injections . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.5.2 Surjections . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.5.3 Bijections . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.6 Sym(X) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.7 Families, Sequences and Cartesian Products . . . . . . . . . . . . 44
4 Relations 47
4.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.2 Equivalence Relations . . . . . . . . . . . . . . . . . . . . . . . . 48
4.3 Partial Orders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.4 Total Orders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5 Induction 57
3
4 CONTENTS
6 Bijections, revisited 61
6.1 Schröder-Bernstein’s Theorem . . . . . . . . . . . . . . . . . . . . 61
6.2 Examples of Sets in Bijection . . . . . . . . . . . . . . . . . . . . 61
7 Some Automorphism Groups 63

7.1 Binary Unirelational Structures . . . . . . . . . . . . . . . . . . . 63
7.2 Automorphism Groups of Graphs . . . . . . . . . . . . . . . . . . 64
7.3 Geometric Automorphism Groups . . . . . . . . . . . . . . . . . 65
7.4 Back and Forth Argument . . . . . . . . . . . . . . . . . . . . . . 67
7.4.1 Dense Total Orderings . . . . . . . . . . . . . . . . . . . . 67
7.4.2 Random Graphs . . . . . . . . . . . . . . . . . . . . . . . 68
8 Formulae 69
9 Miscellaneous Exercises 71
II Axiomatic Set Theory 73

10 Basics 75
10.1 Russell’s Paradox . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
10.2 Easy Axioms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
10.3 Slightly More Complicated Axioms . . . . . . . . . . . . . . . . . 81
10.4 Cartesian Product of Two Sets . . . . . . . . . . . . . . . . . . . 82
10.5 Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
11 Natural Numbers 87
11.1 Definition and Basic Properties . . . . . . . . . . . . . . . . . . . 87
11.2 Well-ordering on ω . . . . . . . . . . . . . . . . . . . . . . . . . . 88
11.3 Peano’s Axioms . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
11.4 Addition of Natural Numbers . . . . . . . . . . . . . . . . . . . . 90
11.5 Multiplication of Natural Numbers . . . . . . . . . . . . . . . . . 92
11.6 Well-Ordering of Natural Numbers . . . . . . . . . . . . . . . . . 94
11.7 Finite and Infinite Sets . . . . . . . . . . . . . . . . . . . . . . . . 96
11.8 Functions Defined by Induction . . . . . . . . . . . . . . . . . . . 96
11.9 Powers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
11.10Divisibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
11.11Uniqueness of N . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
12 Integers 101
12.1 Definition of The Set Z of Integers . . . . . . . . . . . . . . . . . 101
12.2 Operations +, − and × on Z . . . . . . . . . . . . . . . . . . . . 102
12.3 Ordering on Z . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
12.4 Embedding of (N, +, ×, <) in (Z, +, ×, <) . . . . . . . . . . . . . 106
12.5 Additive Structure of Z . . . . . . . . . . . . . . . . . . . . . . . 107
12.6 Multiplicative Structure of Z . . . . . . . . . . . . . . . . . . . . 108
12.7 Ordering on Z, Revisited . . . . . . . . . . . . . . . . . . . . . . . 108
CONTENTS 5
12.8 Divisibility and Subgroups . . . . . . . . . . . . . . . . . . . . . . 109

12.9 Euclidean Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . 112
12.10Quotients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
12.11Chinese Remainder Theorem . . . . . . . . . . . . . . . . . . . . 113
12.12Subgroups of Z ⊕ Z . . . . . . . . . . . . . . . . . . . . . . . . . . 113
13 Rational Numbers 115

13.1 Rational Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . 115
13.2 Some Combinatorics . . . . . . . . . . . . . . . . . . . . . . . . . 118
14 Real Numbers 119
15 Well-Ordered Sets 121

15.1 Definitions and Examples . . . . . . . . . . . . . . . . . . . . . . 121
15.2 Transfinite Induction . . . . . . . . . . . . . . . . . . . . . . . . . 122
15.3 Morphisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
16 Ordinals 125
16.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
16.2 Axiom of Replacement . . . . . . . . . . . . . . . . . . . . . . . . 126
16.3 Classification of Well-Ordered Sets . . . . . . . . . . . . . . . . . 126
16.4 Addition of Ordinals . . . . . . . . . . . . . . . . . . . . . . . . . 127
16.5 Multiplication of Ordinals . . . . . . . . . . . . . . . . . . . . . . 127
17 Cardinals 129
17.1 Addition of Cardinals . . . . . . . . . . . . . . . . . . . . . . . . 129
17.2 Multiplication of Cardinals . . . . . . . . . . . . . . . . . . . . . 129
17.3 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
17.4 Final Exam of Math 112, May 2003 . . . . . . . . . . . . . . . . 129
18 Axiom of Choice and Zorn’s Lemma 131

18.1 Zorn’s Lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
18.2 Some Consequences of Zorn’s Lemma . . . . . . . . . . . . . . . 132
18.2.1 Finite and Infinite Sets . . . . . . . . . . . . . . . . . . . 132
18.2.2 König’s Lemma . . . . . . . . . . . . . . . . . . . . . . . . 132
18.2.3 Some Unexpected Consequences of Zorn’s Lemma . . . . 132
19 Axioms of Set Theory – ZFC 133
20 V = L 135
21 Continuum Hypothesis 137
22 Banach-Tarski Paradox 139
23 First Order Structures 141

6 CONTENTS
24 Ultraproducts and Ultrafilters 143

24.1 Nonstandard Models of N . . . . . . . . . . . . . . . . . . . . . . 143
24.2 Nonstandard Models of R . . . . . . . . . . . . . . . . . . . . . . 143
25 Dimension Theory 145
26 Exams 147
26.1 First Semester Midterm, November 2002 . . . . . . . . . . . . . . 147
26.2 First Semester Final, January 2003 . . . . . . . . . . . . . . . . . 152
26.3 First Semester, Resit, January 2004 . . . . . . . . . . . . . . . . 153
26.4 First Semester, February 2003 . . . . . . . . . . . . . . . . . . . . 158
26.5 First Semester Final and Its Correction, January 2004 . . . . . . 160
26.6 Second Semester, Midterm, May 2003 . . . . . . . . . . . . . . . 163
Part I
Naive Set Theory
7
9
This part is about naive set theory, as opposed to axiomatic set theory that
will be the subject of the next part. Naive set theory does not care so much
about the definition or the existence of a set; on the contrary, in the axiomatic
set theory our primary concern will be the existence of a set.
Here, in this part, we will accept the concept of a “set” without defining
it; further we will assume that any collection of mathematical objects is a set,
whatever this means. Going even further, we will assume that the natural
numbers such as 0, 1, 2, integers such as √ −2, −4, rational numbers such as
3/5, −6/5 and the real numbers such as 2, π, π π exist and that the reader is
familiar with the basic operations of + and × performed with them. In other
words we assume all that has been taught in elementary school, mid-school and
high school.
In the next part we will define all mathematical objects carefully (this means
mathematically). We will prove that all the mathematical objects that are
defined are in fact sets. To prove that these objects are sets, we will use axioms
of set theory, that we will define. But for the moment we are away from this
realm and we work intuitively, without asking the question of existence.
10
Chapter 1
Basic Concepts and

Examples
1.1 Sets, Subsets and Emptyset

A set is just a collection of objects1 . The objects of the collection that form
the set are called the elements of the set.
One should be aware of the fact that there is nothing that prevents a set
from being an element of another set. Indeed, as we will see in the next part,
not only every element is a set, but that everything is a set!
An element of a set is sometimes called a member of the set.
If X is a set, to express the fact that x is an element of X, we write x ∈ X.
If x is not an element of the set X, we write x 6∈ X.
For example, there is a set whose elements are just 0, 1 and 2; we denote
this set by {0, 1, 2}. Clearly 2 ∈ {0, 1, 2}, but2 3 6∈ {0, 1, 2}. In general, a set
that has finitely many elements x1 , . . . , xn is denoted by {x1 , . . . , xn }.
Conversely, given a finite number of n sets x1 , . . . , xn , we can form the set
{x1 , . . . , xn }. The sets x1 , . . . , xn are elements of the set {x1 , . . . , xn }.
In particular, if x is a set, we can form the set {x} whose only element is x.
We have x ∈ {x}. A set with only one element is called a singleton.
Two sets are called equal if they contain the same elements (and not if they
have the same number of elements as some would say). For example {0} 6= {2}
because these two sets have different elements.
A set is called finite if it has finitely many elements. Otherwise we say that
the set is infinite. A finite set cannot be equal to an infinite set. Also two finite
sets that have different number of elements cannot be equal.
1 Notice how loose and unprecise we are.
2 To prove that 3 6∈ {0, 1, 2}, we need to show that 3 6= 0, 3 6= 1 and 3 6= 2. But to
prove these inequalities we need to know the mathematical definitions of 0, 1, 2 and 3. Since
these numbers are not mathematically defined yet, we cannot prove for the moment that
3 6∈ {0, 1, 2}. We will do this in the next part. For the moment we assume that every
natural number is different from the rest.
11
12 CHAPTER 1. BASIC CONCEPTS AND EXAMPLES
The set {0, 1, 2, 3, . . .} is infinite.

In general, we have x 6= {x} because the set {x} has only one element,
whereas the set x could have more than one element. Note that if x = {x}, then
x ∈ x, i.e. x is an element of itself, a quite strange and unexpected phenomena,
which will be forbidden by an axiom that will be introduced in the next part,
in other words such an object will not be a set (it may be something else!)
The sets {x, x, y}, {x, y} and {y, x} are equal because they contain the same
elements, namely x and y. This set has either one or two elements: It has one
element if x = y, otherwise it has two elements. On the other hand the set
{3, 5} has exactly two elements because 3 6= 5 (as we will prove in the second
part of our book once we know what these objects are).
It would be interesting to know what the reader things about the equality
2 = {0, 1}. Does it hold or not? It all depends on the definition of 2. As we will
see in the next part, the integer 2 will be defined as the set {0, 1}, so that the
equality 2 = {0, 1} does indeed hold. But we will not go into these fine points
of set theory in this part.
The number of elements of a set X is denoted by |X|. Thus |{0, 1, 2}| = 3,
|{0, 1, 2, 3, . . .}| = ∞ and 1 ≤ |{x, y}| ≤ 2. The set {x, y} has always at least
one element.
Here is another set: {{0, 1, 2}, {2, 3, 6}}. This set has just two elements,
namely {0, 1, 2} and {2, 3, 6}. The two elements {0, 1, 2} and {2, 3, 6} of this set
are sets themselves and each have three elements.
A set x is called a subset of another set y if every element of x is also an
element of y. We then write x ⊆ y. In this case, one sometimes says that y is a
superset of x.
For example, the set {0, 2, 3} is a subset of the set {0, 1, 2, 3, 4} because each
one of the elements of the first one appears in the second one. But the set
{0, 2, 3} is not an element of the set {0, 1, 2, 3, 4} unless {0, 2, 3} is equal to one
of 0, 1, 2, 3, 4 (which is not the case as we will see in the next part).
Another example: The set of even integers is a subset of the set of integers
divisible by 6.
We will refrain from giving non mathematical examples such as “the set of
pupils in your class is a subset of the set of pupils of your school”.
Another (mathematical) example: The set {0, 1} is a subset of {0, 1, 2}. This
is clear. Is it an element? No if the set {0, 1} is not equal to one of 0, 1 and 2.
Unfortunately, the way we will define numbers in the second part of the book,
2 is equal to the set {0, 1}; therefore {0, 1} is also an element of {0, 1, 2}. But
the reader is not supposed to know these fine points at this stage and may as
well suppose that {0, 1} is not an element of {0, 1, 2}. Anyway, we will try to
convince the reader in the second part that the definition of 2 as the set {0, 1}
is in some sense arbitrary.
Clearly every set is a subset of itself. Thus for all sets x, we have x ⊆ x.
We also note the following fact which is used very often in mathematics:
Two sets x and y are equal3 if and only if x ⊆ y and y ⊆ x. This means that
3 Here we act as if we know what the equality means, because we are doing intuitive set
1.1. SETS, SUBSETS AND EMPTYSET 13
two sets are equal if they have the same elements, i.e. if an element of one of
them is an element of the other and vice versa.
If x ⊆ y and x 6= y, then we write x ⊂ y. For example, {0, 3} ⊂ {0, 2, 3, π}.
If the set x is not a subset of the set y, we write x 6⊆ y. For example,
{π} 6⊆ {3.14, 3.14159, 22/7}.
The statement of our first theorem must be well known by the reader, but
its proof may not be so well known. In fact the proof is very interesting and we
urge the reader to digest it well.
Theorem 1.1.1 A set with no elements is a subset of every set.
Proof: Let ∅ be a set with no elements4 . Let x be any set. Assume ∅ is not
a subset of x. We will obtain a contradiction, proving that ∅ is a subset of x.
Since ∅ is not a subset of x, by the very definition of the concept of “subset”,
there is an element in ∅ which is not in x. But ∅ has no elements. Thus ∅ cannot
have an element which is not in x. Hence ∅ ⊆ x. ¤
Corollary 1.1.2 There is at most one set without elements.
Proof: Let ∅1 and ∅2 be two sets with no elements. By the theorem above,
∅1 ⊆ ∅2 and ∅2 ⊆ ∅1 . Hence ∅1 = ∅2 . ¤
A set that has no elements at all is called emptyset and is denoted by ∅.
As we have seen in the Corollary above there is only one set with no elements5 .
Here we list all the subsets of the set {0, 1, 2}:
∅
{0}
{1}
{2}
{0, 1}
{0, 2}
{1, 2}
{0, 1, 2}
The set of subsets of a set X is denoted by ℘(X). Thus the set ℘({0, 1, 2})
has the eight elements listed above.
Theorem 1.1.3 If |X| = n is finite then |℘(X)| = 2n .

theory. In the second part, in some sense, the equality will be defined so as to satisfy this
property.
4 We do not know yet that there is only one set with no elements. We will prove it right
after this theorem.

5 In fact, all we have shown in the Corollary is that a set with no elements is unique in case
there is such a set. In the next part one of our axioms will state that there is a set with
no elements. This fact cannot be proven, so either the existence of such a set or a statement
implying its existence must be accepted as an axiom.
Proof: Assume |X| = n. To form a subset of X, we must decide which elements

are in the subset, i.e. for each element of X we must give a decision “yes” or
“no”. Since there are n elements in X and since for each element there are two
possible decisions “yes” or “no”, there are 2n possible subsets of X. ¤
Exercises.
i. Let X = {0, 1}. List the elements of {℘(A) : A ∈ ℘(X)}.
ii. Let X = ∅. List the elements of {℘(A) : A ∈ ℘(X)}.
iii. Show that ∅ 6= {∅}.
iv. Write the 16 elements of ℘(℘({0, 1})).
v. Show that ℘(∅) = {∅}.
vi. Show that ℘(℘(∅)) = {∅, {∅}}.
vii. Find ℘(℘(℘(∅))).
viii. Find a set X with an element x such that x ∈ X and x ⊆ X.
ix. Can we have x ⊆ {x}? Answer: According to the definition, x ⊆ {x}

if and only if any element of x is an element of {x}, and this means that
any element of x (if any) is x. This implies that either x = ∅ or x = {x}.
This last possibility is quite curious, and, mush later in the book will be
forbidden with the help of an axiom. We do not want a set to have itself
as an element, because we feel that to define and to know a set we need
to know its elements, so that if x ∈ x then to know x we need to know its
element x!
x. Suppose x = {x}. Find ℘(x). Is x ⊆ ℘(x)? Find ℘(℘(x)). Is x ⊆ ℘(℘(x))?
xi. Show that X = ∅ satisfies the following formula: “for every x ∈ X, x is a

subset of X”.
xii. Can you find a set X 6= ∅ such that for every x ∈ X, x is a subset of X?
xiii. Show that X ⊆ Y if and only if ℘(X) ⊆ ℘(Y ).
xiv. Show that if X ⊆ Y then ℘(X) ∈ ℘(℘(Y )).
xv. Show that for any set X, {∅} ∈ ℘(X) if and only if ∅ ∈ X.
xvi. Show that for any set X, {∅} ∈ ℘(℘(X)).
xvii. Show that for any set X, {∅} ∈ ℘(℘(℘(℘(℘(X))))).
xviii. Show that for any set X, {X} ∈ ℘(℘(X)).
xix. Show that for any set X, {{∅}, {{X}}} ∈ ℘(℘(℘(℘(X)))).

1.2. NOTES ON FORMALISM 15
xx. Show that {℘(A) : A ⊆ X} ∈ ℘(℘(℘(X))).

xxi. Show that {{x}, {x, y}} = {{z}, {z, t}} if and only if x = z and y = t.
∗
xxii. Show that any element of ∅ is ∅. (Hint: See Theorem 1.1.1 and its
proof).
∗
xxiii. Have a philosophical discussion among your friends about whether we
should allow the existence of a set x such that x = {x}.
∗
xxiv. Have a philosophical discussion among your friends about whether we
should allow the existence of two sets x and y such that x ∈ y and y ∈ x.
1.2 Notes on Formalism

Let us look at the definition of “subset” once again: x ⊆ y if and only if every
element of x is an element of y. Here, “if and only if” means that
if x ⊆ y then every element of x is an element of y,
and
if every element of x is an element of y then x ⊆ y.
Some abbreviate the phrase “if and only if” by the symbol ⇔ and write the
sentence above in the following abbreviated form:
x ⊆ y ⇔ every element of x is an element of y,
although it is a bad taste to use such symbols in print, or even in classroom or
exams. Let us continue to abbreviate sentences in this way. Now consider the
phrase
every element of x is an element of y.
We can rephrase this as
for all z, if z is an element of x then z is an element of y
and then as follows:
for all z, if z ∈ x then z ∈ y.
Very often, mathematicians use the symbol
Rightarrow for “if . . . then . . .”, i.e. instead of “if p then q” they write p ⇒ q.
With this formalism, we can write the above phrase as
for all z, z ∈ x ⇒ z ∈ y.
That is not enough: “for all z” is abbreviated as ∀z. Now the phrase
every element of x is an element of y
becomes
∀z (z ∈ x ⇒ z ∈ y).
Finally, the definition of “subset” can be written as
x ⊆ y ⇔ ∀z (z ∈ x ⇒ z ∈ y).
What we have above is really a short cut for the English sentence “x is a subset
of y if and only if every element of x is an element of y”, it is some kind of steno, a
shorthand, as useful as “TGIF”. Although it is convenient, we will refrain from
using such abbreviations without any apparent reason, and we advice the young
reader not to use such symbolism in their papers, unless they are asked to do
so.
Since the above definition is for all sets x and y, we can write,
∀x ∀y (x ⊆ y ⇔ ∀z (z ∈ x ⇒ z ∈ y)).
The fact that every set is a subset of itself is formalized by ∀x x ⊆ x.
Let us formalize the equality of two sets. As we know, two sets x and y are
equal if and only if x ⊆ y and y ⊆ x. Thus
x = y ⇔ (x ⊆ y and y ⊆ x).
(The reader should note the need of parentheses). Instead of the word “and”
we will write ∧, thus we now have
x = y ⇔ (x ⊆ y ∧ y ⊆ x).
We can continue and replace x ⊆ y with ∀z (z ∈ x ⇒ z ∈ y) and y ⊆ x with
∀z (z ∈ y ⇒ z ∈ x). Thus
x = y ⇔ ((∀z (z ∈ x ⇒ z ∈ y) ∧ (∀z (z ∈ y ⇒ z ∈ x))).
A moment of thought will convince the reader that
(∀z (z ∈ x ⇒ z ∈ y) ∧ (∀z (z ∈ y ⇒ z ∈ x))
is equivalent to
∀z ((z ∈ x ⇒ z ∈ y) ∧ (z ∈ y ⇒ z ∈ x)),
and that this one is equivalent to
∀z (z ∈ x ⇔ z ∈ y).
Thus
x = y ⇔ ∀z (z ∈ x ⇔ z ∈ y).
Since this is true for all sets x and y, we have
∀x ∀y (x = y ⇔ ∀z (z ∈ x ⇔ z ∈ y)).
Let us look at the meaning of x 6= y. By definition, x 6= y if it is not true
that x = y, i.e. if there is an element in one of the sets which is not in the
other. Thus, x 6= y if and only if
1.3. NUMBER SETS 17
either there is an element in x which is not in y

or
there is an element in y which is not in z.
The first one, namely “there is an element in x which is not in y” means
there is a z ∈ x such that z 6∈ y,
i.e.,
there is a z such that z ∈ x ∧ z 6∈ y.
We abbreviate “there is a z” by the symbols ∃z. Thus the phrase “there is an
element in x which is not in y” is formalized by
∃z(z ∈ x ∧ z 6∈ y).
Now x 6= y is formalized by
either ∃z(z ∈ x ∧ z 6∈ y) or ∃z(z ∈ y ∧ z 6∈ x).
Finally, the ”either ... or ...” part is shortened by . . .∨. . ., so that x 6= y becomes
equivalent to
(∃z(z ∈ x ∧ z 6∈ y)) ∨ (∃z(z ∈ y ∧ z 6∈ x)).
A moment of reflection will show that this is equivalent to
∃z((z ∈ x ∧ z 6∈ y) ∨ (∃z(z ∈ y ∧ z 6∈ x)).
What about the proposition that x ⊂ y, how is it formalized? By definition,

x ⊂ y if and only if x ⊆ y and x 6= y. We know how to formalize x ⊂ y. What
about x 6= y? By definition, x 6= y if there is an element in one of them
To be completed... This subsection may have to go somewhere else.
1.3 Number Sets

Whole numbers such as 0, 1, 2, 3, 4, 5 are called natural numbers. The set
{0, 1, 2, 3, 4, 5, . . .}
whose elements are natural numbers is denoted by the symbol N. Thus
N = {0, 1, 2, 3, . . .}.
For example 5 ∈ N, but −3 6∈ N, 2/3 6∈ N.

In the next part we will prove using our axioms that N is indeed a set, here
in this part, we do not worry about such questions. Deep questions such as
“what is 0?”, “what is 1?”, “is N a set?” will be asked and answered in the next
part, not in this one.
We can add and multiply two natural numbers to obtain a third natural
number, i.e. “N is closed under addition and multiplication”. On the other
hand substraction is not well-defined in N, because for example 2 − 5 is not a
natural number. To subtract one natural number from another we need the
negative of natural numbers as well.
Any natural number is an integer, as well as −3, −2, −1, which are not
natural numbers. The set of integers is denoted by Z. Thus
Z = {. . . , −3, −2, −1, 0, 1, 2, 3, . . .}.
We have −3, 5 ∈ Z, but 3/5 6∈ Z. Clearly N ⊂ Z. We can add and multiply

two integers, or subtract one integer from another to obtain a third integer. In
other words, the set Z is closed under addition, multiplication and substraction.
On the other hand division is not well-defined in Z, because for example 2/5 is
not an integer. To divide one integer to another (nonzero) integer, we need to
extend the number system to the set of rational numbers, where one can divide
a number by a nonzero number.
Numbers such as 3/5, −10/7 are called rational numbers. Thus every
rational number is of the form a/b for some a and b ∈ Z. But we have to
assume that b 6= 0, because we cannot divide by 0, it is forbidden. The set of
rational numbers is denoted by Q. Thus
Q = {a/b : a, b ∈ Z and b 6= 0}.
Now in the set Q we can add, multiply, subtract any two numbers. We can
also divide one rational number to another nonzero rational number. Thus Q
is closed under addition, multiplication, substraction, and division by nonzero
elements.
By taking b = 1, we see that Z ⊂ Q.
Although the set of rational numbers looks like a perfect set to work with,
it is not, because, there are “holes” in Q, for example, one cannot solve the
equation x2 = 2 in Q as the next lemma shows:
Lemma 1.3.1 There is no q ∈ Q such that q 2 = 2.
Proof: Assume not. Let q ∈ Q be such that q 2 = 2. Let a, b ∈ Z be such that

q = a/b. Simplifying if necessary, we may choose a and b so that they are not
both divisible by 2. From a2 /b2 = (a/b)2 = q 2 = 2 we get a2 = 2b2 . Thus a2 is
even. It follows that a is even (because the square of an odd number is always
odd). Let a1 ∈ Z be such that a = 2a1 . Now 4a21 = (2a1 )2 = a2 = 2b2 and
2a21 = b2 . Hence b is even as well, a contradiction. ¤
Thus 21/2 6∈ Q and Q is not closed under taking (rational) powers. To
overcome this difficulty with the rational numbers, we have to extend the set Q
of rational numbers to a larger set called the set of real√numbers and denoted
by R. All rational numbers are real numbers. But also 2, π, −π π , (π 2 + π)2/3
are real numbers. The intuitive feeling of the concept of real numbers is much
less obvious and much weaker than that of the other numbers defined previously,
1.3. NUMBER SETS 19
in fact ancient Greeks had some considerable difficulty before they became aware
and recognized irrational numbers. In fact even after the Greeks became aware
of the irrational numbers, they had a hard time to admit their existence, for
example the Pythogarian school forbid its members from revealing it. Later on
in this book, we will define each one of these concepts mathematically. Loosely
speaking, we will obtain the set of real numbers by filling the “gaps” in Q. But
for the moment only an intuitive feeling is enough. The following may help the
reader to gain further intuition: A positive real number can be regarded (or
defined) as a distance on a straight line.
Although we know it is not and cannot be the case for most of our readers,
in this part we assume that the reader is familiar with the set R of real numbers
and the four operations performed with these numbers.
In R we can take √ any power of a nonnegative
√ number.√ Unlike in Q, there
2
is a number, called 2, in R such that 2 > 0 and ( 2) = 2. We also
√ √2 π
have real numbers of the form 2 and π . These operations will be defined
mathematically in the next part.
We of course have, N ⊂ Z ⊂ Q ⊂ R.
In between Z and R there are other number sets closed under addition,
substraction and multiplication. For example, the set
√ √
Z[ 2] := {a + b 2 : a, b ∈ Z}
is √
closed under addition, substraction and multiplication. On the other hand
Z[ 2] is not closed under division (by a nonzero element). But the set
√ √
Q[ 2] := {a + b 2 : a, b ∈ Q}
is closed under division (by a nonzero element). We leave this as an exercise to

the reader.
Exercises and Examples.
i. Show that between any two rational numbers there is a rational number.
ii. Show that there is no smallest rational number > 0.
iii. Show that there are positive irrational numbers α and β such that αβ
√ √2
is rational. (Hint: If 2 is rational, we are done. Otherwise consider
√ √2 √2
[ 2 ] .)
iv. Let ² > 0 and α > 0 be rational numbers. Show that there is an n ∈ N
such that n² > α.
v. Show that any nonempty subset of Z closed under substraction is the set
of multiples of a given natural number n.
1.4 Subsets Defined By a Property

Given a set X, we may want to consider the collection of elements of X with
a certain property. For example, we may take take X to be N and consider
the collection of elements of N with the property of “being even”. This new
collection is “of course” a subset of N. It is the set of all even natural numbers.
If we symbolize the property by P and abbreviate the fact that x has property
P by P (x), then we denote this set by
{x ∈ X : P (x)}.
For example, in the example above, the set of even natural numbers we obtained
is written by
{x ∈ N : x is even}.
We may denote the same set as
{2x : x ∈ N}.
We will abbreviate this set as 2N. Similarly, 2Z will denote the set of even
integers. For r ∈ R, the definitions of the sets rN, rZ, rQ should be clear. For
example,
rQ := {rq : q ∈ Q} = {s ∈ R : s = rq for some q ∈ Q}.
Intervals. For a real number a, consider the set
{s ∈ R : s > a}.
This set is denoted by (a, ∞). Similarly for two real numbers a and b we define
the following sets:
(a, b) = {x ∈ R : a < x < b}

(a, b] = {x ∈ R : a < x ≤ b}
[a, b) = {x ∈ R : a ≤ x < b}
[a, b] = {x ∈ R : a ≤ x ≤ b}
(a, ∞) = {x ∈ R : a < x}
[a, ∞) = {x ∈ R : a ≤ x}
(−∞, a) = {x ∈ R : x < a}
(−∞, a] = {x ∈ R : x ≤ a}
(−∞, ∞) = R
These sets are called intervals6 . We also denote the intervals (0, ∞) and [0, ∞)
by R>0 and R≥0 respectively.
6 The reader should be aware that we did not define an object called “infinity” or ∞ here.
We just denoted some subsets with the help of the symbol ∞. We did not use the symbol ∞
in the definition of these subsets, we just used it in the names of the subsets as a meaningless
symbol.
1.5. SETS OF SETS 21
Exercises.
i. Show that if b ≤ a then (a, b) = ∅.
ii. Find the elements of the set {x ∈ R : x > n for all n ∈ N}.
iii. Find the set {x ∈ R : x2 ∈ (0, 1)}.
1.5 Sets of Sets

We recall that the set of all subsets of a set X is denoted by ℘(X). Here, in
this section, we will see some other examples of sets whose elements are sets.
There will be no theorems or results, just a few examples to make the reader
more comfortable with large sets of sets. All our examples will be subsets of
℘(X) for some set X defined by a certain property as in the last section.
Examples.
i. The collection of all sets of the form (r, ∞) for r ∈ R is a set. We denote
it as {(r, ∞) : r ∈ R}. Naming X this set, we have
(5, ∞) ∈ X
7 6∈ X
7 ∈ (5, ∞) ∈ X
∅ 6∈ X
R 6∈ X
[5, ∞) 6∈ X
ii. For a natural number n, define An := {n, n + 1, . . . , 2n}. Thus
A0 = {0}
A1 = {1, 2}
A2 = {2, 3, 4}
A3 = {3, 4, 5, 6}
Consider the set whose elements are all these sets A0 , A1 , A2 , A3 , . . . This
set is denoted as
{An : n ∈ N}.
iii. We may consider the set of subsets X of R such that (−², ²) ⊆ X for some
² ∈ R>0 . This set will be written as
{X ⊆ R : for some ² ∈ R>0 , (−², ²) ⊆ X}.

Exercises.
i. Let X be a finite set with n elements. Find the number of elements of the
set {A ∈ ℘(X) : |A| = n − 1}.
ii. Let X be a finite set with n elements. Find the number of elements of the
set {A ∈ ℘(X) : |A| is even}.
iii. Let X be a finite set with n elements. Find the number of elements of the
set {A ∈ ℘(X) : |A| is divisible by 3}.
iv. Let U be the set of subsets X of R such that for some ² ∈ R>0 , (−², ²) ⊆ X.
Let V be the set of subsets X of R such that for some ² ∈ R>0 , [−², ²] ⊆ X.
Let W be the set of subsets X of R such that for some ² ∈ Q>0 , (−², ²) ⊆
X.
Show that U = V = W .
v. A subset U of R is called open if for any x ∈ U there is an ² ∈ R>0 such
that (x − ², x + ²) ⊆ U .
Show that a subset U of R is open if for any x ∈ U there is an ² ∈ R>0
such that [x − ², x + ²] ⊆ U .
Show that ∅ is open.
Show that the intersection7 of two open subsets is open.
Show that if U 6= ∅, R is open, then {x ∈ R : x 6∈ U is not open.
vi. A subset U of Q is called open if it is the intersection with Q of an open
subset of R. Show that Q can be written as the union of two disjoint
nonempty open subsets. Try to convince yourself that R does not have
this property.
1.6 Parametrized Sets

The two sets defined in Examples i, page 21 and ii, page 21 are parametrized.
For example the set (in fact the elements of the set) {An : n ∈ N} is (are)
parametrized by natural numbers n ∈ N.
Let us consider another example: Let X be any set. Consider the set Y
whose elements are the set of all subsets of subsets of X. Thus each element of
Y is of the form ℘(A) for some subset A of X, i.e.
Y = {℘(A) : A ⊆ X} = {℘(A) : A ∈ ℘(X)}.
Here we parametrized the elements of Y by the elements of ℘(X).

If X is any set, then setting Bx = x, we see that X = {Bx : x ∈ X} and so
every set is parametrized by itself!
7 Term to be defined.
1.6. PARAMETRIZED SETS 23
It is also possible to parametrize a singleton set {a} by natural numbers.

Indeed, for all n ∈ N, define Cn = a. Then {a} = {Cn : n ∈ N} and the
singleton set is parametrized by the infinite set N.
Another mathematical jargon for “parametrized” is “indexed”. If X is a set
indexed by I, we say that I is the index set of X although the set I is far from
being unique.
Sometimes, we may need two or more parameters. For example consider the
set
{(r, s) : r < 1 and s > 2}
of intervals. Here, each element of the set is parametrized by two real numbers
r and s.
Sometimes, a set {xi : i ∈ I} indexed by I is denoted by (xi )i∈I or by (xi )i
when the context is clear enough. Sometimes the indexed set (xi )i is called a
family.
Chapter 2
Operations with Sets
2.1 Difference
Let X and Y be two sets. We want to consider the set of elements of X which
are not in Y , i.e. we want to consider the set {x ∈ X : x 6∈ Y }. This is the set
of elements x of X with the property that x 6∈ Y . We denote this set by X \ Y .
For example {0, 2, 3, 5} \ {2, 5, 7, 8} = {0, 3} and N \ 2N is the set of odd
natural numbers.
Clearly, for all sets x and y,
i. x\x=∅
ii. x \ y ⊆ x
iii. x \ ∅ = x
iv. x \ y = ∅ if and only if x ⊆ y
If a set X is fixed once for all, for a subset Y of X, we write Y c for X \ Y .
Whenever we use the notation Y c , we assume that the superset X is clear from
the context. The set Y c is called the complement of Y (in X).
Given a superset X, it is again clear that
i. (Y c )c = Y
ii. X c = ∅
iii. ∅c = X
2.2 Intersection
Given two sets X and Y we may want to consider the set of common elements
of X and Y . This new set is called the intersection of X and Y and is
denoted X ∩ Y . For example if X = {0, 1, 2, 5, 8} and Y = {1, 3, 5, 6} then
X ∩ Y = {1, 5}.
Two sets that intersect in emptyset are called disjoint.
The following properties are well-known and we leave the proofs to the
reader:
25
26 CHAPTER 2. OPERATIONS WITH SETS
i. x∩x=x
ii. x∩y ⊆x
iii. x∩y =y∩x
iv. x ∩ y = x if and only if x ⊆ y
v. (x ∩ y) ∩ z = x ∩ (y ∩ z)
vi. x∩∅=∅
vii. (x \ y) ∩ (y \ x) = ∅
Property (v) says that there is no need of parentheses when intersecting

several sets. For example the sets x ∩ (y ∩ (z ∩ t)) and (x ∩ y) ∩ (y ∩ t) are equal
and they may be denoted by x ∩ y ∩ y ∩ t.
We can also intersect infinitely many subsets. For example, we may want to
consider all the commonTelements of the intervals (−1/n,
T∞ 1/n) for n = 1, 2, 3, . . .
We denote this set as n=1,2,... (−1/n, 1/n) or as n=1 (−1/n, 1/n); it is the
intersection of all the sets of the form (−1/n, 1/n) for T∞some nonzero natural
number n. Incidentally, the reader should note that n=1 (−1/n, 1/n) = {0}.
Similarly
T∞
Tn=1 [ − 1/n, 1/n) = {0}
∞
Tn=1 [ − 1/n, 1/n] = {0}
∞
Tn=1 [ − 1/n, ∞) = [0, ∞)
∞
Tn=1 (−1/n, ∞) = [0, ∞)
∞
Tn=1 (0, 1/n) =∅
∞
Tn=1 (0, 1/n] = ∅
∞
n=1 [0, 1/n] = {0}
Another example: We may T want to intersect all the sets of form

T (r, ∞) for
r ∈ R. We T denote this set as r∈R (r, ∞). It should be clear that r∈R (r, ∞) =
∅. Also, n∈N (n, ∞) = ∅.
In general if X = {Ai : i ∈ I} is a set indexed by the index set ITand if each
Ai is a set, then the intersection of all the sets in X is denoted by i∈I Ai .
If the set X is not given by
T an index set as above, we denote the intersection
of all the elements of X by x∈X x or sometimes by ∩X. Thus we have
y ∈ ∩X if and only if y ∈ x for all x ∈ X.
For reasons that will be clear in the next part, ∩∅ is left undefined for the
moment. The reader should only note for the moment that if in the definition
y ∈ ∩X if and only if y ∈ x for all x ∈ X
we take X = ∅, then we get
y ∈ ∩∅ if and only if y ∈ x for all x ∈ ∅
and, since the statement on the right hand side holds for all y, every y is in ∩∅.
(See also Exercise viii, page 28).
2.3. UNION 27
Exercises.
i. Show that for any set X, ∩℘(X) = ∅.
ii. Show that x \ y = x if and only if x ∩ y = ∅ and that x \ y = ∅ if and only
if x ⊆ y.
T∞
iii. Show that n=1 (−1/n, 1 + 1/n) = [0, 1].
T
iv. Show that r<1 and 2<s (r, s) = [1, 2].
T T
v. Show that ²>0 (−², ²) = ²>0 [−², ²] = {0}.
2.3 Union
Given two sets x and y we may want to consider the set of elements which
are either in x or in y. This new set is called the union of x and y and is
denoted x ∪ y. For example if x = {0, 1, 2, 5, 8} and y = {1, 3, 5, 6} then
x ∪ y = {0, 1, 2, 3, 5, 6, 8}.
We leave the proof of the following well-known properties to the reader:
i. x∪x=x
ii. x⊆x∪y
iii. x∪y =y∪x
iv. x∪y =x⇔y ⊆x
v. (x ∪ y) ∪ z = x ∪ (y ∪ z)
vi. x∪∅=x
Property (v) says that the parentheses are not needed when intersecting several
sets. For example the sets x ∪ (y ∪ (z ∪ t)) and (x ∪ y) ∪ (y ∪ t) are equal and
they may be denoted by x ∪ y ∪ y ∪ t.
There are two famous relationships between ∩ and ∪ called De Morgan laws:
vii. x ∩ (y ∪ z) = (x ∩ y) ∪ (x ∩ z)
viii. x ∪ (y ∩ z) = (x ∪ y) ∩ (x ∪ z)
We can also take the union of infinitely many subsets. For example, we may
want to consider all the common Selements of the intervals (1/n,
S∞1 − 1/n) for n =
1, 2, 3, . . . We denote this set as n=1,2,... (1/n, 1−1/n) or as n=1 (1/n, 1−1/n);
it is the union of all the sets of the form (1/n, 1−1/n) for some nonzero natural
number n. The reader should prove that
∞
[
(1/n, 1 − 1/n) = (0, 1).
n=1
Similarly S∞
Sn=1 [1/n, 1 − 1/n) = (0, 1)
∞
Sn=1 [1/n, 1 − 1/n] = (0, 1)
∞
Sn=1 [1/n, ∞) = (0, ∞)
∞
n=1 (1/n, ∞) = (0, ∞)
Similarly, we may want to take S the union of all theS sets of form (r, ∞) for
r ∈ R>0 . We denote this set as r∈R S >0 (r, ∞) or by r>0 (r, ∞) if there is no
room for confusion. It is clear that r>0 (r, ∞) = (0, ∞).
In general if X = {Ai : i ∈ I} is a set indexed by the indexSset I and if each
Ai is a set, then the union of all the sets in X is denoted by i∈I Ai .
If the set X is not given
S by an index set as above, we denote the union of
all the elements of X by x∈X x or sometimes by ∪X. Thus we have
y ∈ ∪X if and only if there is an x ∈ X such that y ∈ x.
Lemma 2.3.1 ∪∅ = ∅.
Proof: Since the proposition “y ∈ ∪X if and only if y ∈ x for some x ∈ X”

holds for all sets X, applying this to the particular case X = ∅, we get “y ∈ ∪∅
if and only if y ∈ x for some x ∈ ∅”. Since there is no element x in ∅, there is
no such y. Hence ∪∅ = ∅. ¤
Exercises.
i. For subsets A and B of a superset X, show that (A \ B)c = Ac ∪ B.
ii. Show that for any sets x, y, z,
y ∩ (x \ y) = ∅
(x ∩ y) \ z = (x \ z) ∩ (y \ z)
(x \ y) \ z = x \ (y ∪ z)
iii. Let X 6= ∅ be a set. Show that ∩X ⊆ ∪X.

S
iv. Find n∈N [n, n2 ].
S
v. Find n∈N (n, n2 ).
S T
vi. Let XTbe any set
S and for i ∈ I let Yi ⊆ X. Show that ( i Yi )c = i Yic
and ( i Yi )c = i Yic .
vii. Show that for any set X, ∪℘(X) = X.
viii. For a set X, suppose we define ∩X (changing the previous definition) as
follows:
y ∈ ∩X if and only if y ∈ ∪X and y ∈ x for all x ∈ X.
Show that this coincides with the earlier definition if X 6= ∅ and that with
the new definition, ∩∅ = ∅.
S
ix. Show that r>0 [r, ∞) = (0, ∞).
S S
x. Show that for any set A and for any set X of sets, ( x∈X x)\A = x∈X (x\
A).
2.4. CARTESIAN PRODUCT OF TWO SETS 29
xi. Let X and Y be two nonempty sets of subsets. Show that

¡S ¢ ³S ´ S
x∈X x ∩ y∈Y y = (x ∩ y)
¡T ¢ ³T ´ Tx∈X, y∈Y
x∈X x ∪ y∈Y y = x∈X, y∈Y (x ∪ y)
xii. Symmetric Difference. For two sets X and Y define the symmetric
difference of X and Y as follows: X∆Y = (X \ Y ) ∪ (Y \ X). Show that
for all sets X, Y, Z,
a) X∆(Y ∆Z) = (X∆Y )∆Z.
b) X∆Y = Y ∆X.
c) X∆∅ = ∅∆X = X.
d) X∆X = ∅.
2.4 Cartesian Product of Two Sets

We expect the reader knows from high school the parametrization of the plane
R × R as pairs of real numbers (x, y): Any point P of the plane R × R is
represented by two coordinates x and y and one writes P = (x, y).
What is the pair (x, y)? What do we really mean by this notation? What is
the mathematical object (x, y)? This is the subject of this subsection.
What counts in pure mathematics is not so much how the pair (x, y) is
defined, but its properties that we need. From the pair (x, y) the only property
we need is the following:
(x, y) = (z, t) if and only if x = z and y = t.
Accordingly, we would like to define the pair (x, y) in such a way that the above
property holds.
The set {{x}, {x, y}} has this property, i.e.,
{{x}, {x, y}} = {{z}, {z, t}} if and only if x = z and y = t.
(See Exercise xxi, page 15).

Accordingly, we define the pair (x, y) as the set {{x}, {x, y}}.
The “first” element x of the pair (x, y) is called its first coordinate, the
“second” element y of the pair (x, y) is called its second coordinate.
For two sets X and Y , we define X × Y as the set of all pairs (x, y) for x ∈ X
and y ∈ Y . Thus
X × Y = {(x, y) : x ∈ X, y ∈ Y }.
The set X × Y is called the Cartesian product of X and Y (in that order).
Lemma 2.4.1 If X and Y are sets, then X × Y ⊆ ℘(℘(X ∪ Y )).

Proof: Let x ∈ X and y ∈ Y be any two elements. Clearly x, y ∈ X ∪ Y .

Therefore {x} and {x, y} are subsets of X ∪Y , and so {x} and {x, y} are elements
of ℘(X ∪ Y ). It follows that {{x}, {x, y}} is a subset of ℘(X ∪ Y ), and hence
an element of ℘(℘(X ∪ Y )). Therefore X × Y ⊆ ℘(℘(X ∪ Y )). ¤
Corollary 2.4.2 If X and Y are sets, then X × Y is the set of elements α ∈

℘(℘(X ∪ Y )) such that there are x ∈ X and y ∈ Y for which α = {{x}, {x, y}}}.
In other words, the set X × Y is a definable subset of ℘(℘(X ∪ Y )).
Given three sets x, y and z, we can form (x, y) and (y, z) and then ((x, y), z)
and (x, (y, z)). Unfortunately these two sets are not necessarily equal. We let
(x, y, z) denote the first one of these. Later we will give a better definition of
(x, y) and (x, y, z).
Exercises.
i. Write the elements of ((x, y), z).

ii. Can we ever have ((x, y), z) = (x, (y, z))?
iii. If |X| = n and |Y | = m, how many elements does X × Y have?
iv. Find ∩(x, y), ∪(x, y), ∩ ∩ (x, y), ∩ ∪ (x, y), ∪ ∩ (x, y), ∪ ∪ (x, y).
v. Find ((∪ ∪ (x, y)) \ (∪ ∩ (x, y))) ∪ (∩ ∪ (x, y)).
vi. What is X × ∅?
vii. Show that ∪ ∪ (X × Y ) = X ∪ Y if X 6= ∅ and Y 6= ∅.

viii. Suppose A ⊆ A × A. What can you say about A?
ix. Show that X × Y = ∅ if and only if one of X or Y is empty.
x. Show that (X ∪Y )×Z = (X ×Z)∪(Y ×Z), (X ∩Y )×Z = (X ×Z)∩(Y ×Z)
and (X \ Y ) × Z = (X \ Z) ∩ (Y \ Z).
xi. Find similar equalities for (∪i Xi ) × Z and (∩i Xi ) × Z.

Chapter 3
Functions
3.1 Functions
Loosely and naively speaking, a function or a map f from a set X into a set
Y is a “rule” that assigns to each element x of X a unique element f (x) of Y .
Note that we highlighted the words “each” and “unique”.
For example, the rule that assigns to a real number its square root is not
a function because a negative number does not have a square root. The rule
that assigns the elements x and −x to an element x ∈ R is not a function either
because a function must assign a unique element to each element.
On the other hand, a function may assign the same element of the set of
arrival to two different elements of the domain. For example, to a real number
x we can assign its square x2 and get a function from R into R. Although
the element 1 is assigned to the elements 1 and −1, this does not prevent the
squaring from being a function. As this example also shows, an element of the
set of arrival is not necessarily assigned to an element of the domain, e.g. in this
example, the element −1 is not assigned, since it is not a square.
The set X is called the domain of the function f , and the set Y is called
the arrival set of f .
If f is a function from X into Y , very often we denote this fact as f : X −→
Y . The “rule” f may be made explicit by the notation x 7→ f (x).
If A ⊆ X and f : X −→ Y is a function, we let
f (A) = {f (x) : x ∈ A}
and we call this set the image or the range of A under f . The set f (X) is
called the image or the range of f
Given a function f : X −→ Y and a set Y1 that contains f (X) as a subset,
we can define another function f1 : X −→ Y1 by the rule used to define f . For
each x ∈ X we then have f (x) = f1 (x) by the very definition of f1 . Although
the two functions take the same values with the same input, i.e. although they
are defined by the same rule, their arrival sets are distinct. (But their images
31
32 CHAPTER 3. FUNCTIONS
are the same set f (X)). We will differentiate f and f1 and consider them as
two different functions (but we may denote them by the same letter f ). In other
words, a function is more than a rule; the domain X and the arrival set Y are
also parts of the definition of a function.
Two functions are equal if they have the same domain, the same arrival
set and if they take the same value on any element of the domain. Thus two
functions f : X −→ Y and g : X1 −→ Y1 are equal if and only if X = X1 , Y = Y1
and f (x) = g(x) for all x ∈ X. In particular, two functions f, g : X −→ Y are
equal if and only if f (x) = g(x) for √
all x ∈ X. For example the functions f and
g from R into R defined by f (x) = x2 and g(x) = |x| are equal.
The set of functions from X into Y will be denoted by Func(X, Y ).
Examples and Definitions.
i. Let X be a set. Then there is a function IdX : X −→ X that satisfies

IdX (x) = x. The function IdX : X −→ X is called the identity function.
ii. Let X and Y be two sets. For a fixed element b ∈ Y , the function that
sends each element x of X to the element b of Y is called the constant b
function.
iii. Let X, Y, Z be three sets. Let f : X −→ Y and g : Y −→ Z be two

functions. The function
g ◦ f : X −→ Z
is defined by the rule
(g ◦ f )(x) = g(f (x)).
Note that f ◦ g is not necessarily defined (unless Z = X) and even when

it is, the equality f ◦ g = g ◦ f may not hold.
We could define f ◦ g if g(Y ) were a subset of X, but we will not do it.
The function g ◦ f : X −→ Z is called the composite of f and g. The
operation ◦ is called composition.
If X = Y we may compose f with itself as many times as we wish. We
denote f ◦ . . . ◦ f (n times) by f n . Note that f (n) ◦ f (m) = f (n+m) =
f (m) ◦ f (n) .
iv. Let f : X −→ Y be a function. Let A ⊆ X. Then there is a function

f|A : A −→ Y such that f (a) = f|A (a) for all a ∈ A.
The function f|A is called the restriction of f to A.
v. Let X and Y be two sets. The functions π1 : X × Y −→ X and π2 :

X × Y −→ Y given by π1 (x, y) = x and π2 (x, y) = y are called the first
and second projections respectively.
3.1. FUNCTIONS 33
Lemma 3.1.1 Let X, Y, Z, T be sets and f : X −→ Y , g : Y −→ Z and

h : Z −→ T be three functions. Then,
i. h ◦ (g ◦ f ) = (h ◦ g) ◦ f . (Associativity)
ii. f ◦ IdX = f and IdY ◦f = f .
Proof: i. Let x ∈ X. On the one hand (h ◦ (g ◦ f ))(x) = h((g ◦ f )(x)) =

h(g(f (x))). On the other hand ((h ◦ g) ◦ f )(x) = ((h ◦ g)(f (x)) = h(g(f (x))).
Thus (h ◦ (g ◦ f ))(x) = ((h ◦ g) ◦ f )(x) for all x ∈ X and this means exactly that
h ◦ (g ◦ f ) = (h ◦ g) ◦ f .
ii. Left as an exercise. ¤
If f is a function from X into Y and A ⊆ X, we denote by f (A) the set
values of f at the elements of A. More formally,
f (A) := {y ∈ Y : y = f (a) for some a ∈ A}.
We call f (A) the image of A under f . If B ⊆ Y , f −1 (B) is defined to be the

set f −1 (B) := {x ∈ X : f (x) ∈ B}. It is called the inverse image of B under
f
It is possible that both A ∈ X and A ⊆ X hold, e.g. X may be the set
{0, 1, {0, 1}} and A = {0, 1}, in which case the meaning of f (A) is controversial
depending whether we consider A as an element or as a subset of X. For example
let X be as before, Y = {5, 6, 7} and f : X −→ Y be defined by
f (0) = 5
f (1) = 6
f ({0, 1}) = 7
In this case f (A) = 7 if we consider A as an element of X and f (A) =

{f (0), f (1)} = {5, 6} in case we consider A as a subset of X. Such situations
occur rarely and when they occur we must just be careful.
Given a function f : X −→ Y , we define its graph to be the set of pairs
(x, f (x)) for x ∈ X. More precisely the graph of the function f is
Gph(f ) = {(x, y) ∈ X × Y : y = f (x)}.
The set Gph(f ) has the following properties:

1. Gph(f ) ⊆ X × Y ,
2. For any x ∈ X there is a unique y ∈ Y such that (x, y) ∈ Gph(f ).
Conversely let F be a set that satisfies the two properties above,
1. F ⊆ X × Y ,
2. For any x ∈ X there is a unique y ∈ Y such that (x, y) ∈ F .
Then we can define a function f : X −→ Y such that F = Gph(f ). Indeed,
we can define f : X −→ Y by the rule
f (x) = y if and only if (x, y) ∈ F .
In the next section we will use this fact to give a more mathematical defini-
tion of a function.
Exercises. Below, f denotes a function.
i. How many functions are there from a set of n elements into a set of m
elements?
ii. What can you say about the relationship between f (A \ B) and f (A) \
f (B)?
iii. Show that f (∅) = ∅. (Here ∅ is considered as a subset of the domain).
iv. Let X = Y = N, and let us associate to a natural number x, the natural

number 2x + 1. If we denote by f this function, we have f (x) = 2x + 1.
For example, f (3) = 7, f (f (3)) = f (7) = 15.
a. Find f (N), f (f (N)), f (f (f (N))).
b. Find f (f . . . (f (N)) . . .). (Here there are n f ’s. We denote this set by
f n (N)).
v. Let f : X −→ Y be a function.
i. Show that if (Bi )i is a family of subsets of Y , then
[ [
f −1 ( Bi ) = f −1 (Bi ),
i i
and \ \
f −1 ( Bi ) = f −1 (Bi )
i i
−1 c −1 c
and if B ⊆ Y then f (B ) = f (B) .
S S
Show thatTif (Ai )i is a family of subsets of X, then f ( i Ai ) = i f (Ai ),
ii. T
f ( i Ai ) ⊆ i f (Ai ).
iii. Find an example where the last inclusion fails to be an equality.
iv. If A ⊆ X what is the relationship between f (Ac ) and f (A)c ?
vi. a) Find a function f : ℘(N) −→ N such that if x ∈ ℘(N) \ {∅} then

f (x) ∈ x.
b) Find a function f : ℘(Z) −→ Z such that if x ∈ ℘(Z)\{∅} then f (x) ∈ x.
c) Find a function f : ℘(Q) −→ Q such that if x ∈ ℘(Q) \ {∅} then
f (x) ∈ x.
d) Find a function f : ℘(R) −→ R such that if x ∈ ℘(R) \ {∅} then
f (x) ∈ x. (Don’t even try!)
vii. Let X and Y be two sets. Find a set Z such that the map f : ℘(X) ×
℘(Y ) −→ Z defined by f (A, B) = A × B is a function. (One may just take
Z := {A × B : A ∈ ℘(X), B ∈ ℘(Y )}. But we want something better. Try
to define Z in terms of X and Y using only the set theoretic operations
like ∪, ∩, ℘).
3.2. MORE ON FUNCTIONS 35
3.2 More On Functions

To define a function f : X −→ Y , we said that we need the two sets X and Y
and a “rule” that assigns to each element of X a unique element of Y . This is
a very loose definition, not acceptable by mathematical standards, because we
did not define what we mean by a “rule”. In fact, a function is not defined the
way we defined it in the previous section. Mathematically, a function is a set of
the form (X, Y, F ) such that1 X and Y are two sets and F is a subset of X × Y
that satisfies the following property:
For any x ∈ X there is a unique y ∈ Y such that (x, y) ∈ F .
Then our “rule” f (x) = y is defined by
f (x) = y if and only if (x, y) ∈ F .
Exercises.
i. With the new definition of a function given above, given two functions
(X, Y, F ) and (Y, Z, G) define their composite (X, Y, F ) ◦ (Y, Z, G).
ii. Let (X, Y, F ) be a function. Show that the set X is uniquely determined
by Y and F . In other words if (X, Y, F ) and (X1 , Y, F ) are functions,
then X = X1 . This exercise shows that a function may be defined only
as a pair (Y, F ) such that for every x there is a unique y ∈ Y such that
(x, y) ∈ F .
3.3 Binary Operations

Let X be a set. A function f : X × X −→ X is sometimes called a binary
operation.
J For x, y ∈ X, instead of f (x, y) it is customary to write x ? y, x · y,
x + y, x y, x ⊗ y etc. or even as xy. If it does not have a specific name, the
outcome x ? y may be called the product of the two elements x and y. A set
X with a binary operation ? on it is denoted by (X, ?) most of the time.
Examples.
i. Let X = N, Z, Q or R. The addition and multiplication are binary oper-

ations on X. If X = Z, Q or R then substraction − is a binary operation
on X. If X = Q \ {0} or R \ {0} then division is a binary operation on X.
ii. Let X be a set. Then ∩, ∪, \, ∆ are binary operations on ℘(X).

1 Recall that (X, Y, F ) stands for ((X, Y ), F ).
Potential Properties of Binary Operations. Let ? be a binary relation

defined on a set X.
i. Associativity. If (x ? y) ? z = x ? (y ? z) for all x, y, z ∈ X, then we

say that ? is associative. Associativity means that the parentheses are
useless when taking the product of several elements. If associativity does
not hold, even the products (x ? x) ? x and x ? (x ? x) may be different, in
which case we may have difficulties defining x3 .
When a binary operation is associative, the product of three (or even
more) elements x, y and z (in that order) may be denoted as x ? y ? z
without parentheses.
Almost all the interesting mathematical binary operations are associative
and if they are not, most of the time there is a close relationship between
(x ? y) ? z and x ? (y ? z).
ii. Commutativity. If x ? y = y ? x for all x, y ∈ X, then we say that ? is

commutative.
iii. Identity Element. If e ∈ X is such that x ? e = x for all x ∈ X, then

we say that e is a right identity element for ?. Left identity is defined
similarly. An element which is both left and right identity element is
called an identity element of the binary operation ?.
If e is a left identity and f is the right identity, then e = f , because
f = e ? f = e. It follows that if X has a left and right identity element,
then it has an identity element and this identity element is unique.
iv. Inverse Element. Assume X has an identity element, say e. Let x ∈ X.

If there is a y ∈ X such that x ? y = e, then y is called a right inverse
of x. Left inverse is defined similarly. If ? is associative and if x ∈ X
has both a left inverse y and a right inverse z, then x = z, because,
y = y ? e = y ? (x ? z) = (y ? x) ? z = e ? z = z.
Exercises. For each of the following binary operations determine whether

associativity and commutativity hold and determine the existence of a right or
left identity element, and in which case the existence of right or left inverse.
i. X = ℘(A) for the intersection ∩.
ii. X = ℘(A) for the union ∪.
iii. X = ℘(A) for the difference \.
iv. X = ℘(A) for the symmetric difference ∆.
v. X = N, Z, Q, R for + and ×.
vi. X = N \ {0}, Z \ {0}, Q \ {0}, R \ {0} for the division.

3.4. OPERATIONS WITH FUNCTIONS 37
vii. X = 2Z for +.
viii. X = R, a ∈ R a fixed element and x ? y = x + y − a.
ix. X any set, e ∈ X a fixed element. Define ? as x ? y = e for all x, y ∈ X.
x. X any set. Define ? as x ? y = x for all x, y ∈ X.
xi. X = R and x ? y = max{x, y}.
xii. X = R and x ? y = x − y.
xiii. X = R and x ? y = |x − y|.
3.4 Operations with Functions

In section 3.1 we have seen how to compose two functions when appropriate.
At page 31, we have seen that a function f : X −→ Y gives rise to a function
from ℘(X) into ℘(Y ), that we still denote by f . Here we will see some other
operations on functions, i.e. from one or more functions we will obtain new ones.
Examples and Exercises.
i. Let f and g be two functions from a set X into R. We can define f + g :

X −→ R by the rule (f + g)(x) = f (x) + g(x) for all x ∈ R. We can
also define the function f g : X −→ R by the rule (f g)(x) = f (x)g(x).
More generally if ? is a binary operation on a set Y and f and g are
two functions from X into Y , then we can define f ? g : X −→ Y by
(f ? g)(x) = f (x) ? g(x). Thus the binary operation ? on Y gives rise to
a binary operation on Func(X, Y ), still denoted by ?. Properties of (Y, ?)
reflect to (Func(X, Y ), ?):
a. Show that if (Y, ?) is associative (resp. commutative) then so is (Func(X, Y ), ?).
b. Show that if (Y, ?) has a left (resp. right) identity element then so does
(Func(X, Y ), ?).
c. Show that if (Y, ?) has an identity element and if every element of Y is
invertible, then the same properties hold for (Func(X, Y ), ?).
ii. If f : X −→ X1 and g : Y −→ Y1 are two functions, we can define the

function f × g : X × Y −→ X1 × Y1 by the rule (f × g)(x, y) = (f (x), g(y)).
iii. Any function f : X −→ Y gives rise to a function f˜ : ℘(X) −→ ℘(Y ) by

the rule f˜(A) = f (A) := {f (a) : a ∈ A} for all A ∈ ℘(X).
iv. Any function f : X −→ Y gives rise to a function f˜−1 : ℘(Y ) −→ ℘(X)

by the rule f˜−1 (B) = f −1 (B) := {x ∈ X : f (x) ∈ B} for all B ∈ ℘(Y ).
v. Let f : X −→ X1 and g : Y −→ Y1 be two functions. Assume that

X ∩ Y = ∅. Then we can define the function f ∪ g : X ∪ Y −→ X1 ∪ Y1 ,
the union of f and g, by the rule
½
f (z) if z ∈ X
(f ∪ g)(z) =
g(z) if z ∈ Y
vi. We can generalize the above example. Let f : X −→ X1 and g : Y −→ Y1

be two functions. Assume that for all z ∈ X ∩ Y , f (x) = g(x). Then we
can define the function f ∪ g : X ∪ Y −→ X1 ∪ Y1 , the union of f and g,
by the rule ½
f (z) if z ∈ X
(f ∪ g)(z) =
g(z) if z ∈ Y
vii. We can generalize even further the above example. Let I be an index set.
Let (fi : Xi −→ Yi )i∈I be a family of functions indexed by I. Assume
that for all i, j ∈ I and all x ∈ Xi ∩ Xj , fi (x) = fj (x). Then we can
define the function ∪i∈I fi : ∪i∈I Xi −→ ∪i∈I Yi by the following rule: For
all x ∈ ∪i∈I Xi ,
(∪i∈I fi )(x) = fi (x)
if x ∈ Xi .
Exercises.
i. Let f and g be two functions from a set X into R. Show that if s :

R × R −→ R is given by s(x, y) = x + y, then f + g = s ◦ (f × g) and that
if p : R × R −→ R is given by p(x, y) = xy, then f g = p ◦ (f × g).
3.5 Injections, Surjections, Bijections

3.5.1 Injections
A function f : X −→ Y is called one-to-one or an injection if for all x1 , x2 ∈
X, f (x1 ) = f (x2 ) implies x1 = x2 . That means that if f : X −→ Y is an
injective function, then no two different elements of X can go to the same
element of Y under f . For example the function f : R −→ R defined by
f (x) = x2 is not injective. But the function f : R≤0 −→ R defined by the same
rule f (x) = x2 is injective. These examples show that “to be injective” is a
property that has to do more with the domain than with the arrival set.
Further Examples.
i. If X and Y are nonempty sets and x0 ∈ X is fixed, then the map f :

Y −→ X × Y defined by f (y) = (x0 , y) is one-to-one.
ii. The map f : N −→ N defined by f (n) = n + 1 is one to one.

3.5. INJECTIONS, SURJECTIONS, BIJECTIONS 39
iii. If f1 : X1 −→ Y1 and f2 : X2 −→ Y2 are one-to-one then the function

f1 ×f2 : X1 ×X2 −→ Y1 ×Y2 defined by (f1 ×f2 )(x1 , x2 ) = (f1 (x1 ), f2 (x2 ))
is also one- to-one.
Lemma 3.5.1 i. The composition of two injections is an injection.

ii. If f ◦ g is an injection then g is an injection.
Proof: i. Let f and g be two injections. Assume (f ◦ g)(x1 ) = (f ◦ g)(x2 ), i.e.

assume f (g(x1 )) = f (g(x2 )). Since f is an injection, g(x1 ) = g(x2 ). Since g is
an injection, x1 = x2 .
ii. Assume g(x1 ) = g(x2 ). Applying f we get f (g(x1 )) = f (g(x2 )), i.e.
(f ◦ g)(x1 ) = (f ◦ g)(x2 ). Since f ◦ g is an injection, this implies x1 = x2 . ¤
3.5.2 Surjections
A function f : X −→ Y is called onto or a surjection if for all y ∈ Y there is
an x ∈ X such that f (x) = y, i.e. if f (X) = Y . That means that if f : X −→ Y
is an surjection, then every elements of Y is “attained” by an element of X
under f . For example the function f : R −→ R defined by f (x) = x2 is not
onto because -1 is not attained. But the function f : R −→ R≥0 defined by the
same rule f (x) = x2 is onto. These examples show that “to be surjective” is a
property that has to do more with the arrival set than with the domain.
Further Examples.
i. If X and Y are nonempty sets, then the projection maps π1 : X ×Y −→ X
and π2 : X ×Y −→ Y are surjections. The map π1 is not one to one unless
|Y | = 1.
ii. The map f : N −→ N defined by f (n) = [n/2] where [x] stands for the
integer part of x is onto (but not one to one).
iii. The map f : N −→ N defined by
½ n
2 if n is even
f (n) = n−1
2 if n is odd
is onto.
iv. If f1 : X1 −→ Y1 and f2 : X2 −→ Y2 are onto then the function f1 × f2 :
X1 × X2 −→ Y1 × Y2 defined by (f1 × f2 )(x1 , x2 ) = (f1 (x1 ), f2 (x2 )) is also
onto.
Lemma 3.5.2 i. The composition of two surjections is an surjection.

ii. If g ◦ f is a surjection then g is a surjection.
iii. If f : X −→ Y is any function, then the function f : X −→ f (X) is a
surjection.
iv. The projection map π1 : X × Y −→ X (resp. π2 : X × Y −→ Y ) is a
surjection if Y 6= ∅ (resp. X 6= ∅).
Proof: i. Let f : X −→ Y and g : Y −→ Z be two surjections. Clearly

(g ◦ f )(X) = g(f (X)) = g(Y ) = Z. So g ◦ f is a surjection.
ii. Assume f : X −→ Y and g : Y −→ Z are functions. Assume g ◦ f is a
surjection. Then Z = g(f (X)) ⊆ g(Y ).
iii and iv are clear and are left as exercise. ¤
3.5.3 Bijections
A function f : X −→ Y is called bijection if it is one-to-one and onto. Two
sets X and Y for which there is such a bijection are said to be in one-to-one
correspondence or in bijection.
Lemma 3.5.3 i. The composition of two bijections is a bijection.

ii. If f ◦ g is a bijection then g is an injection and f is a surjection.
iii. If f : X −→ Y is an injection, then f : X −→ f (X) is a bijection.
Proof: Follows from lemmas 3.5.1 and 3.5.2. ¤
Lemma 3.5.4 (Inverse) Let f : X −→ Y be a bijection. Then there is a

unique bijection f −1 : Y −→ X such that for all x ∈ X and y ∈ Y ,
f −1 (y) = x ⇐⇒ f (x) = y.
We have
f ◦ f −1 = IdY and f −1 ◦ f = IdX .
Furthermore if g : Y −→ X satisfies
f ◦ g = IdY or g ◦ f = IdX
then g = f −1 . It follows that (f −1 )−1 = f .
Proof: Left as an exercise. ¤

Given a bijection f : X −→ Y , the bijection f −1 : X −→ Y of the lemma
above is called the inverse of f .
Lemma 3.5.5 i. Let f : X −→ Y be a bijection. Then (f −1 )−1 = f .

ii. Let f : X −→ Y and g : Y −→ Z be bijections. Then (g◦f )−1 = f −1 ◦g −1 .
Exercises.
i. Find all bijections from {0, 1, 2} into {0, 1, 2}. Find all bijections from
{0, 1, 2, 3} into {0, 1, 2, 3}. Find all injections from the set {0, 1, 2} into
the set {0, 1, 2, 3}. Find all surjections from {0, 1, 2, 3} into {0, 1, 2}.
3.6. SYM(X) 41
ii. Let f : X −→ Y be a bijection. Show that for all A, B ⊆ X,

a. f (A ∩ B) = f (A) ∩ f (B)
b. f (A ∪ B) = f (A) ∪ f (B)
c. f (A \ B) = f (A) \ f (B)
d. Show that the induced map f : ℘(X) −→ ℘(Y ) is also a bijection.
iii. Let X be a set. The set of functions from X into {0, 1} is denoted by X 2.
The purpose of this exercise is to show that there is a bijection between
the sets ℘(X) and X 2.
a) For a subset A of X, define fA : X −→ {0, 1} by the rule
½
1 if x ∈ A
fA (x) =
0 if x 6∈ A
X
Thus fA ∈ 2. (The function fA is called the characteristic function
of A).
b) Show that the rule A 7→ φ(A) := fA defines a function φ from ℘(X)
into X 2.
c) Let f ∈ X 2 be a function. Let Xf := {x ∈ X : f (x) = 1}. Show that
the rule f 7→ ψ(f ) = Xf defines a function ψ from X 2 into ℘(X).
e) For A ∈ ℘(X), show that ψ(φ(A)) = A.
A
f) For f ∈ 2, show that ψ(φ(f )) = f .
g) Conclude that the functions A 7→ φ(A) and f 7→ ψ(f ) are bijections
and that they are inverses of each other.
iv. Let X be a set. The set of functions from {0, 1} into X is denoted by 2 X.
Show that there is a bijection between the sets X × X and 2 X.
3.6 Sym(X)
Let X be any set and G = Sym(X), the set of all bijections from X into X.
Consider the composition operation ◦ on Sym(X). The triple (Sym(X), ◦, IdX )
has the following properties:
0. For all f, g ∈ Sym(X), f ◦ g, f −1 , IdX ∈ Sym(X)

1. Associativity. For all f, g, h ∈ Sym(X), (f ◦ g) ◦ h = f ◦ (g ◦ h).
2. Identity Element. For all f ∈ Sym(X), f ◦ IdX = IdX ◦f = f.
3. Inverse Element. For all f ∈ Sym(X), f ◦ f −1 = f −1 ◦ f = IdX .
We call Sym(X) the symmetric group on X. If X = {1, . . . , n}, the set

Sym(X) is simply denoted by Sym(n) and is called the symmetric group on
n letters. The group Sym(n) has n! many elements. Sometimes, composition
will also be called multiplication.
Let us first get acquainted with the groups of the form Sym(n) for n ∈ N.
We consider the following element g of Sym(7):
g(1) = 2
g(2) = 5
g(3) = 3
g(4) = 7
g(5) = 1
g(6) = 6
g(7) = 4
We can represent this element as

µ ¶
1 2 3 4 5 6 7
g= ,
2 5 3 7 1 6 4
where we put the domain on the first line, and the image of the elements of the
domain just below it. We can represent this element also as (1, 2, 5)(3)(4, 7)(6) or
as (1, 2, 5)(4, 7). We will prefer the last representation. The last representation
can be read as follows: “1 goes to 2, 2 goes to 5, 5 goes back to 1; 4 goes to 7,
7 goes back to 4; the rest of the numbers (3 and 6) are fixed”.
Written this way, the element g = (1, 2, 5)(4, 7) may be seen as an element of
Sym(7) as well as an element of Sym(8). If one is careful enough when needed,
this never causes a problem.
It is impossible to represent IdSym(6) with the representation we adopted, so
we will write simply IdSym(6) or just 1 (by abuse of language).
Such a representation is called the cyclic representation of the bijection.
The elements such as (1, 2, 5), (3) and (4, 7) of a cyclic representation are called
cycles. The cycle (1, 2, 3) is a 3-cycle. The cycle (4, 7) is a 2-cycle. The element
(1, 2, 5)(4, 7) of Sym(7) has four cycles, namely (1, 2, 5), (3), (4, 7) and (6). On
the other hand, the element (1, 2, 5)(4, 7) of Sym(8) has five cycles, namely
(1, 2, 5), (3), (4, 7), (6) and (8). The element IdSym(6) or 1 of Sym(6) has six
cycles. But the element 1 of Sym(8) has eight cycles.
Clearly (1, 2, 3) = (2, 3, 1) = (3, 1, 2) 6= (1, 3, 2).
The multiplication (more precisely, the composition) of disjoint cycles is very
simple: One just juxtaposes them, e.g. the multiplication of (1, 2, 5) and (4, 7)
is just (1, 2, 5)(4, 7), or (4, 7)(1, 2, 5). Similarly, the product of the two elements
(1, 5, 7)(3, 4) and (2, 6)(8, 9) is just (1, 5, 7)(2, 6)(3, 4)(8, 9).
If the cycles are not disjoint, then the multiplication (i.e. the product or the
composition) of elements is slightly more complicated. For example to compute
(1, 2, 3)(1, 2, 4, 3, 5) we start from the right and see where 1 goes to: 1 first goes
to 2 (the right cycle) and the left cycle takes 2 to 3. Thus the product starts
as (1, 3, . . .. Now we restart the same procedure with 3. As a result we obtain
(1, 2, 3)(1, 2, 4, 3, 5) = (1, 3, 5, 2, 4). As an exercise the reader should check that
(1, 2, 3, 4)(3, 5, 4) = (1, 2, 3, 5)

(1, 3, 2, 4)(1, 4)(3, 5, 4) = (2, 4)(3, 5)
3.6. SYM(X) 43
The inverse of an element is easy to find. For example (1, 2, 3, 4, 5)−1 =

(1, 5, 4, 3, 2) and ((1, 2, 3, 4)(5, 6, 7))−1 = (1, 4, 3, 2)(5, 7, 6).
We can use the same notation for Sym(N). For example the element g :=
(0, 1)(2, 3)(4, 5)(6, 7) . . . is in Sym(N) and g 2 = IdN . On the other hand the
function represented by (0, 1, 2, 3, 4, 5, . . .) is not in Sym(N) because the number
0 is not in the range and this function is not onto, hence not a bijection. An
infinite cycle in Sym(N) cannot have a beginning, nor an end.
On the other hand the element (. . . , 7, 5, 3, 1, 2, 4, 6, 8, . . .) is in Sym(N).
The reader should not that, we do not see the element (0, 1)(2, 3)(4, 5)(6, 7) . . .
as the product of infinitely many elements (0, 1), (2, 3), (4, 5) etc. We can only
multiply finitely many elements of Sym(X).
The type of the element g of Sym(n) is a sequence of n natural numbers
a1 − a2 − . . . − an where ai is the number of i-cycles in the cyclic representation
of g. For example the type of (1, 2, 5)(4, 7) ∈ Sym(8) is 3 − 1 − 1, the type
(1, 2, 5)(4, 7) ∈ Sym(9) is 4 − 1 − 1, the type of (1, 2, 3)(4, 5, 6) ∈ Sym(6) is
0 − 0 − 2. The following formalism is more convenient and expressive: We
will say that (1, 2, 5)(4, 7) ∈ Sym(8) is of type (1)(2)(3, 4)(5, 6, 7)(8), or even of
type (1, 2)(3, 4, 5) if there is no possible confusion. For example the elements of
Sym(5) of type (1, 2, 3)(4, 5) are:
(1, 2, 3)(4, 5), (1, 3, 2)(4, 5), (1, 2, 4)(3, 5), (1, 4, 2)(3, 5), (1, 2, 5)(3, 4),
(1, 5, 2)(3, 4), (1, 3, 4)(2, 5), (1, 4, 3)(2, 5), (1, 3, 5)(2, 4), (1, 5, 3)(2, 4),
(1, 4, 5)(2, 3), (1, 5, 4)(2, 3), (2, 3, 4)(1, 5), (2, 4, 3)(1, 5), (2, 3, 5)(1, 4),
(2, 5, 3)(1, 4), (2, 4, 5)(1, 3), (2, 5, 4)(1, 3), (3, 4, 5)(1, 2), (3, 5, 4)(1, 2).
Exercises.
i. Show that Sym(n) has n! elements.
ii. Write the elements of Sym(n) for n = 1, 2, 3, 4. Draw the multiplication

table of these groups.
iii. Find elements of each type of Sym(n) for n = 2, 3, 4, 5, 6.
iv. How many types of elements are there in Sym(9) whose square is 1?
v. Show that if g := (. . . , 7, 5, 3, 1, 2, 4, 6, 8, . . .) ∈ Sym(N), then g n 6= 1 for

any n ∈ Z \ {0}.
vi. Let a := (0, 1) ∈ Sym(N). Show that CSym(N) (a) := {g ∈ Sym(N) : ga =

ag} = {g ∈ Sym(N) : g({0, 1}) = {0, 1}}.
vii. Multiply the elements
(0, 1)(1, 2)(3, 4) . . . and (1, 2)(3, 4)(5, 6) . . .

of Sym(n) both ways. Are they equal? How many cycles do the products
have?
viii. Show that it is not true that for every g ∈ Sym(N) there is an h ∈ Sym(N)
such that h2 = g.
∗∗
ix. Is it true that every element of Sym(N) is a product of finitely many
squares?
∗∗
x. Let p be an integer. Is it true that every element of Sym(N) is a product
of finitely many p-th powers?
xi. For an element g ∈ Sym(X), define o(g) to be the least positive natural
number m such that g m = IdX if there is such an m, otherwise define
o(g) = ∞. Thus o(g) = IdX if and only if g = IdX .
Let G = Sym(N). Show that
o(1, 2, 3) = 3
o(1, 2, 3, 4) = 4
o((1, 2, 3)(4, 5)) = 6
o(. . . , 7, 5, 3, 1, 2, 4, 6, 8, . . .) = ∞
For each n = 1, 2, . . . , 12 find max{o(g) : g ∈ Sym(n)}.
3.7 Families, Sequences and Cartesian Products

In Section 1.6, we have defined a family (Xi )i∈I as set indexed by the “index
set” I. We are now in a position to redefine the same concept in a more precise
(mathematical) way. Let X and I be two sets. Let x : I −→ X be a surjection.
Thus
X = {x(i) : i ∈ I}
so that X now becomes a family indexed by I. Sometimes we will write xi
instead of x(i) and (xi )i∈I instead of X.
A sequence from a set X is a subset of X indexed by N, thus a sequence is
just a function x : N −→ X. A sequence usually is written as (xi )i∈N or as (xi )i
for short where xi stands for x(i) for all i ∈ N. We may also write a sequence
as
x0 , x1 , x2 , . . .
which more visual and more convenient.
Let (Xi )i∈I be a family of sets. The set of all sequences x from the set ∪i Xi
such that xi ∈ Xi for all i ∈ I is called the Cartesian
Q product of the family
(Xi )i∈I . We denote the Cartesian Q product by Q i If all the sets Xi are
i∈I X .
equal to a set X, then instead of i∈I Xi we write i∈I X.
We have already defined the Cartesian product of two sets X1 and X2 in
chapter 2.4. We should check that the concept we have defined then coincides
3.7. FAMILIES, SEQUENCES AND CARTESIAN PRODUCTS 45
with the concept we have defined here when I has just two elements. But
thisQ
is not true. The set X1 × X2 defined in chapter 2.4 is not equal to the
set i∈{1,2} Xi that we have defined above. Nevertheless there is a bijection
between them.
Q
Lemma 3.7.1 There is a bijection between X1 × X2 and i∈{1,2} Xi .
Proof: Let α ∈ X1 × X2 . Define φα : {1, 2} −→ X1 ∪ X2 by the rule φα (i) =

πi (α) where Q
π1 and π2 are the two projections of X1 × X2 onto X1 and X2 .
Then φa ∈ i∈{1,2} Xi . We leave to the reader the easy fact that the map
Q
φ : X1 × X2 −→ i∈{1,2} Xi defined by φ(a) = φa is a bijection between the
two sets. ¤
Let I be an index set and (Xi )i∈I and (Yi )i∈I be two Q families.
Q For each
iQ∈ I, let fi : Xi −→ Y
Q i be a function. Then the function f
i∈I i : i∈I Xi −→
Y
i∈I i given by ( f
i∈I i )((x )
i i∈I ) = (f (x ))
i i∈I is called the product of the
functions (fi )i∈I . The product f1 × f2 of two functions f1 : X1 −→ Y1 and
f2 : X2 −→ Y2 is of course defined by (f1 × f2 )(x1 , x2 ) = (f1 (x1 ), f2 (x2 )).
Chapter 4
Relations
4.1 Definitions
Let X be a set. A binary relation R of X is just a subset of X × X. For
x, y ∈ X, rather than writing (x, y) ∈ R, one often writes xRy. The usual
symbols for relations are <, ≤, ≺, ¹, ¿, ⊂, ⊆, v, ∼, ', ≈, ≡, ⊥, k, C, E etc.
Examples.
i. Let X = {0, 1, 2, 3}. Define
< = {(0, 1), (0, 2), (1, 2), (0, 3), (1, 3), (2, 3)}.
Then < is a relation. It is the order relation that we are familiar with.
ii. Let X be any set. Then ∅ and X × X are binary relations on X. In

the first one no two elements of X are related, in the second one any two
elements of X are related.
iii. Let X be any set. Then δ(X × X) := {(x, y) ∈ X × X : x = y} is a binary

relation on X (called equality!).
iv. Let X = N. Then {(x, y) ∈ N × N : x divides y in N} is a binary relation

on N.
v. Let U be any set. Let X be the set of finite subsets of U . Then {(A, B) ∈
X × X : |A| = |B|} is a relation on X.
vi. Let U be any set. Let X = ℘(U ). Then
{(A, B) ∈ X × X : there is a bijection f : A −→ B}
is a relation on X.
47
48 CHAPTER 4. RELATIONS
vii. Let U be any set. Let X = ℘(U ). Then
{(A, B) ∈ X × X : there is an injection f : A −→ B}
is a relation on X.
viii. Let U be any set. Let X = ℘(U ). Then {(A, B) ∈ X × X : A ⊆ B} is a

relation on X.
ix. Let U be any set. Let X = ℘(U ). Then {(A, B) ∈ X × X : A ∩ B 6= ∅} is

a relation on X.
x. Let U be any set. Let X = ℘(U ). Then {(A, B) ∈ X × X : A ∩ B = ∅} is

a relation on X.
xi. If f : X −→ X is a function, the graph of f is a relation on X.
Potential Properties of Relations. i. A binary relation R is called re-

flexive if xRx for all x ∈ X.
ii. A binary relation R is called irreflexive if ¬xRx for all x ∈ X.
iii. A binary relation R is called symmetric if xRy implies yRx for all
x, y ∈ X.
iv. A binary relation R is called antisymmetric if xRy implies ¬yRx for
all x, y ∈ X.
v. A binary relation R is called transitive if xRy and yRz implies xRz for
all x, y, z ∈ X.
Exercises.
i. Determine whether or not the binary relations defined in the examples

above are reflexive, irreflexive, symmetric, antisymmetric or transitive.
ii. Let R be a binary relation on a set X. Show that there is a smallest

reflexive relation S that contains R.
iii. Let R be a binary relation on a set X. Show that there is a smallest

symmetric relation S that contains R.
iv. Let R be a binary relation on a set X. Show that there is a smallest

transitive relation S that contains R.
4.2 Equivalence Relations

An equivalence relation is a reflexive, symmetric and transitive binary rela-
tion.
An equivalence relation is very often denoted by one of the following symbols:
≡, ∼, ', ≈, ∼
=.
4.2. EQUIVALENCE RELATIONS 49
According to the definition, for a binary relation ≡ on a set X to be an

equivalence relation, the following should hold:
EqR1. Reflexivity. For all x ∈ X, x ≡ x.
EqR2. Symmetry. For all x, y ∈ X, if x ≡ y then y ≡ x.
EqR3. Transitivity. For all x, y, z ∈ X, if x ≡ y and y ≡ z then x ≡ y.
Examples and Exercises.
i. The equality is an equivalence relation on any set.
ii. The relation R defined by xRy for all x, y on any set is an equivalence
relation.
iii. On any subset of R the relation defined by xRy ⇐⇒ x2 = y 2 is an

equivalence relation.
iv. Let n > 0 be any natural number. On Z, the relation ≡n defined by
x ≡n y ⇐⇒ n divides x − y
is an equivalence relation.
v. On the set R or Q the relation ≡ defined by x ≡ y if and only if x − y ∈ Z

vi. On the set R the relation ≡ defined by x ≡ y if and only if x − y ∈ Q is

an equivalence relation.
vii. Let X and Y be two sets. Let A ⊆ X. The relation ≡A defined on

Func(X, Y ) by
f ≡ g ⇐⇒ f|A = g|A
viii. On X := R × R × R, the relation ≡ defined by (x, y, z) ≡ (x1 , y1 , z1 ) by

2(x − x1 ) + 3(y − y1 ) − (z − z1 ) = 0 is an equivalence relation.
ix. On C, the relation defined by
x ≡ y ⇐⇒ |x| = |y|
x. More generally, if X and Y are sets and F is a set of functions from X

into Y , then the relation ≡ defined on X by
x ≡ y if and only if f (x) = f (y) for all f ∈ F
is an equivalence relation on X.
Given an equivalence relation ≡ on a set X and an element a ∈ X, we define

the equivalence class a of a as follows:
a = {x ∈ X : a ≡ x}.
The equivalence class of a may also be denoted as [a], ã or some other similar
way.
A partition P of a set X is a set of subsets of X (i.e. P is an element of
℘℘(X)) such that
i. ∪P = X, and
ii. For any A, B ∈ P , if A 6= B then A ∩ B = ∅.
Examples.
i. The set {{0, 3}, {1, 2, 4}} is a partition of {0, 1, 2, 3, 4}.
ii. The sets (n, n + 1] for n ∈ Z partition R, meaning that {(n, n + 1] : n ∈ Z}
is a partition of R.
Lemma 4.2.1 Let ≡ be an equivalence relation on a set X. Then the set {x :
x ∈ X} is a partition of X. Conversely, any partition P of X gives rise to an
equivalence relation on X via
x ≡ y if and only if there exists A ∈ P such that x ∈ A and y ∈ A.
Thus there is a one-to-one correspondence between the set of partitions on X
and the set of equivalence relations on X.
Proof: Let ≡ be an equivalence relation on X. Since x ≡ x for all x ∈ X, we
have x ∈ x. It follows that ∪{x : x ∈ X} = X. Now we prove that for any
x, y ∈ X, either x = y or x ∩ y = ∅. We proceed to show that if x ∩ y 6= ∅, then
x = y. Let z ∈ x ∩ y. Thus z ≡ x and z ≡ y. Let t ∈ x. Thus t ≡ x. Since ≡ is
symmetric we have
t≡x
x≡z
z≡y
By transitivity we get t ≡ y, i.e. t ∈ y. We proved that any element t of x is an
element of y. The other direction holds as well. Thus x = y. Thus {x : x ∈ X}
is a partition of X.
Conversely, let P be a partition of X and define the relation ≡ as follows:
x ≡ y if and only if there exists A ∈ P such that x ∈ A and y ∈ A.
It is a matter of triviality to check that this is an equivalence relation. ¤
Quotient Set. Let ≡ be an equivalence on a set X. The set

{x : x ∈ X}
is called the quotient set and it is denoted by X/ ≡. It is important to note
that each element of X/ ≡ is a subset of X and any two distinct elements of
X/ ≡ are disjoint subsets of X.
4.2. EQUIVALENCE RELATIONS 51
Examples
i. Consider the equality as an equivalence relation on the set X. Then the
quotient set is {{x} : x ∈ X} and it is in one to one correspondence with
X.
ii. On any set X consider the equivalence relation defined by x ≡ y for all
x, y ∈ X. Then |X/ ≡ | = 1.
iii. On R consider the equivalence relation defined by xRy ⇐⇒ x2 = y 2 .
Then, for each x ∈ R, x = {x, −x}. Thus there is a natural one to one
correspondence between R/ ≡ and R≥0 given by r 7→ |r|.
iv. Let n > 0 be any natural number. On Z, consider the equivalence relation
≡n defined by
x ≡n y ⇐⇒ n divides x − y.
Then for each i ∈ Z, i = nZ + i = Z + j where j = 0, 1, . . . , n − 1
is the remainder when we divide i by n. Therefore we have Z/ ≡n =
{0, 1, . . . , n − 1} and |Z/ ≡n | = n.
v. On the set R, consider the equivalence relation ≡ defined by x ≡ y if and
only if x − y ∈ Z. Then for each r ∈ R, r = r + Z = s + Z for some unique
s ∈ [0, 1). Therefore R/ ≡ is in a natural one to one correspondence with
the interval [0, 1).
vi. On the set R, consider the relation ≡ defined by x ≡ y if and only if
x − y ∈ Q. Then for each r ∈ R, r = r + Q. In this case, we cannot find a
well-known set with which R/ ≡ is in a one-to-one correspondence1 .
vii. Let X and Y be two sets. Let a ∈ X. Consider the equivalence relation
≡a defined on Func(X, Y ) by
f ≡ g ⇐⇒ f|A = g|A .
Given f ∈ Func(X, Y ), f = {g ∈ Func(X, Y ) : f (a) = g(a)}. There
is a one to one correspondence between Func(X, Y ) ≡ and Y given by
f 7→ f (a). There is also a correspondence between Func(X, Y ) ≡ and the
set of constant functions from X into Y .
viii. On X := R × R × R, consider the equivalence relation ≡ defined by
(x, y, z) ≡ (x1 , y1 , z1 ) by 2(x − x1 ) + 3(y − y1 ) − (z − z1 ) = 0. Then
(x, y, z) = {(x1 , y1 , z1 ) : 2x + 3y − z = 2x1 + 3y1 − z1 }.
There is a (not so natural) correspondence between R3 / ≡ and R2 given
by (x, y, z) 7→ (x, y).
ix. On C, consider the equivalence relation defined by x ≡ y ⇐⇒ |x| = |y|.
Then x = {y ∈ C : |x| = |y|} and there is a natural one to one correspon-
dence between C/ ≡ and R≥0 given by x 7→ |x|. Each equivalence class x
is the circle of center 0 and radius |x|.
1 But there is such a set as we will see later in the second part of the book.
Canonical Surjection. If ≡ is an equivalence relation on the set X, then

the map from X onto the quotient set X/ ≡ given by x 7→ x is called the
canonical surjection from X into X/ ≡. Most often we will denote the
canonical surjection with symbol π. Thus, for all x ∈ X, π(x) = x = {y ∈ X :
x ≡ y} ∈ X/ ≡.
Induced Map. Let X be a set and ≡ be an equivalence relation on X. Let

Y be a set and let f : X −→ Y be a function such that for all x, y ∈ X, if x ≡ y
then f (x) = f (y). Then the map f : X −→ Y induces a map f˜ : X ≡−→ Y by
the rule f˜(x) = f (x).
Exercises.
i. Let X be a set.
a) For i ∈ I (index set), let Ri be an equivalence relation on X. Thus
Ri ⊆ X × X for each i ∈ I. Show that ∩i∈I Ri is an equivalence relation.
b) Conclude that for any relation R on X there is a smallest equivalence
relation containing R. This equivalence relation is called the equivalence
relation generated by R.
c) Let X = ∪n∈N {(x, y) ∈ R : 2n + 1 ≤ x2 + y 2 ≤ 2n + 2}. Define R on
X as follows: For A, B ∈ X, A R B if and only the line segment AB is
in X. Show that this is not an equivalence relation. Find the equivalence
relation generated by this relation. Find its set of equivalence classes.
ii. Let X be the set of functions from R into R. Let a ∈ R. Define the
following relation on X: f ≡ g if and only if there is an ² > 0 such that
f (x) = g(x) for all x ∈ (a − ², a + ²). Show that this is an equivalence
relation. Show that the functions induced by two distinct polynomials are
not equivalent.
iii. Let X be a set. For two subsets A and B of X define
X ≡ Y ⇐⇒ There is a bijection f : X −→ Y.
Show that this is an equivalence relation on ℘(X).
iv. Let X be a set. For two subsets A and B of X define
X ≡ Y ⇐⇒ X∆Y is finite.
Show that this is an equivalence relation on ℘(X).
v. Let a ∈ R be fixed. We will say that two functions f and g from R into R
have the same germ around a, and we write f 'a g, if there is an ² ∈ R>0
such that f (x) = g(x) for all x ∈ (a − ², a + ²).
a. Show that 'a is an equivalence relation on the set X := R R of all
functions from R into R. For f ∈ X, let [f ] denote the equivalence class
of f with respect to this equivalence relation.
4.3. PARTIAL ORDERS 53
b. For f and g in X, we have already defined f + g and f g at page

37. Show that if f1 'a g1 and f2 'a g2 then f1 + f2 'a g1 + g2 and
f1 f2 'a g1 g2 . It follows that one can define addition and multiplication
on X/ 'a .
c. For r ∈ R, define cr : R −→ R by cr (x) = r for all x ∈ R, i.e. cr is the
constant function that takes always the value r. Show that the function
c : R −→ X/ 'a defined by c(r) = [cr ] is one-to-one and that it satisfies
the equalities c(r + s) = c(r) + c(s) and c(rs) = c(r)c(s) for all r, sR.
vi. Let X be a set. For two subsets A and B of X, define the relation A ≡ B
by the condition “A∆B is finite”.
a) Show that this is an equivalence relation on ℘(X).
b) Show that ℘(X)/ ≡ has only one element if X is finite.
c) Conversely show that if ℘(X)/ ≡ has only one element then X is finite.
d) Show that ℘(N)/ ≡ is infinite.
Now let A, B, A1 , B1 ⊆ X be such that A ≡ A1 and B ≡ B1 , then
e) A ∩ B ≡ A1 ∩ B1 .
f) A ∪ B ≡ A1 ∪ B1 .
g) Ac ≡ Ac1 .
h) A \ B ≡ A1 \ B1 .
i) A∆B ≡ A1 ∆B1 .
4.3 Partial Orders

A partial order on a set is a irreflexive and transitive binary relation on this
set. A set together with a partial order is called a partially ordered set or a
poset for short.
Partial orders are denoted by such symbols as <, ≺, ⊂, C in general.
Examples.
i. Let < be a partial order on a set X and let A ⊆ X. Then the partial order
restricted to A, i.e. the binary relation < ∩(A × A), is a partial order on
A.
ii. Let X be a set. On ℘(X) define the relation
A < B ⇐⇒ A ⊆ B and A 6= B.
This is a partial order on ℘(X).
iii. The natural order on (any subset of) R is a partial order.

iv. On Z define
x ≺ y ⇐⇒ y < x.
This is a partial order. In general, the inverse of a partial order is a partial
order.
v. On R define
x < y ⇐⇒ |x| < |y|.
This is a partial order.
vi. On Z define
x ≺ y ⇐⇒ y divides x ⇐⇒ there is a z ∈ Z such that x = yz.
If < is a partial order on a set X, we define the relation ≤ on X by
x ≤ y ⇐⇒ x < y or x = y.
The binary relation ≤ is reflexive and transitive. Conversely, if ≤ is a reflexive

and transitive binary relation on a set X, then the relation < on X defined by
x < y ⇐⇒ x ≤ y and x 6= y
is irreflexive and transitive, i.e. is a partial order. Thus there is no much differ-
ence between irreflexive and transitive binary relation (i.e. partial orders) and
reflexive and transitive binary relations. For this reason, sometimes we will de-
fine a partial order as a reflexive and transitive binary relation, in which case,
we use a notation of the form ≤, ¹, ⊆, v, E.
If < is a binary relation on a set X, we define x ≥ y by y ≤ x and “x > y or
x = y” and “y < x or y = x” respectively. Given a binary relation ≤, we could
also define < by “x ≤ y and x 6= y”. It is easy to show that given a partial
order < on X, the relation ≤ defined on X is reflexive, transitive and satisfies
∀x ∀y ((x ≤ y ∧ y ≤ x) → x = y).
Conversely if a relation ≤ on X satisfies these three properties then < is a

partial order. Therefore, we may switch from < to ≤ as it suits us.
Define upper bound, lower bound, least upper bound = supremum and great-
est lower bound = infimum, maximum and minimum
Exercises.
i. Let < and ≺ be two partial orders on two sets X and Y respectively.
Define the partial order on X × Y via
(x, y) < (z, t) ⇐⇒ x < z or (x = z and y ≺ t).
Is this a partial order?

4.4. TOTAL ORDERS 55
ii. Let (Y, <) be a poset. Let f : X −→ Y be any function. Show that the
relation < on X defined by
x1 < x2 ⇐⇒ f (x1 ) < f (x2 )
is a partial order on X.
iii. Let (Y, <) be a poset, X a set and A ⊆ X. For two functions f and g
from X into Y define
f < g ⇐⇒ f (a) < g(a) for all a ∈ A.

X
Show that this is a partial order on Y.
iv. Let (Y, ≤) be a poset (reflexive and transitive), X a set and A ⊆ X. For
two functions f and g from X into Y define
f ≤ g ⇐⇒ f (a) ≤ g(a) for all a ∈ A.
Show that this is a (reflexive) partial order.

v. Find supremum and infimum (if they exist) of the following sets:
(a) {x ∈ R : x2 + x − 1 < 0},

(b) {x ∈ R : x2 + x − 1 > 0},
(c) { n1 + (−1)n : n ∈ N, n > 0},
1
(d) { 1−x : x ∈ R, x > 1},
1
(e) { 1−x : x ∈ R, x < 1}.
4.4 Total Orders

A partial order < on a set X is called a total order if for all x, y ∈ X one of
x < y, x = y, y < x
holds.
Lemma 4.4.1 Let < be a total order on a set X. For x, y ∈ X only one of
x < y, x = y, y < x
holds.
Proof: Assume x < y and x = y hold. Then x < x contradicting irreflexivity.

Similarly x = y and y < x cannot hold simultaneously. Assume finally that
x < y and y < x hold. Then by transitivity x < x, contradicting irreflexivity
again. ¤
Chapter 5
Induction
“Induction” is a method of proving statements about natural numbers. Suppose

φ(x) is a statement about the natural number x, and that we want to prove the
statement “φ(x) for all x ∈ N”. We can proceed as follows:
1) We first show that the statement φ(x) holds for n = 0, i.e. we prove φ(0).
2) Next, we show that whenever φ(n) holds for some natural number n, then
φ(n + 1) holds; in other words, we show that if φ(n) holds then φ(n + 1) holds.
If we can successfully do these two steps, then φ(x) will hold for all x ∈ N.
Because, by the first step, we will know that φ(0) holds. Since φ(0) holds, by
the second step (take n = 0) φ(1) will hold. Now that we know φ(1) holds, by
the second step (take n = 1) φ(2) will hold. Now that we know φ(2) holds, by
the second step (take n = 2) φ(3) will hold, etc. So φ(x) will hold for all x ∈ N,
as we can check one by one all numbers1 .
Example. As an exercise, let us prove that for all x ∈ N, 0 + 1 + 2 + . . . + x =

x(x + 1)/2. We let φ(x) denote “0 + 1 + 2 + . . . + x = x(x + 1)/2”.
Step 1. φ(0) holds. Indeed, φ(0) says that the sum of natural numbers
from 0 to 0 is 0(0 + 1)/2, i.e. is 0. This is of course true.
Step 2. If φ(n) holds then φ(n + 1) holds. Assume φ(n) holds, i.e.
assume that 0 + 1 + 2 + . . . + n = n(n + 1)/2. Under this assumption, we will
show that 0 + 1 + 2 + . . . + n + (n + 1) = (n + 1)(n + 2)/2 (just replace x by
n + 1 in φ(x)): 0 + 1 + 2 + . . . + n + (n + 1) = (0 + 1 + 2 + . . . + n) + (n + 1) =
n(n + 1)/2 + (n + 1) = (n + 1)(n + 2)/2. (Here the second equality follows from
the assumption that φ(n) holds. The rest are purely algebraic).
Therefore φ(x) holds for all x ∈ N.
Definition by Induction. We can also define functions or sets by induction.

For example, let us consider the function f : N −→ N by the rule f (0) = 1 and
1 If this last statement sounds strange to you, then you are a pretty good mathematician.
What we have here is not a precise mathematical proof, but only a heuristic argument why
φ(x) should hold for all x ∈ N. In the next part, we will prove that steps (1) and (2), do
indeed proof “φ(x) for all x ∈ N”.
57
58 CHAPTER 5. INDUCTION
f (n + 1) = 2f (n) for all n ∈ N. Then,

f (0) = 1
f (1) = 2f (0) = 2
f (2) = 2f (1) = 2 × 2 = 4
f (3) = 2f (2) = 2 × 4 = 8
More generally, one sees that f (n) = 2n .
Let us consider the following function defined by induction: g(0) = 3, g(1) =
2, g(2) = 1 and g(n + 1) = g(g(n)) for all n ≥ 2. Then
g(0) = 3
g(1) = 2
g(2) = 1
g(3) = g(g(2)) = g(1) = 2
g(4) = g(g(3)) = g(2) = 1
g(5) = g(g(4)) = g(1) = 2
g(6) = g(g(5)) = g(2) = 1
On the other hand the following attempt of definition by induction fails:
h(0) = 0, h(1) = 2, h(2) = 3 and h(n + 1) = h(h(n)) for all n ≥ 2, because
h(3) = h(h(2)) = h(3) = h(h(2)) = h(3) = . . .
We can use induction to define addition from the successor function S :
N −→ N given by S(n) = n + 1 as follows: n + 0 = n for all n ∈ N and
n + (m + 1) = (n + m) + 1
for all n, m ∈ N.
We can use + to define multiplication: n × 0 = n for all n ∈ N and
n × (m + 1) = n × m + n
for all n, m ∈ N.
We can use × to define exponentiation: n0 = 1 if n 6= 0, 0m = 0 for all
m 6= 0 and
nm+1 = nm × n
for all n, m ∈ N.
We also define factorial as follows: 0! = 1 and
(n + 1)! = n! × (n + 1)
for all n ∈ N.
n Choose k. For n, k ∈ N and k ≤ n, define

µ ¶
n n!
= .
k k!(n − k)!
µ ¶
n n!
A priory = k!(n−k)! ∈ Q. We will see in fact that it is in N.
k
59
Exercises.
i. Show that for any natural number n and for any real number x ∈ [0, 1],
n(n − 1) 2
(1 − x)n ≤ 1 − nx + x .
2
ii. Show that n! > 2n for all n large enough.
iii. Show that (x − 1)n ≥ xn − nxn−1 for all x > 1.
iv. Show that if 0 < x < 1 and n > 0 is a natural number, then (1 − x)n ≤
1 − nx + n(n−1)
2 x2 .
v. Show that for any n ∈ N \ {0}, 13 + 23 + . . . + n3 = (1 + 2 + . . . + n)2 .

n(n+1)(6n3 +9n2 +n−1)
vi. Show that for any n ∈ N \ {0}, 14 + 24 + . . . + n4 = 30 .
vii. Given a set X, define ℘n (X) as follows by induction on n: ℘0 (X) = X

and ℘n+1 (X) = ℘(℘n (X)). Show that ℘(℘n (X)) = ℘n (℘(X)) for all sets
X and all natural numbers n.
viii. Given a set X and a natural number n, define ℘n (X) as follows by induc-
tion on n: ℘0 (X) = X and ℘n+1 (X) = ℘(℘n (X)). Find all n such that
for any set X, {{∅}, {{X}}} ∈ ℘n (X).
µ ¶ µ ¶
n n
ix. Show that = .
k n−k
µ ¶ µ ¶
n n
x. Show that = = 1.
0 n
µ ¶
n
xi. Show that = n.
1
µ ¶ µ ¶ µ ¶
n+1 n n
xii. Show that = + .
k+1 k k+1
µ ¶
n
xiii. Deduce that ∈ N. (Hint: By induction on n).
k
µ ¶
n
xiv. Show that for n ∈ N and 0 ≤ k ≤ n, a set with n elements has
k
subsets with k elements.
xv. Show
µ ¶that for n ∈ N and k ∈ N with k ≤ n, a set with n elements has
n
subsets with k elements.
k
60 CHAPTER 5. INDUCTION
xvi. Show that for x, y ∈ R and n ∈ N,

n µ
X ¶
n
(x + y)n = xk y n−k .
k
k=0
(Hint: By induction on n).

µ ¶
Pn n
xvii. Show that k=0 = 2n .
k
µ ¶
Pn k n
xviii. Show that k=0 (−1) = 0.
k
xix. Compute (x + y + z)3 in terms of x, y and z.

xx. Compute (x + y + z)4 in terms of x, y and z.
xxi. Show that for x > 1 and n ∈ N, (x − 1)n ≥ xn − nxn−1 .
xxii. Show that for x < 1, (1 − x)n ≥ 1 − nx.
xxiii. Show that for n ∈ N \ {0}, (1 + n1 )n ≤ (1 + 1
n+1 )
n+1
.
xxiv. Find and prove (by induction) formulas for the sums
3 5 2n + 1
+ 2 2 + ··· + 2
12 ·2 2 2 ·3 n · (n + 1)2
and
1 + 9 + 25 + · · · + (2n + 1)2 .
xxv. Prove by induction that for all real numbers a and b and natural number
n,
(an + an−1 b + an−2 b2 + · · · abn−1 + bn )(a − b) = an+1 − bn+1 .

Chapter 6
Bijections, revisited
6.1 Schröder-Bernstein’s Theorem

The next result is probably the first nontrivial result of the book.
Theorem 6.1.1 (Schröder-Bernstein) Let A and B be two sets. If there are

injections from A into B and from B into A, then there is a bijection between
A and B.
Proof: Let f : A −→ B and g : B −→ A be two injections. Given a ∈ A,

consider the finite or infinite sequence (a, f (a), g(f (a)), f (g(f (a))), . . .). If this
sequence ends at B, call the point a a B-point.
6.2 Examples of Sets in Bijection

Two sets are said to be in bijection if there is a bijection between them. It
is clear that two finite sets are in bijection if and only if they have the same
number of elements. It is also clear that a finite set cannot be in bijection with
an infinite set.
Although a finite set cannot be in bijection with one of its proper subsets,
an infinite set can be in bijection with one of its subsets. For example the map
x 7→ x + 1 is a bijection between N and N \ {0}. The map x 7→ 2x is a bijection
between Z and 2Z.
We do not have enough tools for the moment to prove the following theorems:
Theorem 6.2.1 If X is an infinite set and x ∈ X, then there is a bijection

between X and X \ {x}.
We will prove this theorem in the next part. But we can prove the following
results – although naively.
Theorem 6.2.2 If X is an infinite set then there is an injection from N into

X.
61
62 CHAPTER 6. BIJECTIONS, REVISITED
Proof: We will define an injection f : N −→ X. Let x0 ∈ X, x1 ∈ X \ {x0 } and

for each n ∈ N, let xn+1 ∈ X \ {x0 , . . . , xn }. Since X is infinite we can continue
this procedure of choosing an element xn+1 not equal to any of the previously
chosen elements x0 , . . . , xn . Set f (n) = xn . Then clearly f is an injection from
N into N. ¤
Theorem 6.2.3 If A is an infinite subset of N, then there is a bijection between

N and A.
We will give a naive proof of this result later on in this part. Now we will
find some bijections between unexpected sets. If there is a bijection between
two sets X and Y , we will represent this fact as X ∼ Y . Note that
X∼X
If X ∼ Y then Y ∼ X
If X ∼ Y andY ∼ Z then Y ∼ X
Chapter 7
Some Automorphism
Groups
7.1 Binary Unirelational Structures

A binary unirelational structure Γ is a set X together with a binary relation
R. Thus a binary unirelational structure Γ is a pair (X, R) where X is a set
and R is a binary relation Γ.
An isomorphism of a binary unirelational structure (X, R) onto another
binary unirelational structure (Y, S) is a bijection f : X −→ Y such that for any
x1 x2 ∈ X, x1 Rx2 if and only if f (x1 )Rf (x2 ). Two binary relational structures
among which there is an isomorphism are called isomorphic. An isomorphism
from a graph onto itself is called an automorphism of the binary unirelational
structure. The set of all automorphisms of a binary unirelational structure Γ is
called the automorphism group of Γ and is denoted by Aut(Γ).
The dual of a binary unirelational structure (X, R) is the binary unirela-
tional structure (X, S) defined as follows:
xSy ⇐⇒ x 6 Ry,
i.e. xSy if and only if xRy does not hold.
Exercises.
i. Show that the set Aut(Γ) of all automorphisms of a binary unirelational
structure Γ is is closed under composition and inversion.
ii. Show that the automorphism groups of a binary unirelational structure
and of its dual are equal.
iii. Let (X, R) and (Y, S) be two binary unirelational structures. Let φ be an
isomorphism from (X, R) onto (Y, S). Show that
Aut(Y, S) = φ Aut(X, R)φ−1 .
63
64 CHAPTER 7. SOME AUTOMORPHISM GROUPS
7.2 Automorphism Groups of Graphs

A graph Γ is a set X together with a symmetric and irreflexive binary relation
X. Thus a graph Γ is a pair (X, R) where X is a set and R is a symmetric and
irreflexive binary relation X.
If (X, R) is a graph and x, y ∈ X are such that xRy then we say that x and
y are connected. We often write x − y instead of xRy. We also say that x and
y are related.
The complete graph on a set X is the graph where any two distinct x, y ∈
X are related.
Exercises.
i. Show that the automorphism group of a complete graph on a set X is

Sym(X).
ii. Let us find all graph structures on the set X = {1, 2, 3}. For any two
distinct points x and y we know that if x − y then y − x, so to shorten our
writing, we will write only one of the two relations.
Γ∅ : no relations at all
Γ3 : only 1 − 2 (and 2 − 1 of course)
Γ2 : only 1 − 3
Γ1 : only 2 − 3
Γ13 : only 1 − 2 and 2 − 3
Γ12 : only 1 − 3 and 2 − 3
Γ23 : only 1 − 2 and 1 − 3
Γ123 : all possible relations 1 − 2, 2 − 3 and 1 − 3
iii. ¶ Show that Γ1 , Γ2 and Γ3 are isomorphic. Show that Γ13 , Γ12 and Γ3
are isomorphic. Show that G1 and G23 are duals of each other.
Thus on X there are only fundamentally different (i.e. nonisomorphic)
graph structures, Γ∅ , Γ1 , Γ23 and Γ123 .
iv. Let X be a set. Let Γ be the set of subsets of X with two elements. On
Γ define the relation αRβ if and only if |α ∩ β| = 1. Then Γ becomes a
graph with this relation.
a) Calculate Aut(Γ) when |X| = 4.
b) Draw the dual of Γ when |X| = 5.
c) Show that for any α, β ∈ Γ there is an automorphism φ ∈ Aut(Γ) such
that φ(α) = β.
d) Calculate Aut(Γ) when |X| = 5.
v. Show that Aut(Γ∅ ) = Aut(Γ123 ) = Sym(3) and Aut(Γ1 ) = Aut(Γ23 ) =

{1, (2, 3)}.
7.3. GEOMETRIC AUTOMORPHISM GROUPS 65
vi. Show that the automorphism group of a graph Γ on X = {1, 2, 3, 4} is the

automorphism group of one of the following graphs:
Γ∅ : no relations at all
Γ12 : only 1 − 2 (and 2 − 1 of course)
Γ123 : only 1 − 2 and 2 − 3
Γ12−34 : only 1 − 2 and 3 − 4
Γ1234 : only 1 − 2, 2 − 3 and 3 − 4
Γ∗ : only 1 − 2, 1 − 3 and 1 − 4
Show that
Aut(Γ∅ ) = Sym(4)
Aut(Γ12 ) = Aut(Γ12−34 ) = {1, (1, 2), (3, 4), (1, 2)(3, 4)}
Aut(Γ123 ) = {1, (1, 3)}
Aut(Γ12−34 ) = {1, (12), (34), (12)(34), (13)(24), (23)(14), (1324), (1223)}
Aut(Γ1234 ) = {1, (14)(23)}
Aut(Γ∗ ) = {1, (12), (13), (23), (123), (132)}
Draw the multiplication table of Aut(Γ12−34 )
vii. Let Γ be a graph such that for any α, α1 , β, β1 , if α 6= α1 and β 6= β1 ,

then there is a φ ∈ Aut(G) such that φ(α) = β and φ(a1 ) = β1 . What
can you say about Γ?
viii. Find a graph Γ on six points such that Aut(Γ) = {Id}.
ix. Find a finite graph Γ such that | Aut(Γ)| = 3.
x. Let n ≥ 3 and let Γ be the cyclic graph on {1, 2, . . . , n}, i.e. the only
relations are 1 − 2 − 3 − . . . − (n − 1) − n − 1. Show that ρ := (1, 2, . . . , n) ∈
Aut(Γ). Show that τ = (2, n)(3, n−1) . . . ∈ Aut(Γ). (For example if n = 6
then τ = (2, 6)(3, 5), if n = 7 then τ = (2, 7)(3, 6)(4, 5)). Show that
Aut(Γ) = {ρi τ j : i = 0, 1, . . . , n − 1 and j = 0, 1}
and that | Aut(Γ)| = 2n.
7.3 Geometric Automorphism Groups

Problem 7.3.1 (Isometries of R) Let G := {g : R −→ R : |x − y| = |g(x) −
g(y)|}. We will show that G is a group under composition and we will find all
the elements of G. Note that G is clearly closed under composition and IdR ∈ G.
Also if g ∈ G is a bijection, then g −1 ∈ G as well. To show that G is a group,
we just need to show that the elements of G are bijections.
i. Show that any element of G is one-to-one.
ii. For a ∈ R, let τa : R −→ R be given by τa (x) = x + a (translation). Show
that τa ∈ G. Show that the set T = {τa : a ∈ R} is a group under composition.
iii. For ² = ±1, let ρ² : R −→ R be given by ρ² (x) = ²x. Show that ρ² ∈ G.

Show that the set R = {ρ1 , ρ−1 } is a group under composition.
iv. Show that for any g ∈ G there are unique ² ∈ {1, −1} and a ∈ R such
that g = τa ◦ ρ² . Conclude that the elements of G are bijections.
v. Conclude that G is a group under composition.
The group G is called the group of isometries of R.
Problem 7.3.2 (Isometries of R2 ) Let G := {g : R2 −→ R2 : |x − y| =

|g(x) − g(y)|}. We will show that G is a group under composition and we
will find all the elements of G explicitly. Note that G is clearly closed under
composition and IdR ∈ G. Also if g ∈ G is a bijection, then g −1 ∈ G as well.
To show that G is a group, we just need to show that the elements of G are
bijections.
Note that G is the set of functions from R2 into R2 that send any circle into
a circle of the same radius.
i. Show that any element of G is one-to-one.
ii. For (a, b) ∈ R2 , let τ(a,b) : R2 −→ R2 be given by τ(a,b) (x, y) = (x+a, y+b)
(translation). Show that τ(a,b) ∈ G. Show that the set T := {τ(a,b) : (a, b) ∈ R}
is a group under composition.
iii. For θ ∈ [0, 2π), let ρθ : R2 −→ R2 be the clockwise rotation of angle θ
around the center (0, 0), i.e. ρθ (x, y) = (x cos(θ) − y sin(θ), x sin(θ) + y cos(θ)).
Show that ρθ ∈ G. Show that the set R := {ρθ : θ ∈ [0, 2π)} is a group under
composition.
iv. For ² = ±1, let σ² : R −→ R be given by σ² (x, y) = (x, ²y). Show that
σ² ∈ G. Show that the set S = {σ1 , σ−1 } is a group under composition.
v. Show that for any g ∈ G there is a unique (a, b) ∈ R2 such that such that
(τa,b ◦ g)(0, 0) = (0, 0).
vi. Show that for any g ∈ G such that g(0, 0) = (0, 0) there is there is a
unique θ ∈ [0, 2π) such that (ρθ ◦ g)(0, 0) = (0, 0) and (ρθ ◦ g)(1, 0) = (1, 0)
vii. Show that for any g ∈ G such that g(0, 0) = (0, 0) and g(1, 0) = (1, 0)
we have g(0, 1) = (0, ²) for some ² = ±1. Conclude that for such g there is
a unique ² = ±1 such that (s² ◦ g)(0, 0) = (0, 0), (s² ◦ g)(1, 0) = (1, 0) and
(s² ◦ g)(0, 1) = (0, 1).
vii. Show that if g ∈ G is such that g(0, 0) = (0, 0), g(1, 0) = (1, 0) and
g(1, 0) = (1, 0), then g = IdR2 .
viii. Conclude that for any g ∈ G there are unique (a, b) ∈ R2 , θ ∈ [0, 2π)
and ² = ±1 such that g = τ(a,b) ◦ ρθ ◦ σ² . Conclude that g is a bijection.
ix. Conclude that G is a group under composition.
The group G is called the group of isometries of R.
x. Conclude that an element of G sends a line onto a line.
Problem 7.3.3 (Affine Transformations of R2 ) Let G := {g : R2 −→ R2 :

g is a bijection and sends a line into a line}. It is clear that G is a group under
composition. We will find all the elements of G explicitly. As we have noticed
in the end of Problem 7.3.2, the elements of the group defined there are in G.
Thus the groups T and R are subsets of G. We adopt the same terminology.
7.4. BACK AND FORTH ARGUMENT 67
i. Show that an element of G respect the parallelism, i.e. sends two parallel
lines onto two parallel lines.
ii. For c ∈ R>0 , let hc : R2 −→ R2 be defined by hc (x, y) = (cx, y). Show
that hc ∈ G and that the set H := {hc : c ∈ R>0 } is a group under composition.
iii. For d ∈ R define ud : R2 −→ R2 by ud (x, y) = (x + dy, y). Show that
ud ∈ G and that the set U := {td : d ∈ R} is a group under composition.
iv. For a bijection phi : R −→ R satisfying the properties φ(x + y) =
φ(x) + φ(y), φ(xy) = φ(x)φ(y) and φ(1) = 1, define the map αφ : R2 −→ R2
by αφ (x, y) = (φ(x), φ(y)). Show that αφ ∈ G and that the set A := {αφ :
φ as above} is a group under composition.
v. Show that for any g ∈ G there is a unique (a, b) ∈ R2 such that such that
(τa,b ◦ g)(0, 0) = (0, 0).
vi. Show that for any g ∈ G such that g(0, 0) = (0, 0) there is there is a
unique θ ∈ [0, 2π) such that (ρθ ◦ g)(0, 0) = (0, 0) and (ρθ ◦ g)(1, 0) is on the
positive side of the x-axis.
vii. Show that for any g ∈ G such that g(0, 0) = (0, 0) and g(1, 0) is on the
positive side of the x-axis, there is a unique c ∈ R>0 such that (hc ◦ g)(0, 0) =
(0, 0) and (hc ◦ g)(1, 0) = (1, 0).
viii. Show that for any g ∈ G such that g(0, 0) = (0, 0) and g(1, 0) = (1, 0)
there is a unique d ∈ R such that (ud ◦ g)(0, 0) = (0, 0), (ud ◦ g)(1, 0) = (1, 0)
and (ud ◦ g)(0, 1) = (0, 1).
ix. Show that if g ∈ G satisfies g(0, 0) = (0, 0), g(1, 0) = (1, 0) and
g(0, 1) = (0, 1), then g ∈ A. (Hint: Express the addition and multiplication
of real numbers geometrically and use part (i)).
x. Show that for any g ∈ G there are unique (a, b) ∈ R2 , θ ∈ [0, 2π),
c ∈ R>0 , d ∈ R and bijection phi : R −→ R satisfying the properties φ(x + y) =
φ(x)+φ(y), φ(xy) = φ(x)φ(y) and φ(1) = 1, such that g = τ(a,b) ◦ρθ ◦hc ◦ud ◦αφ .
The group G is called the affine transformations of R2 .
Problem 7.3.4 On R2 define the metric d((x, y), (z, t)) := |z − x| + |t − y|. Let
G be the group of isometries of this metric. Find the elements of G.
Problem 7.3.5 On R2 define the metric d((x, y), (z, t)) := max(|z − x|, |t − y|).
Let G be the group of isometries of this metric. Find the elements of G.
7.4 Back and Forth Argument

In this subsection we will prove two theorems using an interesting method called
back and forth.
7.4.1 Dense Total Orderings

Theorem 7.4.1 Any two countable dense total orderings without end points
are isomorphic.
Corollary 7.4.2 A countable dense total ordering is ω-homogeneous.

7.4.2 Random Graphs

Theorem 7.4.3 Any two countable random graphs are isomorphic.
Corollary 7.4.4 A countable random graph is ω-homogeneous.

Chapter 8
Formulae
In this chapter, we will see the formulae of set theory.

Formulae are like sentences of a language. Like any language, mathematical
languages have alphabets. All mathematical languages include the following
symbols:
Logical Connectives. There are two of those: ∧ called conjunction and
¬ called negation.
Existential Quantifier. There is only one: exists.
Parentheses. There are two kinds of parentheses, left parenthesis ( and
right parenthesis ).
Equality Symbol. The equality is symbolized by =, or by ≡ sometimes.
Variables. There are an infinity of them: v0 , v1 , v2 , v3 , . . .
Instead of the infinitely many variables, we could have adapted only two
symbols v and | and use
v for v0
v| for v1
v|| for v2
v||| for v3 etc.
Apart from these symbols, each mathematical theory has its own specific
symbols. Set theory has only one specific symbol: ∈, called the membership
relation.
Thus the symbols we will use to write the formulae in set theory are the
following:
∧, ¬, ∃, (, ), =, ∈, v0 , v1 , v2 , v3 , . . .
These symbols – although do have an intended meaning – do not have a
meaning. They are only symbols and nothing more.
The formulae will be certain “words” in this alphabet, i.e. a finite string of
symbols written only using these symbols.
69
70 CHAPTER 8. FORMULAE
Chapter 9
Miscellaneous Exercises
i. Let φ : R −→ R be such that φ(x + y) = φ(x) + φ(y) for all x, y ∈ R.

a) Assume φ(x2 ) = φ(x)2 . Show that φ(xy) = φ(x)φ(y) for all x, y ∈ R
and φ(q) = q for all q ∈ Q.
b) Assume φ(x−1 ) = φ(x)−1 . Show that φ(xy) = φ(x)φ(y) for all x, y ∈ R.
71
72 CHAPTER 9. MISCELLANEOUS EXERCISES
Part II
Axiomatic Set Theory
73
Chapter 10
Basics
10.1 Russell’s Paradox

A century ago, mathematicians, as common people do, believed that any set
of objects could be a set in the mathematical sense. For example, they could
speak about the set of all sets without much questioning the use of the term
“set”.
After reading the first volume of a work of Frege, Bertrand Russell found
the following mathematical paradox:
Russell’s Paradox. Let X be the set of all sets. Let
Y := {x ∈ X : x 6∈ x}.
Thus a set x is in Y if and only if x 6∈ x. In symbols:
∀x (x ∈ Y ⇐⇒ x 6∈ x).
Since Y is a set, the formula above – that holds for all sets x – should also
hold for Y . Replacing x by Y we get:
Y ∈ Y ⇐⇒ Y 6∈ Y.
In English this translates as follows: Y is an element of itself if and only if it
is not an element of itself. This is certainly a contradiction, and a contradiction
in set theory, the basis of all mathematics, therefore of all sciences... This is
very serious, and one cannot just ignore it.
Russell’s Paradox changed the way mathematicians looked at mathematics;
from naive mathematicians – that believed naively that any collection could be
a set – they turned into more careful creatures.
A careful analysis of the above paradox (like the one a careful creature
would carry out) reveals that the acceptance of the collection of all sets as a set
is the source of the problem. The collection of all sets is certainly something, a
75
76 CHAPTER 10. BASICS
“collection” for example, but is not allowed to be a set itself. Something that
contains all sets cannot – or should not be – a set, because otherwise we arrive
at a contradiction...
What to do?
The solution is the following: be much more careful before naming something
a set. Set the rules for something to be a set and insure that each naive set is
really a set, i.e. that it passes the test of being a set.
The experience shows that collections which are too large cause problems.
Therefore our test should exclude too large collections from being sets. The
easiest way to avoid big collections from being sets is to start from the smallest
of all sets, the emptyset, and to construct the other sets step by step from the
emptyset, by taking the unions, the intersections, the products, the “set” of
subsets of the previously constructed sets to construct new sets. By doing so,
one should never have enough time to arrive at large collections for them to
be sets. For example, if the steps never end, if after each step we can take
the next step to invent new sets, during this procedure, we will never meet the
collection of all sets. This is the idea behind Russell’s theory of types as well as
the Zermelo-Fraenkel set theory that we will expose in this book.
10.2 Easy Axioms

We accept the two undefined terms ”set” and ”being an element of”. If x and
y are sets and x is an element of y, we write x ∈ y, otherwise we write x 6∈ y.
In this case, we also say that “x is in y” or “x is not in y” respectively. The
elements of a set will be sets themselves. Every object we will define will be a
set, including objects like “function” and “equivalence relation”.
We do not know whether any set exists yet. They may or may not exist. We
create our first set:
A1. Emptyset. There exists a set with no elements.

This is an axiom and we accept it without any discussions.
Formally this axiom is written as
∃x∀y y 6∈ x.
Now the following question arises: How many sets without elements are
there? We would like to say that there is only one set without elements. This
would be saying that the set of humans with five heads on their shoulders is equal
to the set of apple trees of 3000 meters long, since both sets are - presumably -
empty.
To prove that there is only one set with no elements, we assume that x and
y are two sets with no elements and we proceed to show that x = y. But since
we do not know the meaning of the equality of two sets, we cannot proceed!
The second axiom will tell us exactly when two sets are equal.
10.2. EASY AXIOMS 77
A2. Equality. Two sets which have the same elements are equal.
Formally this axiom is written as
∀x∀y (∀z (z ∈ x ⇐⇒ z ∈ y) → x = y).
Now we can prove our first theorem:
Theorem 10.2.1 There is only one set without elements.
Proof: By A1 we know that there is at least one set without elements. Let
x and y be “two” sets without elements. Assume that x 6= y. Then by A2,
there is an element in one of them which is not in the other. But this is absurd
because none of the sets can have elements. Thus it is not true that x 6= y.
Hence x = y. ¤
The statement of the theorem can be written as follows:
∃x(∀y y 6∈ x ∧ ∀z((∀t t 6∈ z) → z = x)).
The unique set without elements is called emptyset and is denoted by the
symbol ∅.
We will use the symbol ∅ in our formal formulas as an abbreviation.
We prove
√ a statement which is not supposed to make any sense at this point
(because 2 is not defined yet):
√
Theorem 10.2.2 For any x ∈ ∅, x = 2.
√
Proof: Assume not. Then ∅ contains an element which is not equal to 2.
But this is absurd because ∅ does not contain any elements at all! ¤
What saves us from a contradiction in the statement above is that emptyset√
has no elements. But if it had one, this element would have to be equal to 2,
and as you must have guessed to π as well. √Fortunately ∅ has no elements and
we do not get a contradiction of the form “ 2 = π”.
A set x is called a subset of another set y if any element of x is an element
of y. We then write x ⊆ y.
Formally we write this as
x ⊆ y ⇐⇒ ∀z(z ∈ x → z ∈ y).
We use the symbol ⇐⇒ as an abbreviation for the English phrase “if and
only if”, while the symbol ↔ is used as a mathematical symbol.
To prove that x is a subset of y we need to prove that x and y are sets and
that every element of x is an element of y.
Clearly every set is a subset of itself. Formally ∀x x ⊆ x.
Theorem 10.2.3 ∅ is a subset of every set.

Proof: Let x be a set. Assume ∅ is not a subset of x. Then by definition, there

is an element in ∅ which is not in x. But ∅ has no elements. Thus ∅ cannot have
an element which is not in x. Hence ∅ ⊆ x. ¤
Formally, the above statement can be written as follows: ∀x ∅ ⊆ x, where
∅ ⊆ x is in fact an abbreviation of a longer sentence.
Theorem 10.2.4 ∅ is the only subset of ∅.
Proof: Let x be a subset of ∅. If x 6= ∅, then there is an element y in x. Since

x ⊆ ∅, y is also an element of ∅. Contradiction. Thus x = ∅. ¤
A3. Definable Subsets. Let x be a set and φ(y) a “property”. Then there
is a set whose elements are exactly the elements of x that satisfy the property φ.
We will be unprecise about the meaning of the word “property”. The reader
should only be aware that “properties” are given by finite formulas that involve
only the mathematical symbols
∀, ∃, ∧, ∨, →, ⇐⇒ , (, ), =, ¬,
the abbreviations like ∅ and ⊆, variables like x, y and z and constants (param-
eters, the existing sets).
By A2, given x and φ, the set as in A3 is unique. We denote this set as
{y ∈ x : φ(y)}.
Using this axiom we can prove that the intersection of two sets is a set.
Theorem 10.2.5 If x and y are sets then there is a set whose elements are
exactly the common elements of x and y.
Proof: Let φ(z) be z ∈ y. Use A3 to see that {z ∈ x : φ(z)} is a set. This set
is exactly the set of common elements of x and y. ¤
By A2, given x and y, the set given in Theorem 10.2.5 is unique. We call
this set the intersection of x and y and we denote it by x ∩ y.
Thus z ∈ x ∩ y ⇐⇒ (z ∈ x) ∧ (z ∈ y).
Axiom A3 allows us to prove that the difference of two sets is a set:
Theorem 10.2.6 If x and y are sets then there is a set whose elements are
exactly the elements of x which are not in y.
Proof: Let φ(z) be z 6∈ y. Use A3 to see that {z ∈ x : φ(z)} is a set. This set
is exactly the set of elements of x which are not in y. ¤
By A2, given x and y, the set given in Theorem 10.2.6 is unique. We denote
this set by x \ y. It can be called “x minus y”.
Now we prove that a certain collection cannot be a set:
10.2. EASY AXIOMS 79
Theorem 10.2.7 (Bertrand Russell) There is no set that contains all sets
as elements.
Proof: Assume not. Let x be a set that contains all sets as elements. Let φ(y)
be “y 6∈ y” Consider the set
z := {y ∈ x : φ(y)} = {y ∈ x : y 6∈ y}.
z is a set by A3. For any set y, we have
y ∈ z ⇐⇒ y 6∈ y.
Since the above statement holds for any set y and since z is a set, we may replace
y of the formula above by z:
z ∈ z ⇐⇒ z 6∈ z.
This is a clear contradiction. Thus x is not a set, or the set of all sets does not
exist. ¤
At this point we cannot prove that the union of two sets is a set. We need
the following axiom for this.
A4. Union of Two Sets. If x and y are two sets, then there is a set whose
elements are exactly the elements that are in either x or y. By A2 such a set is
unique. It is called the union of x and y and is denoted by x ∪ y.
Thus we have z ∈ x ∪ y ⇐⇒ (z ∈ x) ∨ (z ∈ y).

Later on we will adopt a more general axiom than A4.
There are some relationships between ∪ and \:
(x \ y) \ z = x \ (y ∪ z)
y ∪ (x \ y) = x ∪ y
With the first four axioms in hand, we can only be sure of the existence of
∅, indeed:
1) The first axiom tells us that emptyset exists.
2) The second axiom cannot be used to create new sets.
3) The new sets obtained by A3 are subsets of the old sets. Thus, starting
from ∅, we cannot obtain new sets using A3.
4) If x and y are emptysets, the new set obtained using A4 is also emptyset,
hence not new.
Another way of seeing this is by noticing that the universe that contains
only emptyset as a set satisfies all four axioms.
Now we will adopt a new axiom that will allow us to define new sets.
A5. Sets with One Elements. If x is a set, then there is a set which has
only x as an element.
We write this set as {x}.
Now, starting from ∅, we can obtain new sets:
∅
{∅}
{{∅}}
{{{∅}}}
...
These sets have only one element. Such sets are called singleton sets. A set
whose elements are x and y is denoted by {x, y}. If we can name all the elements
of a set one by one and enumerate them as a1 , . . . , an , then we will denote this
set by {a1 , . . . , an }.
Using A4, we can obtain sets with more than one element:
{∅} ∪ {{∅}} = {∅, {∅}}
{∅, {∅}} ∪ {{{∅}}} = {∅, {∅}, {{∅}}}
Now we define “zero” as the emptyset:
0 := ∅.
Given a set x, we define its successor S(x) as
S(x) = x ∪ {x}.
By axioms A4 and A5, S(x) is a set. Next we define 1, 2, 3 etc. as follows:
1 := S(0) = 0 ∪ {0} = {0}

2 := S(1) = 1 ∪ {1} = {0} ∪ {1} = {0, 1}
3 := S(2) = 2 ∪ {2} = {0, 1} ∪ {2} = {0, 1, 2}
4 := S(3) = 3 ∪ {3} = {0, 1, 2} ∪ {3} = {0, 1, 2, 3}
...
By using our axioms and results we can form sets that contain any finite number
of these sets. For example, {0, 3} is a set because
{0, 3} = 1 ∪ (4 \ 3).
To prove that {0, 3} is a set we can also use A3 with x = 4 and φ(y) as “y =
0 ∨ y = 3”.
Although we now have sets with as many elements as we wish, we are still
not sure that there are sets with infinitely many elements. For example, we do
not know that yet that there is a set that contains 0, 1, 2, 3, 4,... In fact, we
cannot prove yet that there is such a set. We need a new axiom for this. Before
creating sets with infinitely many elements, we will state a few more axioms and
we will define some basic notions of set theory.
10.3. SLIGHTLY MORE COMPLICATED AXIOMS 81
10.3 Slightly More Complicated Axioms

A40 . Union. If x is a set, then there is a set whose elements are exactly the
elements of the elements of x.
This axiom tells us that the union of the elements (remember that the ele-
ments of a set are sets as well) of a set is a set.
Although we do not know whether there are sets with infinitely many ele-
ments, the set x in A40 could have - hypothetically - infinitely many elements.
If x = {a, b}, then the set obtained by A40 is just a ∪ b. Similarly, if x =
{a, b, c}, then the set obtained by A40 is just a ∪ b ∪ c.
We denote the set given by A40 by ∪x or ∪y∈x y. Note that
z ∈ ∪x ⇔ ∃y(y ∈ x ∧ z ∈ y).
0
For the moment, A4 is not stronger than A4, because, given two sets x and
y, we do not know yet that {x, y} is a set. We now adopt an axiom that will
solve this problem:
A50 . Sets with Two Elements. If x and y are sets then there is a set whose
elements are only x and y.
We denote this set as {x, y} of course. By taking x = y in A50 , we see
that A5 is a consequence of A50 , therefore we can erase A5 from the list of our
axioms. Also, as we have noticed A4 is a consequence of A40 and A50 , and we
can also erase A4 from our list.
Lemma 10.3.1 ∪∅ = ∅.
Proof: By definition, for any set x, z ∈ ∪x if and only if there exists a y ∈ x
such that z ∈ y. Apply this to x = ∅: z ∈ ∪∅ if and only if there exists a y ∈ ∅
such that z ∈ y. Thus ∪∅ cannot have an element. ¤
We can also form the intersection of all sets which are the elements of a
given set:
Theorem 10.3.2 Let x be a set. Then there is a set whose elements are exactly
the elements of ∪x which are elements of all the elements of x.
Proof: Let φ(z) be “∀y(y ∈ x → z ∈ y)”. Now use A3: {z ∈ ∪x : φ(z)} is the
set we are looking for. ¤
This set is denoted by ∩x or ∩y∈x y. Thus
z ∈ ∩x ⇔ ∀y (y ∈ x → z ∈ x).
The following statement cannot be proven: “Let x be a set. Then there is

a set whose elements are exactly the common elements of all the elements of
x.” Because to prove that this is a set we need x to be nonempty. Indeed, if
we adopted this definition, ∩∅ would be the set of all sets, creating Russell’s
Paradox.
A6. Power Set. If x is a set then there is a set whose elements are exactly
the subsets of x.
We denote this set with the symbol ℘(x) and call it the power set of x.
Thus
y ∈ ℘(x) ⇐⇒ y ⊆ x.
For example if x = 2 = {0, 1}, then
℘(x) = {∅, {0}, {1}, {0, 1}}.
Note that ∅ and x are always elements of ℘(x).

Even after these more complicated axioms, we still cannot prove that infinite
sets exist, because given a finite set whose elements are finite sets all our axioms
create only finite sets. In the next section we will adopt an axiom that will allow
us ”to be sure” of the existence of infinite sets. We should also say that we do
not really know what “finite” or “infinite” means at this point since we did not
define these concepts yet. Finiteness is not such an easy concept as we will later
see.
10.4 Cartesian Product of Two Sets

If X and Y are two sets, one can form their Cartesian product X × Y . In this
section we will define X × Y mathematically. The elements of X × Y will be
entities denoted as (x, y) for x ∈ X and y ∈ Y . We will first define these pairs
(x, y). The only property of the pairs (x, y) which will interest us will be the
following:
(x, y) = (z, t) ⇐⇒ x = z and y = t.
So we will define the pair (x, y) as a set satisfying the above property.
For two sets x and y, we let (x, y) = {{x}, {x, y}. We call (x, y) the (or-
dered) pair of x and y.
Lemma 10.4.1 i. (x, y) is a set.

ii. (x, y) = (z, t) if and only if x = z and y = t.
Proof: See Exercise xxi, page 15. The second part follows also from Exercise
vi, page 83. ¤
Lemma 10.4.2 Let X and Y be two sets. Then there is a set X × Y whose
elements are exactly the ordered pairs (x, y) for x ∈ X and y ∈ Y .
Proof: Since x ∈ X, we have x ∈ X ∪ Y . Similarly y ∈ X ∩ Y . It follows that

the sets {x} and {x, y} are subsets of X ∪ Y and hence elements of ℘(X ∪ Y ).
Therefore {{x}, {x, y}} is a subset of ℘(X ∪ Y ). Hence (x, y) = {{x}, {x, y}} ∈
℘(℘(X ∪ Y )). Thus all the elements of X × Y are elements of ℘(℘(X ∪ Y )).
Now among the elements of ℘(℘(X ∪ Y )) we will distinguish the ones of the
10.4. CARTESIAN PRODUCT OF TWO SETS 83
form (x, y) by a formula φ(z), i.e. by a property. In other words, we will find a
formula φ(z) such that for all z, the following will hold:
φ(z) ⇐⇒ ∃x∃y (x ∈ X ∧ y ∈ Y ∧ z = {{x}, {x, y}}).
If we can do this, then we will have
X × Y = {z ∈ ℘(℘(X ∪ Y )) : φ(z)},
and so by A3, X × Y will be a set.

Now we find the formula φ(z) logically equivalent to
∃x∃y (x ∈ X ∧ y ∈ Y ∧ z = {{x}, {x, y}}).
In this “pseudo-formula” we can replace z = {{x}, {x, y}} by
∃t∃u (t ∈ z ∧ u ∈ z ∧ t = {x} ∧ u = {x, y} ∧ ∀w (w ∈ z → (w = t ∨ w = u))).
In this ”formula”, the strings t = {x} and u = {x, y} are still nonacceptable.
We replace t = {x} by x ∈ t ∧ ∀s (s ∈ t → s = x). And we replace u = {x, y} by
x ∈ u ∧ y ∈ u ∧ ∀s (s ∈ u → (s = x ∨ s = y)). After doing all these replacements,
we get φ(z), a legitimate formula. ¤
Exercises.
i. Find X × ∅.
ii. Let x and y be two sets.

a) Show that {{x}, {x, y}} is a set. Call this set (x, y).
b) Show that (x, y) = (z, t) if and only if x = z and y = t.
iii. Find ∪1, ∪2, ∪3, ∪4.
iv. Find ∪ ∪ 4.
v. Show that ∪℘(x) = x.
vi. Show that ∩∩(x, y) = x. Show that y = (∪∪(x, y)\∩∩(x, y))∪(∩∪(x, y)).
vii. Let X be a set. Show that the diagonal {(x, x) : x ∈ X} is a set.
viii. Let A ⊆ X × Y be a subset of X × Y . Show that the object A−1 defined

by
A−1 = {(y, x) : (x, y) ∈ A}
is a set.
ix. Find a formula φ(Z, T ) such that φ(X × Y, T ) holds if and only if T = X.
Find a formula ψ(Z, T ) such that φ(X × Y, T ) holds if and only if T = Y .
10.5 Functions
In naive set theory, a function f from a set X into a set Y is a rule that assigns
to each element x of X a unique element f (x) of Y . The graph of a function is
the set of pairs (x, y) such that x ∈ X, y ∈ Y and y = f (x).
Since everything should be a set in set theory, we may attempt to define a
function as its graph, hence as a subset F of X × Y with the property that for
every x ∈ X there is a unique y ∈ Y such that (x, y) ∈ F . This attempt is not
too far from the accepted definition of a function. We also need the sets X and
Y in the definition.
Given a function f : X −→ Y , we can extend the set Y to a larger set Y1
to get another function f1 : X −→ Y1 : For each x ∈ X, just let f1 (x) = f (x).
Although the two functions f and f1 take the same values with the same input,
we would like to differentiate them, they should be two different functions. But
this is impossible if we define a function only as its graph. So, rather than
defining a function as its graph, we will define a function as a triple (F, X, Y )
where F ⊆ X ×Y will be the graph of the function. Here is the formal definition:
Let X and Y be two sets. A function or a map is a triple (F, X, Y ) with
the following properties:
i. F is a subset of X × Y .
ii. For each x ∈ X there is a unique y ∈ Y such that (x, y) ∈ F .
The set X is called the domain and the set Y is called the set of arrival
of the function. The set F is called the graph of the function.
In fact we do not need the domain X in the definition of a function, because
given a graph F ⊂ X × Y , the set X can be found back as follows: First of
all note that ∪ ∪ F = X ∪ Y . Now X = {x ∈ ∪ ∪ F : (x, y) ∈ G for some y}.
Given a a set F , we define π1 (F ) as the set {x ∈ ∪ ∪ F : (x, y) ∈ F for some y}.
Clearly π1 (G) is a set. With this notation in hand, a reformed definition of a
graph may be read as follows:
A function or a map is a pair (F, Y ) such that
i. F ⊆ pr1 (F ) × Y ,
ii. For each x ∈ pr1 (F ) there is a unique y ∈ Y such that (x, y) ∈ F .
If f = (F, Y ) is a function, given x ∈ pr1 (F ) the unique element y of Y for
which (x, y) ∈ F is denoted by f (x)
It may happen that we are not so much interested in Y , in which case we
drop Y from the notation and we only say that f is a function (from X into Y ).
If (f, Y ) is a function and (x, y) ∈ f , we write y = f (x).
If (f, Y ) is a function from X into Y , we very often denote this fact as
f : X −→ Y . The “rule” f may be made explicit by the notation x 7→ f (x).
When we say that f : X −→ Y is a function, we assume implicitly that X
and Y are sets and that X 6= ∅.
Very often the condition (ii) is easily checked. Condition (i) may be more
tedious to check. One should not forget to prove (at this beginning stage of our
mathematical education) that f is a set. The fact that f is a subset of X × Y
will then be clear most of the time.
10.5. FUNCTIONS 85
Lemma 10.5.1 (Identity Function) Let X be a set. Then there is a function

IdX : X −→ X that satisfies IdX (x) = x.
Proof: We need to check condition (i) of the definition. For this we just need
to show that the diagonal
δ(X) := {(x, x) : x ∈ X}
is a set. This is given in Exercise vii, page 83. ¤

The function IdX : X −→ X is called the identity function.
Lemma 10.5.2 (Composition) Let X, Y, Z be three sets. Let f : X −→ Y

and g : Y −→ Z be two functions. Define
g ◦ f := {(x, z) : ∃y(y ∈ Y ∧ (x, y) ∈ f ∧ (y, z) ∈ g}.
Then g ◦ f is a function from X into Z. We have
(g ◦ f )(x) = g(f (x)).
Proof: It is clear that g ◦ f is a set and that it is a subset of X × Z. We need

to show condition (ii) of the definition of a function.
Given x ∈ X, there is a y ∈ Y such that (x, y) ∈ f . Since y ∈ Y , there is a
z ∈ Z such that (y, z) ∈ g. Thus (x, z) ∈ g ◦ f .
Assume now that (x, z) and (x, z1 ) are two elements of g ◦ f . We need to
show that z = z1 . By definition of g ◦ f there are y, y1 nY such that
(x, y) ∈ f and (x, y1 ) ∈ f and (y, z) ∈ g and (y1 , z1 ) ∈ g.
The first two conditions imply y = y1 , and now, the last two conditions imply
z = z1 . ¤
The function g ◦ f : X −→ Z is called the composition of f and g.
Lemma 10.5.3 (Restriction) Let f : X −→ Y be a function. Let A ⊆ X.

Then there is a function f|A : A −→ Y such that f (a) = f|A (a) for all a ∈ A.
Proof: Define f|A = f ∩ (A × Y ). ¤

The function f|A is called the restriction of f to A.
Lemma 10.5.4 (Image Restriction) Let f : X −→ Y be a function. Let

Y1 ⊆ Y be such that f (X) ⊆ Y1 . Then there is a function g : X −→ Y1 such
that g(x) = f (x) for all x ∈ X.
Proof: Clearly (f, Y1 ) is a function. ¤

We denote this function as f : X −→ Y1 .
Exercises.
i. Let X and Y be two sets. Show that there is a set whose elements are
exactly the functions from X into Y . We denote this set by X Y .
X
ii. Find ∅.
iii. Show that if we allowed X to be emptyset in the definition of a function
from X into Y , then ∅ Y would be the set of all sets.
iv. Let X be a set. Let π1 (X) := {x : fo some y, (x, y) ∈ X}. Show that
π1 (X) is a set.
v. Let X and Y be sets. Show that pr1 : X × Y −→ X given by pr1 (x, y) = x
and pr2 : X × Y −→ Y given by pr1 (x, y) = y are functions. These are
called first and second projections.
vi. Let f : X −→ Y be a function.
For A ⊆ X, show that f (A) := {f (a) : a ∈ A} is a set. It is called the
image of A under f . The set f (X) is called the image of f .
For two subsets A and B of X show that f (A∪B) = f (A)∪f (B) and that
f (A ∩ B) ⊆ f (A) ∩ f (B). Show that the equality in the second formula
does not always hold.
vii. Let X and Y be two sets.
a) Show that there is a set whose elements are bijections from X into Y .
b) Show that there is a set whose elements are injections from X into Y .
c) Show that there is a set whose elements are surjections from X into Y .
viii. Let X and Y be two sets. Show that there is a function f : X × Y −→

Y × X such that f (x, y) = (y, x). Show that f is a bijection.
ix. Let X, Y and Z be three sets. Show that there is a function f : (X ×

Y ) × Z −→ X × (Y × Z) such that f ((x, y), z) = (x, (y, z)). Show that f
is a bijection.
x. Let f : X −→ Y be a function.
For A ⊆ Y , show that f −1 (A) := {x ∈ X : f (x) ∈ A} is a set. It is called
the inverse image of A under f .
Show that f −1 (Y ) = X.
For two subsets A and B of Y show that f −1 (A ∪ B) = f −1 (A) ∪ f −1 (B)
and that f −1 (A ∩ B) = f −1 (A) ∩ f −1 (B).
Chapter 11
Natural Numbers
11.1 Definition and Basic Properties

Recall that if x is a set, S(x) is defined to be the set x ∪ {x}.
A set x is called inductive if 0 ∈ x and for all y ∈ x, S(y) ∈ x.
Lemma 11.1.1 If x is a set of inductive sets then ∩y∈x y is an inductive set.
Proof: By Theorem 10.3.2, ∩y∈x y is a set. It remains to prove that it is an
inductive set. Since every element of x is inductive, 0 is in every element of
x. Hence 0 ∈ ∩y∈x y. Assume now z ∈ ∩y∈x y. Then z is in every element of
x. Since every element of x is inductive. S(z) is in every element of x as well.
Thus S(z) ∈ ∩y∈x y. ¤
Until now, there was no axiom to ensure the existence of an infinite set. Our
next axiom will give rise to an infinite set (the term “infinite” is to be defined
mathematically later on).
A6. Inductive Sets. There is an inductive set.

Theorem 11.1.2 There is a smallest inductive set. In other words, there is an
inductive set which is a subset of all inductive sets. This smallest inductive set
is (of course!) the intersection of all inductive sets.
Proof: The intersection of all inductive sets is inductive (as in the proof of the
lemma above), but a priori it is not a set. We will prove that it is in fact a set.
Let x be an inductive set (whose existence is insured by A6). The collection c of
inductive subsets of x is a set by Axioms A6 and A3. Then by Theorem 10.3.2,
ω := ∩z∈c z is a set, and by Lemma 11.1.1 ω is an inductive set. We claim that
ω is a subset of all inductive sets. Indeed, let t be any inductive set. By A50 ,
{x, t} is a set. By lemmas 10.3.2 and 11.1.1, x ∩ t is an inductive subset of x.
Thus x ∩ t ∈ c and ω is a subset of x ∩ t. It follows that ω is a subset of t. ¤
From now on we let ω or N denote the smallest inductive set.
87
88 CHAPTER 11. NATURAL NUMBERS
11.2 Well-ordering on ω
A well-ordered set is a linear order with the property that every nonempty
subset has a least element. In this subsection, we will show that ω is a well-
ordered set with respect to the relation ∈.
Lemma 11.2.1 For all x, y ∈ ω, if x ∈ y then x ⊆ y.
Proof: We proceed by induction on y. The statement holds trivially if y =

0 = ∅, since there is no such x. Assume the statement holds for y. We proceed
to show that the statement holds for S(y). Let x ∈ S(y). Since S(y) = y ∪ {y},
either x ∈ y or x ∈ {y}. In the first case x ⊆ y by our assumption. In the
second case x = y. Thus in both cases x ⊆ y, proving our lemma. ¤
Lemma 11.2.2 For all x ∈ ω, either ∅ ∈ x or x = ∅.
Proof: We proceed by induction on x. The statement holds trivially if x = 0 =

∅. Assume the statement holds for x. Since x ⊆ S(x), if ∅ ∈ x, then ∅ ∈ S(x).
If ∅ = x, then S(x) = S(∅) = {∅} and it contains ∅ as an element. ¤
Lemma 11.2.3 For all x ∈ ω, x 6∈ x.
Proof: We proceed by induction on x. The statement is clear if x = 0. Assume

the statement holds for x, i.e. that x 6∈ x. We prove that S(x) 6∈ S(x). Assume
otherwise, i.e. that S(x) ∈ S(x). Thus S(x) ∈ x ∪ {x}. If S(x) ∈ {x}, then
S(x) = x and x ∈ S(x) = x, contradicting our assumption that x 6∈ x. Hence
S(x) ∈ x. By Lemma 11.2.1, S(x) ⊆ x. Since x ∈ S(x), we again get x ∈ x. ¤
Lemma 11.2.4 For all x, y ∈ ω, if y ∈ x then either S(y) ∈ x or S(y) = x.
Proof: We proceed by induction on x. The statement is clear if x = 0. Assume

the statement holds for x. We prove that the statement holds for S(x). Let
y ∈ S(x) = x ∪ {x}. If y ∈ {x}, then y = x and so S(y) = S(x). If y ∈ x, then
by induction either S(y) ∈ x or S(y) = x. In either case S(y) ∈ S(x). ¤
Theorem 11.2.5 The relation ∈ is a linear order on ω.
Proof: We first show that (ω, ∈) is a linear order.

i. By Lemma 11.2.3, for all x ∈ ω, x 6∈ x.
ii. Let x, y, z ∈ ω be such that x ∈ y and y ∈ z. By Lemma 11.2.1, y ⊆ z.
Thus x ∈ z.
iii. Let x, y ∈ ω. We want to show that either x ∈ y or x = y or y ∈ x. We
do this by induction on x. In other words we show by induction on x that “for
all y ∈ ω, either x ∈ y or x = y or y ∈ x”. If x = 0, then this is given by Lemma
11.2.2. Assume the statement holds for x. We show it for S(x). Let y ∈ ω.
Case 1. If y ∈ x, then y ∈ x ⊆ S(x), so y ∈ S(x).
Case 2. If y = x, then y ∈ S(x).
11.3. PEANO’S AXIOMS 89
Case 3. If x ∈ y, then by Lemma 11.2.4, either S(x) ∈ y or S(x) = y. This

proves the lemma. ¤
We now show that any nonempty subset of ω has a least element (for ²). Let
X be a subset of ω without a least element. We will prove by induction on x
the following statement: “For all y ∈ ω, if y ∈ x then y 6∈ X”. Let φ(x) be this
statement.
Clearly φ(x) holds for x = 0. Assuming φ(x) holds, we show that φ(S(x)).
Assume not. Let y ∈ S(x) = x ∪ {x} be an element of X. If y ∈ x then by
induction y 6∈ X, a contradiction. Thus y = x. But then, x is the least element
of X, a contradiction again. This proves the statement ∀x φ(x).
Now we show that X = ∅. Assume otherwise. Let x ∈ X. Then φ(S(x))
does not hold. A contradiction. ¤
Exercises.
i. Show that if x ∈ ω then x ⊆ ω.
Proof: If x = 0 = ∅, that is clear. Assuming it holds for x, we show that it

holds for S(x). Since S(x) = x ∪ {x} and since x ⊆ ω by induction, this is clear.
¤
11.3 Peano’s Axioms

In this subsection, we will prove the so-called “Peano’s Axioms” (Theorem
11.3.1). After the proof of the Peano’s Axioms, we will not need for a long
time what has been done until now in the preceding sections and subsections.
To develop arithmetic and to construct other number sets, we will only use the
Peano’s Axioms.
As it is customary, we will let N = ω.
Theorem 11.3.1 (Peano’s Axioms) PA1. S is a one to one function from

N into N.
PA2. If X is a subset of N with the property that

a) 0 ∈ X,
b) For all x ∈ X, S(x) ∈ X,
then X = N.
Proof: PA2. By hypothesis X is an inductive set. By Theorem 11.1.2, X = N.

PA1. We first prove that S is a function. We only need to prove that
{(x, S(x)) : x ∈ N} is a set. This is easy and we leave it to the reader.
We now show that S is one to one. Assume S(x) = S(y) and x = y. Then
x ∪ {x} = y ∪ {y}. Then x, which is an element of the left hand side is an
element of the right hand side. Thus x ∈ y ∪ {y}. If x ∈ {y} then x = y. Thus
x 6∈ {y} and x ∈ y. By Lemma 11.2.1, x ⊆ y. Similarly y ⊆ x.
Lemma 11.3.2 S(N) = N \ {0}.

Proof: Let X = {0} ∪ {x ∈ N : there exists y ∈ N such that S(y) = x}. We
will show that X = N by using PA2. By definition 0 ∈ X. Assume x ∈ X.
Then x ∈ N. So S(x) ∈ X by definition of X. By PA2, X = N.
Thus N \ {0} ⊆ S(N). Since S(N) ⊆ N, for the reverse inclusion, we just
need to show that 0 6∈ S(N). But x ∈ S(x) for any x ∈ N, so that S(x) 6= ∅ = 0.
¤
From now on we will forget about the construction of ω and we will only use
the properties expressed in Theorem 11.3.1.
11.4 Addition of Natural Numbers

A subset A of N × N × N will be called additive (only for this subsection) if
i. (x, 0, x) ∈ A for all x ∈ N.
ii. If (x, y, z) ∈ A then (x, S(y), S(z)) ∈ A.
Theorem 11.4.1 (Addition) There is a smallest additive set and it is the
graph of a function from N × N into N.
Proof: Let A be the intersection of all additive subsets of N × N × N. (Note
that N × N × N is an additive set). Clearly A is a set and is an additive set.
Therefore it is the smallest additive set. We will show that for all (x, y) ∈ N × N
there exists a unique z ∈ N such that (x, y, z) ∈ A. We proceed by induction on
y.
Suppose first y = 0. Then (x, 0, x) ∈ A. Assume (x, 0, z) ∈ A with z 6= x.
Consider the set A1 := A \ {(x, 0, z)}. We now show that A1 is an additive set.
Clearly (x1 , 0, x1 ) ∈ A for all x1 ∈ N because none of the elements of this kind
was taken away from A and they are all in A1 . This proves (i) for A1 . Assume
(x, y, z) ∈ A1 . Then (x, y, z) ∈ A. Thus (x, S(y), S(z)) ∈ A. Since S(y) 6= 0,
(x, S(y), S(z)) ∈ A1 as well. This proves (ii). Therefore A1 is an additive set.
Since A is the smallest additive set, A1 = A and (x, 0, z) 6∈ A for z 6= x.
Suppose now we know the result for y and we proceed to prove it S(y).
Let x ∈ N. By induction, there is a z ∈ N such that (x, y, z) ∈ A. Thus
(x, S(y), S(z)) ∈ A. Assume there exists a z1 ∈ N such that z1 6= S(z) and
(x, S(y), z1 ) ∈ A. Consider the set A1 = A \ {(x, S(y), z1 )}. We will show
that A1 is an additive set. Let x2 ∈ N. Then (x2 , 0, x2 ) ∈ A. Since the
element (x, S(y), z1 ) missing in A1 cannot have the second coordinate equal to
0, (x2 , 0, x2 ) ∈ A. This proves (i). To prove (ii), let (x2 , y2 , z2 ) ∈ A1 . Then
(x2 , y2 , z2 ) ∈ A, so that (x2 , S(y2 ), S(z2 )) ∈ A. Assume (x2 , S(y2 ), S(z2 )) 6∈ A.
Then (x2 , S(y2 ), S(z2 )) = (x, S(y), z1 ). Thus x2 = x and S(y2 ) = S(y) and
S(z2 ) = z1 . Since S is one to one, x2 = x and y2 = y and S(z2 ) = z1 . So
(x, y, z2 ) = (x2 , y2 , z2 ) ∈ A. Since (x, y, z) ∈ A, by induction, we get z2 = z.
Thus S(z) = S(z2 ) = z1 , a contradiction to the fact that S(z) 6= z1 . This proves
(ii). Hence A1 is an additive set. But this contradicts the fact that A is the
smallest additive set. ¤
11.4. ADDITION OF NATURAL NUMBERS 91
We call this function addition and let x + y = z if (x, y, z) ∈ A. We have
A1. x + 0 = x for all x ∈ N.
A2. x + S(y) = S(x + y) for all x, y ∈ N.
Lemma 11.4.2 For all x ∈ N, if S(x) = x + 1.
Proof: x + 1 = x + S(0) = S(x + 0) = S(x). ¤
Theorem 11.4.3 2 + 2 = 4.
Proof: 2 + 2 = 2 + S(1) = S(2 + 1) = S(2 + S(0)) = S(S(2 + 0)) = S(S(2)) =

S(3) = 4. ¤
Lemma 11.4.4 (Associativity of Addition) For all x, y, z ∈ N, x + (y +

z) = (x + y) + z.
Proof: By induction on z. Assume first z = 0. Then x + (y + z) = x + (y + 0) =

x + y = (x + y) + 0 = (x + y) + z. Now we assume the statement holds for z
and we prove it for S(z): x + (y + S(z)) = x + S(y + z) = S(x + (y + z)) =
S((x + y) + z) = (x + y) + S(z). ¤
Lemma 11.4.5 For all x ∈ N, 0 + x = x.
Proof: This is clear if x = 0. Assume the result for x. Then 0 + S(x) =

S(0 + x) = S(x). ¤
Lemma 11.4.6 (Right Simplification for Addition) For all x, y, z ∈ N,

if x + z = y + z then x = y.
Proof: This is clear if z = 0. Assume the result for z. Assume also that
x + S(z) = y + S(z). Then S(x + z) = S(y + z). Since S is one to one, this
implies that x + z = y + z. By induction x = y. ¤
Lemma 11.4.7 For all x, y ∈ N, if y + S(x) = S(y) + x.
Proof: If x = 0, then the statement says that y + 1 = S(y). But y + 1 =

y + S(0) = S(y + 0) = S(y) = S(y) + x. Assume the result for x. Then
y + S(S(x)) = S(y + S(x)) = S(S(x) + y) = S(x) + S(y) = ... Do it together
with next lemma... ¤
Lemma 11.4.8 (Commutativity of Addition) For all x, y ∈ N, if x + y =

y + x.
Proof: If x = 0, then this is just Lemma 11.4.5. Assume the result for x. Then
by Lemma 11.4.7 and induction, S(x) + y = x + S(y) = S(x + y) = S(y + x) =
y + S(x). ¤
Corollary 11.4.9 (Left Simplification for Addition) For all x, y, z ∈ N,

if x + y = x + z then y = z.
Proof: By Lemmas 11.4.8 and 11.4.6. ¤
Lemma 11.4.10 For all x, y ∈ N, if x + y = 0 then x = y = 0.
Proof: If y = 0, the result is clear. Otherwise, y = S(z) for some z and

x + y = x + S(z) = S(x + z) 6= 0. ¤
Exercises.
i. Suppose your computer knows how to add 1 to a natural number, knows

how to subtract 1 from a positive natural number and recognizes 0. Teach
how to add two natural numbers to your computer.
11.5 Multiplication of Natural Numbers

A subset M of N×N×N will be called multiplicative (only for this subsection)
if
i. (x, 0, 0) ∈ M for all x ∈ N.
ii. If (x, y, z) ∈ M then (x, S(y), z + x) ∈ M .
Theorem 11.5.1 (Multiplication) There is a smallest multiplicative set and

it is a function
Proof: Let M be the intersection of all the multiplicative subsets of N × N × N.

(Note that N × N × N is a multiplicative set). Clearly M is a set and is a
multiplicative set. Therefore it is the smallest multiplicative set. We will show
that for all (x, y) ∈ N × N there exists a unique z ∈ N such that (x, y, z) ∈ M .
We proceed by induction on y.
Suppose first y = 0. We know that (x, 0, 0) ∈ A. Assume (x, 0, z) ∈ M with
z 6= 0. Consider the set M1 := M \ {(x, 0, z)}. We will show that M1 is an
inductive set. Clearly (x1 , 0, 0) ∈ M for all x1 ∈ N because none of the elements
of this kind was taken away from M and they are all in M1 . This proves (i) for
M1 . Assume (x, y, z) ∈ M1 . Then (x, y, z) ∈ M . Thus (x, S(y), z + x) ∈ M .
Since S(y) 6= 0, (x, S(y), z + x) ∈ M1 as well. This proves (ii). Therefore M1
is a multiplicative set. Since M is the smallest multiplicative set, M1 = M and
(x, 0, z) 6∈ M for z 6= 0.
Suppose now we know the result for y. We prove it S(y). Let x ∈ N. By
induction, there is a z ∈ N such that (x, y, z) ∈ M . Thus (x, S(y), z + x) ∈ M .
Assume there exists a z1 ∈ N such that z1 6= z + x and (x, S(y), z1 ) ∈ M .
Consider the set M1 := M \ {(x, S(y), z1 )}. We will show that M1 is a multi-
plicative set. Let x2 ∈ N. Then (x2 , 0, 0) ∈ M . Since the element (x, S(y), z1 )
missing in M1 cannot have the second coordinate equal to 0, (x2 , 0, 0) ∈ M1 .
This proves (i). To prove (ii), let (x2 , y2 , z2 ) ∈ M1 . Then (x2 , y2 , z2 ) ∈ M ,
11.5. MULTIPLICATION OF NATURAL NUMBERS 93
so that (x2 , S(y2 ), z2 + x2 ) ∈ M . Assume (x2 , S(y2 ), z2 + x2 ) 6∈ M1 . Then

(x2 , S(y2 ), z2 + x2 ) = (x, S(y), z1 ). Thus x2 = x and S(y2 ) = S(y) and
z2 + x2 = z1 . Since S is one to one, x2 = x and y2 = y and z2 + x = z1 .
So (x, y, z2 ) = (x2 , y2 , z2 ) ∈ M . Since (x, y, z) ∈ M , by induction, we get z = z2
and z1 = z2 + x = z + x, a contradiction to the fact that z + x 6= z1 . This proves
(ii). Hence M1 is a multiplicative set. But this contradicts the fact that M is
the smallest multiplicative set. ¤
We call this function multiplication and let x × y = z if (x, y, z) ∈ M . As
it is customary, instead of x × y, we may just write xy. We have,
M1. x0 = 0 for all x ∈ N.

M2. xS(y) = xy + x for all x, y ∈ N.
Theorem 11.5.2 2 × 2 = 4.
Proof: We directly compute using definitions and the previous results about
addition: 2 × 2 = 2 × S(1) = 2 × 1 + 2 = 2 × S(0) + 2 = (2 × 0 + 2) + 2 =
(0 + 2) + 2 = 2 + 2 = 4. ¤
Lemma 11.5.3 (Right Identity Element for Multiplication) For all x ∈

N, if x1 = x.
Proof: This follows from the definition of the multiplication and Lemma 11.4.5:
x1 = xS(0) = x0 + x = 0 + x = x. ¤
Lemma 11.5.4 For all x ∈ N, 0x = 0.
Proof: If x = 0 this is just the definition. We assume 0x = 0 and prove that

0S(x) = 0: 0S(x) = 0x + 0 = 0 + 0 = 0. ¤
Lemma 11.5.5 (Left Identity Element for Multiplication) For all x ∈

N, 1x = x.
Proof: If x = 0 this is just the definition. We assume 1x = x and prove that

1S(x) = S(x): 1S(x) = 1x + 1 = x + 1 = S(x). ¤
Lemma 11.5.6 (Left Distributivity) For all x, y, z ∈ N, x(y+z) = xy+xz.
Proof: Assume first z = 0. Then x(y + 0) = xy = xy + 0 = xy + x0 = xy + xz.

Now assume the statement holds for z. We show it holds for S(z): x(y +S(z)) =
xS(y + z) = x(y + z) + x = (xy + xz) + x = xy + (xz + x) = xy + xS(z). ¤
Lemma 11.5.7 (Associativity of Multiplication) For all x, y, z ∈ N, x(yz) =

(xy)z.
Proof: If z = 0, then x(yz) = x(y0) = x0 = 0 = (xy)0 = (xy)z. Assume now

the statement holds for z. We show it holds for S(z) by using Lemma 11.5.6:
x(yS(z)) = x(yz + y) = x(yz) + xy = (xy)z + xy = (xy)S(z). ¤
Lemma 11.5.8 (Right Distributivity) For all x, y, z ∈ N, (y + z)x = yx +

zx.
Proof: If z = 0, then (y + z)x = (y + z)0 = 0 = 0 + 0 = y0 + z0 = yx +

zx. Assume now the statement holds for x. We show it holds for S(x) by
using Lemma 11.5.5 and the commutativity and the associativity of addition:
(y + z)S(x) = (y + z)(x + 1) = (y + z)x + (y + z)1 = (yx + zx) + (y + z) =
(yx + y) + (zx + z) = yS(x) + zS(x). ¤
Lemma 11.5.9 (Commutativity of Multiplication) For all x, y ∈ N, if

xy = yx.
Lemma 11.5.10 (Simplification for Multiplication) For all x, y, z ∈ N,

if xy = xz and x 6= 0, then y = z.
Lemma 11.5.11 For all x, y ∈ N, if xy = 0, then either x = 0 or y = 0.
Proof: Assume x 6= 0. Since xy = 0 = x0, by Lemma 11.5.10, y = 0. ¤
Exercises.
i. Suppose your computer knows how to add 1 to a natural number, knows

how to subtract 1 from a positive natural number and recognizes 0. Teach
how to multiply two natural numbers to your computer. (Hint: use Exer-
cise i, page 92).
11.6 Well-Ordering of Natural Numbers

In Theorem 11.2.5, we showed that ∈ is a well-ordering of ω = N. Here in this
section, we will give another definition of the same ordering and reprove that
this is a well-ordering using only the Peano’s Axioms and its consequences.
For x, y ∈ N, define x < y if there exists an a ∈ N \ {0} such that x + a = y.
Lemma 11.6.1 For all x ∈ N, x < S(x) = x + 1.
Proof: By Lemma 11.4.2, S(x) = x + 1. Since 1 = {0}, 1 6= ∅ = 0. Thus

x < x + 1. ¤
Theorem 11.6.2 (Well-ordering of N) < is a well-ordering on N.
Proof: i. Assume x ∈ N and x < x. Then x + a = x for some a ∈ N \ {0}.

Then, since x + a = x = x + 0, by Lemma 11.4.9, a = 0, a contradiction.
ii. Assume now x, y, z ∈ N, x < y and y < z. Then there are nonzero
natural numbers a and b such that x + a = y and y + b = z. Then x + (a + b) =
(x + a) + b = y + b = z and a + b 6= 0 by Lemma 11.4.10. Thus x < z.
It follows that (N, <) is a poset.
iii. Let x, y ∈ N. We want to show that either x < y or x = y or y < x.
11.6. WELL-ORDERING OF NATURAL NUMBERS 95
Suppose first x = 0. If y 6= 0, then, since x + y = 0 + y = y, we have x < y.

Assume now that we know the result for x. We will prove it for S(x). Let
y ∈ N. Either y < x or y = x or x < y.
Case 1. If y < x, then by Lemma 11.6.1 and part (ii), y < S(x).
Case 2. If y = x, then by Lemma 11.6.1 and part (ii), y < S(x).
Case 3. If x < y, then there exists an a ∈ N \ {0} such that x + a = y.
Subcase 3.1. If a = 1, then by Lemma 11.6.1, y = x + 1 = S(x).
Subcase 3.2. If a 6= 1, then by Lemma 11.3.2, a = S(b) for some b 6= 0. Then
y = x + a = x + S(b) = S(x) + b by Lemma 11.4.7. Thus y > S(x).
It follows that (N, <) is a total order.
iv. We now show that given x, y ∈ N, y ∈ N, if y < S(x) then y ≤ S(x).
Since y < S(x), there is an a ∈ N\{0} such that y +a = S(x). By hypothesis
a 6= 0. Thus a = S(b) for some b. Thus S(y + b) = y + S(b) = y + a = S(x) and
so y + b = x. If b = 0 then y = x. If b 6= 0 then y < x. This proves the claim.
v. Let X ⊆ N be a subset that does not have a least element. We will show
that X = ∅. We will show by induction on x that “for all y ∈ N if y < x then
y 6∈ X”. This statement clearly holds for x = 0. Assume it holds for x. Let
y < S(x) be an element of X. then, by induction y 6< x. By part (iv), y = x.
But then x is the least element of X, a contradiction. This proves the statement
“for all y ∈ N if y < x then y 6∈ X”.
Now if a ∈ X, then this statement would be false for S(a). Thus X = ∅. ¤
Lemma 11.6.3 For all x, y, z ∈ N, x < y if and only if x + z < y + z.
Proof: (⇒) Suppose x < y. Let a 6= 0 be such that x + a = y. Then
(x + z) + a = (x + a) + z = y + z. So x + z < y + z.
(⇐) Suppose x + z < y + z. Let a 6= 0 be such that x + z + a = y + z. Then
(x + a) + z = y + z. By simplifying, we get x + a = y. So x + z < y + z.
Lemma 11.6.4 For all x, y, z ∈ N, if y < z and x 6= 0 then xy > xz.
Proof: Let a 6= 0 be such that y + a = z. Now xz = x(y + a) = xy + xa and
xa 6= 0 by Lemma 11.5.11. So xy < xz. ¤
Exercises.
i. Let x ∈ N. Show that there is no y ∈ N such that x < y < S(x).
ii. Show that x + x = 2 × x for all x ∈ N.
iii. Show that 2N ∩ (2N + 1) = ∅.
iv. Show that if the structure (N1 , S1 , 01 ) satisfies the Peano Axioms, then
(N1 , S1 , 01 ) ' (N, S, 0) and the isomorphism is unique.
v. Show that given x, y ∈ N, y ∈ N, if x < y then S(x) ≤ y.
Proof: Since x < y, there is an a ∈ N \ {0} such that x + a = y. By
hypothesis a 6= 0. Thus a = S(b) for some b. Thus y = x + a = x + S(b) =
S(x) + b by Lemma 11.4.7 and so S(x) ≤ y. ¤
vi. Show that if x < y < x + 2, then y = x + 1.
vii. Show that the ordering < defined in this section is the same as the relation
∈. I.e. show that for x, y ∈ N, x > y if and only if x ∈ y.
11.7 Finite and Infinite Sets

A set is called finite if it is in bijection with a natural number. A set which
is not finite is called an infinite set. If A is a finite set and if f : A −→ n is
a bijection between A and the natural number n, then letting ai = f −1 (i) for
i ∈ n, we will write
A = {a0 , a1 , . . . , an−1 }.
Given a finite set A, the natural number n as in the definition is unique as the
next lemma shows:
Lemma 11.7.1 Let n, m ∈ ω. If there is a bijection between n and m then

n = m.
Proof: Assume first n = 0. Let f : n −→ m be a bijection. If m 6= 0, then

0 ∈ m and f (x) = 0 for some x ∈ n = 0 = ∅, a contradiction.
Suppose we know the result for n. Let us prove the result for n + 1. Let
f : n + 1 −→ m be a bijection. Let f (n) = x ∈ m. In particular m 6= 0 and
m = S(k) for some k. Define g : n \ {x} −→ k by
½
y if y < x
g(y) =
y−1 if y > x
It is easy to check that g is a bijection. Now the map g ◦ f|n is a bijection from
n onto k. By induction n = k, and so n + 1 = S(n) = S(k) = m. ¤
Thus if A is a finite set, then there is a unique natural number n such that
A and n are in bijection. We set |A| = n and call n the cardinality of A.
Exercises.
i. Show that if A and B are finite sets with |A| = n and |B| = m, then
|A ∪ B| ≤ n + m.
ii. Show that ω is an infinite set.
11.8 Functions Defined by Induction

Theorem 11.8.1 To be announced.
11.9. POWERS 97
11.9 Powers
Corollary 11.9.1 There is a unique map f : N×N\{(0, 0)} such that f (n, 0) =
1 for all n 6= 0, f (0, 1) = 0 and f (n, m + 1) = f (n, m)n for all n, m ∈ N.
We let f (n, m) = nm .
Lemma 11.9.2 For all a, b, c ∈ N for which the operations below are defined,
ab+c = ab ac , (ab )c = abc .
Lemma 11.9.3 Let X be the set of finite subsets of N × N. Then there is a
unique function f : X −→ N such that
f (∅) = 1
and
f ({(n1 , m1 ), . . . , (nk , mk )}) = f ({(n1 , m1 ), . . . , (nk−1 , mk−1 )})nm
k .
k
We let
f ({(n1 , m1 ), . . . , (nk , mk )}) = nm mk
k . . . nk .
k
Exercises.
i. Let X be the set of finite subsets of N. Show that there is a unique function
f : X −→ N such that
f (∅) = 0
and
f ({n1 , . . . , nk }) = f ({n1 , . . . , nk−1 }) + nk .
11.10 Divisibility
Theorem 11.10.1 (Division) Let a, b ∈ N. Assume b 6= 0. Then there are
unique q, r ∈ N such that a = bq + r and r < b.
Proof: We proceed by induction on a to show the existence. If a < b, we can
let q = 0 and r = a. Assume a ≥ b. Let c be such that a = b + c. Since c < a,
by induction there are q1 and r such that c = bq1 + r and r < b. Now, we have,
a = b + c = b + bq1 + r = b(1 + q1 ) + r. We let q = 1 + q1 . This proves the
existence.
Now we prove the uniqueness. Assume bq + r = bq1 + r1 and r < b, r1 < b.
If q = q1 , then, by simplifying we get r = r1 . Assume q 6= q1 . We may assume
without loss of generality that q < q1 . Let p > 0 be such that q + p = q1 . Then
bq + r = bq1 + r1 = b(q + p) + r1 = bq + bp + r1 , so r = bp + r1 ≥ bp ≥ b, a
contradiction. ¤
We say that a natural number a divides another natural number b if ac = b
for some c ∈ N. We then write a|b. Note that all natural numbers divide 0, but
that only 1 divides 1. Note also that 1 divides all natural numbers.
Lemma 11.10.2 The relation “divisibility” is a partial order on N with 1 as

the least element and 0 as the largest element.
This partial order has elements which are immediate successors of 1, as we
will show soon. These numbers are called irreducible numbers. Thus a
natural number p is irreducible if p > 1 and if the only natural numbers that
divide p are 1 and p. In other words, p is irreducible if p > 1 and if p = xy
implies either x or y is p (equivalently 1).
Proposition 11.10.3 Every natural number > 1 is divisible by an irreducible
number.
Proof: Let a > 1 be a natural number. If a is not divisible by a natural
number 6= 1, a, then a is irreducible and so a is divisible by an irreducible
number, namely a. Otherwise, a is divisible by a natural number b 6= 1, a.
By induction b is divisible by an irreducible number. This irreducible number
divides a as well. ¤
Lemma 11.10.4 If a divides x and x + y, then a divides y.
This is clear if a = 0, because then x = x + y = 0. Assume a > 0. Let b, c ∈ N
be such that ab = x and ac = x + y. Since ab = x ≤ x + y = ac, we have b ≤ c.
Thus c = b + z for some z. Now ab + az = a(b + z) = ac = x + y = ab + y, so
az = y and a divides y. ¤
The reader may be used to the term prime instead of irreducible. The
term prime, in general, has another meaning, which happens to coincide with
“irreducible” in the context of natural numbers. We say that a natural number
p > 1 is prime if, for all x, y ∈ N, if p|xy then either p|x or p|y. We will now
show that a natural number is a prime number if and only if it is irreducible.
Theorem 11.10.5 A number is prime if and only if it is an irreducible number.
Proof: Let p be prime. Assume that a|p. Then p = ab for some b. It follows
that p divides ab. Thus p divides either a or b. Assume – without loss of
generality – that p divides a. Then px = a some x. Hence p = ab = pxb. Since
p 6= 0, it follows that xb = 1. Thus b = 1, and so a = p.
Let now p be an irreducible. We will prove that p is a prime. Let p divide
xy. We will show that p divides either x or y. We proceed by induction on
p + x + y. By Theorem 11.10.1, x = pq1 + x1 and y = pq2 + y1 where x1 , y1 < p.
Since xy = (pq1 + x1 )(pq2 + y1 ) = p(pq1 q2 + q1 y1 + q2 x1 ) + x1 y1 , by Lemma
11.10.4, p divides x1 y1 . Assume x1 y1 6= 0. Thus p ≤ x1 y1 < p2 . It follows that
x1 y1 = rp for some r = 1, . . . , p − 1. If r = 1, then either p = x1 or p = y1 , a
contradiction. Let q be an irreducible dividing r. Thus q ≤ r < p. By induction
q divides either x1 or y1 , say q divides x1 . Write x1 = qx2 and r = qr0 . We
have qx2 y1 = x1 y1 = rp = qr0 p and x2 y1 = r0 p. By induction p divides either
x2 or y1 , in which case it divides x or y (respectively). Thus we may assume
that x1 y1 = 0. Hence one of x1 or y1 is 0, say x1 = 0. Then x = pq1 + x1 = pq1
and p divides x. ¤
11.10. DIVISIBILITY 99
Theorem 11.10.6 (Euclid) The set of prime numbers is infinite (i.e. there
are infinitely many prime numbers).
Proof: Suppose that there are only n prime numbers. Call them p1 , . . . , pn .
Consider the number k = p1 . . . pn + 1. By Proposition 11.10.3, some prime
number pi for i = 1, . . . , k divides k. Since pi divides the product p1 . . . pn + 1,
by Lemma 11.10.4, pi divides 1, a contradiction. ¤
Theorem 11.10.7 Every natural number 6= 0 is a product of finitely many

prime numbers. In other words, for every n ∈ N \ {0}, there are finitely many
distinct prime numbers p1 , . . . , pk and nonzero natural numbers n1 , . . . , nk such
that n = pn1 1 · · · pnk k . Furthermore the decomposition is unique if we convene
that p1 < p2 < . . . < pk .
We first show that the decomposition is unique if it exists:
Lemma 11.10.8 Let p1 < . . . < pr and q1 < . . . < qs be two finite sets of
prime numbers. Let n1 , . . . , nr and m1 , . . . , ms be natural numbers. Assume that
pn1 1 . . . pnr r = q1m1 . . . qsms . Then r = s, pi = qi and ni = mi for all i = 1, . . . , r.
Proof: Assume the hypotheses. We proceed by induction on n1 + . . . + nr .

We may assume that all ni and mj are positive. If n1 + . . . + nr = 0 (or 1),
then the lemma is obvious. Let N = pn1 1 . . . pnr r = q1m1 . . . qsms . Let p be the
smallest prime dividing N . Then p divides one of the p1 , . . . , pr (this can be
proven by induction on m1 + · · · + ms ). Thus p = pi some i = 1, . . . , r. But p
is the smallest prime dividing N . So p = p1 . Similarly p = q1 . Simplifying, we
get pn1 1 −1 . . . pnr r = q1m1 −1 . . . qsms . By induction we get the result. ¤
Now we prove the existence part of Theorem 11.10.7. If n = 1, then take
k = 0. If n is an prime number then take k = 1, p1 = n, n1 = 1. Otherwise
there is a prime p < n that divides n. Let n = pm. By induction there are
finitely many distinct prime numbers p1 , . . . , pk and nonzero natural numbers
n1 , . . . , nk such that m = pn1 1 · · · pnk k . If p = pi for some i = 1, . . . , k, then
n = pn1 1 · · · pni i +1 . . . pnk k . If p 6= pi for any i = 1, . . . , k, then n = pn1 1 . . . pnk k p.
This proves Theorem 11.10.7. ¤
Exercises.
i. Suppose n|m and m 6= 0. Show that n ≤ m.
ii. Suppose n|m and m|n. Show that n = m.
iii. Show that if a|b then ac |bc .
iv. Analyzing carefully the inductive nature of Theorem 11.10.1, write a com-
puter program that knows how to divide a natural number into another
(nonzero) natural number. Hint: See Exercise i, 92 and Exercise i, 94.
11.11 Uniqueness of N
In this subsection, we show that if (N1 , S1 , O1 ) is a structure satisfying the
Peano’s Axioms (defined below), then there is a (unique) bijection f : N −→ N1
such that f (S(x)) = S(f (x)) and f (0) = 01 for all x ∈ N. This will show that a
structure satisfying Peano’s Axioms is unique “up to isomorphism”. We make
this more precise:
Let N0 be a set, S 0 : N0 −→ N0 a function and 00 ∈ N0 an element. Suppose
that the triple (N0 , S 0 , 00 ) satisfies the following:
PA1. S 0 is a one to one function from N0 into N0 .
PA2. If X is a subset of N0 with the property that
a) 00 ∈ X,
b) For all x ∈ X, S 0 (x) ∈ X,
then X = N0 .
Then we say that the triple (N0 , S 0 , 00 ) satisfies the Peano axioms. Of course,
for a triple that satisfies the Peano axioms, all the results of the previous sections
xxx hold.
Theorem 11.11.1 Any two triples (N0 , S 0 , 00 ) and (N00 , S 00 , 000 ) satisfying the
Peano Axioms are “isomorphic”, i.e. there is a bijection f : N00 −→ N0 such
that f (000 ) = 00 and f (S 00 (x)) = S 0 (f (x)) for all x ∈ N00 .
Proof: Since the composition and the inverse of isomorphisms is clearly an
isomorphism, we may suppose that (N00 , S 00 , 000 ) = (N, S, 0). Define f : N −→ N0
by the rule,
f (0) = 00
and
f (S(x)) = S 0 (f (x)).
By Theorem 11.8.1 f is a function. We just need to prove that it is one to one
and that it is unique.
We first show that f is one to one. Assume f (x) = f (y). We will show that
x = y. We proceed by induction on x.
Assume x = 0. Then f (y) = 00 . If y 6= 0, then y = S(x1 ) and so 00 = f (y) =
f (S(x1 )) = S 0 (f (x1 )), contradicting the fact that 00 is not in the range of S 0 .
Thus y = 0 = x.
Assume now x 6= 0. Then x = S(x1 ). So S 0 (f (x1 )) = f (S(x1 )) = f (x) =
f (y). In particular f (x) and f (y) are nonzero. Therefore y 6= 0 as well. Let
y = S(y1 ). Then S 0 (f (x1 )) = f (y) = f (S(y1 )) = S 0 (f (y1 )). Since S 0 is one to
one, it follows that f (x1 ) = f (y1 ). By the inductive hypothesis x1 = y1 . Hence
x = S(x1 ) = S(y1 ) = y. This proves that f is one to one.
We now show that if f 0 is another such function then f = f 0 . By assump-
tion f (0) = 00 = f 0 (0). Assume f (x) = f 0 (x). Then f (S(x)) = S 0 (f (x)) =
S 0 (f 0 (x)) = f 0 (S(x)). This proves that f = f 0 . ¤
Chapter 12
Integers
12.1 Definition of The Set Z of Integers

Prologue. We can add and multiply two elements of N, but in general we
cannot subtract one element from another, for example, although we could
subtract 5 from 6, we cannot subtract 6 from 5. In order to do subtraction, we
need to introduce another number system, namely the ring (term to be defined)
Z of integers. In other words, we want to invent (or create) the number system
Z = {..., −2, −1, 0, 1, 2, 3, ...}
that we are all familiar since our elementary school years.

We want to be able to speak about 5 − 6 and we cannot. Instead of 5 − 6,
let us try to write (5, 6); we want (5, 6) to stand for 5 − 6. There is one problem
though: 5 − 6 = 8 − 9, but (5, 6) 6= (8, 9).
We can try to solve this problem as follows: We may define a relation ≡ on
N × N by
(x, y) ≡ (z, t) ⇐⇒ x − y = z − t,
show that this is an equivalence relation and define −1 as the equivalence class
of (5, 6) (which is the same as the equivalence class of (8, 9)). But there is
a problem: The operation − appears in the definition of ≡, and we did not
define it yet. Here is the clever solution to this problem: Instead of defining the
relation ≡ as above, we define it as follows:
(x, y) ≡ (z, t) ⇐⇒ x + t = z + y.
We first show that this is an equivalence relation.
Lemma 12.1.1 The relation ≡ on N × N defined by
(x, y) ≡ (z, t) ⇐⇒ x + t = z + y
101
102 CHAPTER 12. INTEGERS
Proof: It is clear that (x, y) ≡ (x, y). It is also clear that if (x, y) ≡ (z, t) then
(z, t) ≡ (x, y). It remains to prove the transitivity. Assume (x, y) ≡ (z, t) and
(z, t) ≡ (u, v). Thus x + t = z + y and z + v = u + t. Adding these two equalities
term by term, we get,
(x + t) + (z + v) = (z + y) + (u + t).
Using the associativity and the commutativity of the addition, we get,
x + v + t + z = u + y + z + t.
Simplifying, we find x + v = u + y, i.e. (x, y) ≡ (u, v). ¤
Now we are ready to define the set Z of integers:
Z := N × N/ ≡ .
Note that
(0, 0) ≡ (1, 1) ≡ (2, 2) . . .
(1, 0) ≡ (2, 1) ≡ (3, 2) . . .
(0, 3) ≡ (1, 4) ≡ (2, 5) . . .
The class of (0, 0) will stand for 0, the class of (1, 0) will stand for 1 and the
class of (0, 3) will stand for −3, but not now, in due time.
12.2 Operations +, − and × on Z

In this subsection, we will define three operations on the set Z that we will call
“addition”, “subtraction” and “multiplication”. Recall that the element (x, y)
of Z is supposed to stand for the element “x − y”. Let us, in a moment of
weakness, write
(x, y) = x − y.
Assuming we know the properties of the three operations +, − and × on Z, we
then have,
(x, y) + (z, t) = (x − y) + (z − t) = (x + z) − (y + t) = (x + z, y + t),
(x, y) − (z, t) = (x − y) − (z − t) = (x + t) − (y + z) = (x + t, y + z),
(x, y) × (z, t) = (x − y)(z − t) = (xz + yt) − (xt + yz) = (xz + yt, xt + yz).
So, getting the hint from above, let us (attempt to) define the three operations
+, − and × as
(x, y) + (z, t) = (x + z, y + t),
(x, y) − (z, t) = (x + t, y + z),
(x, y) × (z, t) = (xz + yt, xt + yz).
But in order to be allowed to give such definitions, we should first prove that
we are allowed to do so. Let us explain what we mean on the first example +.
It is very possible that
(x, y) = (x1 , y1 )
12.2. OPERATIONS +, − AND × ON Z 103
and
(z, t) = (z1 , t1 ),
but that
(x + z, y + t) 6= (x1 + z1 , y1 + t1 ),
in which case we are not allowed to define the addition of (x, y) and (z, t) as
(x + z, y + t), because this definition depends not on the equivalence classes
(x, y) and (z, t), but on the choice of the elements (x, y) and (z, t). Fortunately,
this problem does not occur, all that should work does work:
Lemma 12.2.1 Assume (x, y) = (x1 , y1 ) and (z, t) = (z1 , t1 ). Then,

i. (x + z, y + t) = (x1 + z1 , y1 + t1 ).
ii. (x + t, y + z) = (x1 + t1 , y1 + z1 ).
iii. (xz + yt, xt + yz) = (x1 z1 + y1 t1 , x1 t1 + y1 z1 ).
Proof: Assume (x, y) = (x1 , y1 ) and (z, t) = (z1 , t1 ). Thus (x, y) ≡ (x1 , y1 )
and (z, t) ≡ (z1 , t1 ), i.e.
x + y1 = x1 + y and z + t1 = z1 + t.
i. Adding term by term, and using associativity and commutativity, we get
x + z + y1 + t1 = x1 + z1 + y + t,
which means exactly that (x+z, y+t) ≡ (x1 +z1 , y1 +t1 ) i.e. that (x + z, y + t) =
(x1 + z1 , y1 + t1 ).
ii. Take the first equality as it is and reverse the second one:
(∗) x + y1 = x1 + y and z1 + t = z + t1 .
Now add them:

x + y1 + z1 + t = x1 + y + z + t1 ,
which means
x + t + y1 + z1 = x1 + t1 + y + z,
which means (x + t, y + z) = (x1 + t1 , y1 + z1 ).
iii. This is slightly more difficult. We want to show that
xz + yt + x1 t1 + y1 z1 = x1 z1 + y1 t1 + xt + yz.
We will eliminate y1 and t1 from the above equality and arrive at a known
equality. Then, going backwards, the equality will be proven.
The quantities y1 and t1 occur four times, twice on the left, twice on the
right. It is easy to deal with the ones on the left where they appear linearly. It
is slightly harder to get rid of the ones at the right hand side where they appear
as a product. We first deal with the right hand side. We add xz + xt1 + zy1
to both sides: (xz + yt + x1 t1 + y1 z1 ) + (xz + xt1 + zy1 ) = (x1 z1 + y1 t1 +
xt + yz) + (xz + xt1 + zy1 ) = (x1 z1 + xt + yz) + (xz + xt1 + zy1 + y1 t1 ) =
(x1 z1 + xt + yz) + (x + y1 )(z + t1 ) = (x1 z1 + xt + yz) + (x1 + y)(z1 + t).

Multiplying out and simplifying, we get
2xz + x1 t1 + y1 z1 + xt1 + zy1 = 2x1 z1 + xt + yz + x1 t + yz1 .
Now there is no more y1 nd t1 left on the right hand side. Now we deal with
the left hand side. We add x1 z, xz1 and 2zx to both sides, factor out x1 and
z1 from the left hand side and use the known equalities (*): (2x1 z1 + xt + yz +
x1 t + yz1 ) + x1 z + xz1 + 2zx = (2xz + x1 t1 + y1 z1 + xt1 + zy1 ) + x1 z + xz1 + 2zx =
2xz + x1 (t1 + z) + (y1 + x)z1 + x(t1 + z) + z(x + y1 ) = 2xz + x1 (t + z1 ) + (y +
x1 )z1 + x(t + z1 ) + z(x1 + y) = 2xz + x1 t + x1 z1 + yz1 + x1 z1 + xt + xz1 + zx1 + zy,
i.e. 2x1 z1 + xt + yz + x1 t + yz1 + x1 z + xz1 + 2zx = 2xz + x1 t + x1 z1 + yz1 +
x1 z1 + xt + xz1 + zx1 + zy. Simplifying, we see that this last equality holds.
Now, we can go backwards. ¤
At this point, we can define the three operations +, − and × on Z as above,
i.e.
(x, y) + (z, t) = (x + z, y + t),

(x, y) − (z, t) = (x + t, y + z),
(x, y) × (z, t) = (xz + yt, xt + yz).
For (x, y) ∈ Z, we also let
−(x, y) = (0, 0) − (x, y) = (y, x).
Very often, for α, β ∈ Z, we will write αβ instead of α × β.

Before investigating the properties of these operations, we will define an
ordering on Z.
Exercises.
i. Show that for all α, β ∈ Z,
α−β = α + (−β)
−(α − β) = −α + β
−(−α + β) = α−β
−(−α − β) = α + β.
ii. Show that the equation α + α = (1, 0) has no solution in Z.
iii. Show that the equation α + α = (0, 1) has no solution in Z.
iv. Show that the equation α + α + α = (2, 0) has no solution in Z.
v. Find all solutions of αβ = (1, 0) in Z.

12.3. ORDERING ON Z 105
12.3 Ordering on Z
We would like to define the ordering on Z as follows: For (x, y), (z, t) ∈ Z,
(x, y) < (z, t) ⇐⇒ x + t < z + y.
But for this definition to make sense, we need to prove the following result:
Lemma 12.3.1 Assume (x, y) ≡ (x1 , y1 ) and (z, t) ≡ (z1 , t1 ). Assume that
x + t < z + y. Then x1 + t1 < z1 + y1 .
Proof: The hypotheses say that
(1) x + y1 = x1 + y
(2) z + t1 = z1 + t
(3) x+t<z+y
We have to prove that x1 + t1 < z1 + y1 . We compute using (1), (2) and (3):
x1 + y + t1 + z = x + y1 + z1 + t = z + y + y1 + z1 ,
so, simplifying, we get x1 + t1 < y1 + z1 . ¤

According to the lemma, we have the right to define the relation < on Z as
above:
(x, y) < (z, t) ⇐⇒ x + t < y + z.
Now we show that this is a total order on Z.
Lemma 12.3.2 < is a total order on Z. Furthermore, any element of Z has

an immediate successor and an immediate predecessor.
Proof: Irreflexivity. Easy.

Transitivity. Assume (x, y) < (z, t) and (z, t) < (u, v). Thus x + t < z + y
and z + v < v + t. Hence x + t < v + t and so (x, y) < (u, v).
Comparability. We will use the fact that < is a total order on N. Let
(x, y) and (z, t) be two elements of Z. If x + t < z + y then (x, y) < (z, t). If
z + y < x + t then (z, t) < (x, y). If x + t = z + y then (x, y) = (z, t).
Immediate Successor. Let (x, y) be an element of Z. It is easy to show
that (x, y) < (x + 1, y) and that there are no elements of Z in between.
Immediate Predecessor. Let (x, y) be an element of Z. It is easy to show
that (x, y + 1) < (x, y) and that there are no elements of Z in between. ¤
Exercises. α, β, γ stand for the elements of Z.

i. Show that if α < β then α + γ < β + γ.
ii. Show that if α < β then −β < −α.
iii. Show that α < β if and only if β − α > 0.
iv. Assume α < β and 0 < γ. Show that αγ < βγ.
v. Assume α < β and γ < 0. Show that βγ < αγ.
vi. Assume α < β and γ < 0. Show that βγ < αγ.
vii. Assume α > 0 and β > 0. Show that αβ > 0.
viii. Assume α < 0 and β < 0. Show that αβ > 0.
12.4 Embedding of (N, +, ×, <) in (Z, +, ×, <)

In “real life”, N is a subset of Z, but as we defined it here, N ∩ Z =. In this
subsection we will remedy this this by transporting N into Z.
Consider the function i : N −→ Z defined by the rule i(x) = (x, 0). Then, i
is an injection and for all x, y ∈ N,
i. i(x + y) = i(x) + i(y).
ii. i(xy) = i(x)i(y).
iii. x < y ⇐⇒ i(x) < i(y).
These properties are easy to check and we leave them to the reader.
This shows that a copy of the structure (N, +, ×, <) exists in Z, namely
the subset i(N) of Z together with the operations + and × and the relation <
restricted to to the subset i(N) of Z.
Lemma 12.4.1 Z = i(N) ∪ −i(N) and i(N) ∩ −i(N) = {0}.
Proof: Let (x, y) ∈ Z. Assume x ≥ y. Then there is an a ∈ N such that y +a =

x. Thus (x, y) = (y + a, y) = (a, 0) = i(a). Assume x < y. Then there is an
a ∈ N such that x + a = y. Thus (x, y) = (x, x + a) = (0, a) = −(a, 0) = −i(a).
This proves that Z = i(N) ∪ −i(N). The last part is easy as well. ¤
From now on we identify the elements of N with the elements of i(N) via the
injection i, in other words, we rename the elements of i(N) by the elements of
N: For x ∈ N, the element (x, 0) will be denoted by x from now on. With this
convention, we get Z = N ∪ −N and for all x ∈ N,
0 stands for (0, 0)

x stands for (x, 0)
−x stands for (0, x).
Now we can write the multiplication and the order on Z in terms of the elements
of N. For all x, y ∈ N,
x(−y) = −(xy) = (−x)y

−x ≤ y
−x ≤ −y ⇐⇒ y ≤ x.
These properties are also easy to check and we leave them to the reader.
12.5. ADDITIVE STRUCTURE OF Z 107
Exercises.
i. Show that 0 × α = 0 for all α ∈ Z. (Here 0 stands for the element

i(0) = (0, 0) of Z).
ii. Show that 1 × α = α for all α ∈ Z. (Here 1 stands for the element
i(1) = (1, 0) of Z).
iii. Show that −1 × α = −α for all α ∈ Z. (Here −1 stands for the element
−i(1) = (0, 1) of Z).
iv. Show that 2 × α = α + α for all α ∈ Z. (Here 2 stands for the element
i(2) = (2, 0) of Z).
v. Show that Z is the smallest subset of Z containing 1 and closed under

subtraction.
12.5 Additive Structure of Z

In this subsection, we will investigate the properties of + and − on Z. From
the above subsection, we will only retain the convention that the element (0, 0)
is denoted by 0.
Theorem 12.5.1 i. For all α, β, γ ∈ Z, (α + β) + γ = α + (β + γ),

ii. For all α ∈ Z, α + 0 = 0 + α = α
iii. For all α ∈ Z there is a β ∈ Z (namely −α) such that α + β = β + α = 0.
iv. For all α, β ∈ Z, α + β = β + α.
Proof: i. Let α = (x, y), β = (x1 , y1 ) and γ = (x2 , y2 ). Then
(α + β) + γ = ((x + x1 ) + x2 , (y + y1 ) + y2 )
and
α + (β + γ) = (x + (x1 + x2 ), y + (y1 + y2 )).
The equality is now clear.
ii, iii, iv. Easy. ¤
We now define the following terms:
−x + y := (−x) + y
−x − y := (−x) + (−y) = (−x) − y
Lemma 12.5.2 For all x, y ∈ Z,
−(x − y) = −x + y
−(−x + y) = x−y
−(−x − y) = x+y

Exercises. We compute in Z. The elements α, β, γ stand for the elements of

Z.
i. Show that if α + γ = β + γ then α = β.
ii. Show that if α + β = α for some α ∈ Z then β = 0.
12.6 Multiplicative Structure of Z

As we let 0 stand for (0, 0) above, in this subsection, we will also let 1 stand for
(1, 0)
Theorem 12.6.1 For all α, β, γ ∈ Z we have i. α(βγ) = (αβ)γ.

ii. α × 1 = 1 × α = α.
iii. αβ = βα.

The theorem above gives the multiplicative properties of Z. The next theo-
rem is about the relationship between addition and multiplication.
Theorem 12.6.2 For all α, β, γ ∈ Z, α(β + γ) = αβ + αγ.
i. Assume αβ = 1. Show that α = β = ±1.
ii. Assume αβ = 0. Show that either α = 0 or β = ±1.
iii. Show that 2x = 3 has no solution in Z.
12.7 Ordering on Z, Revisited

In the two last subsections above we investigated the additive and the multi-
plicative properties of Z. In this subsection, we investigate the relationship of
these two operations to the order relation .
Theorem 12.7.1 Let α, β, γ ∈ Z. Assume that α < β. Then

i. α + γ < β + γ.
ii. If γ > 0, then αγ < βγ.
Absolute Value. For α ∈ Z, we define |α| = max{α, −α}. We call |α|, the
absolute value of α.
12.8. DIVISIBILITY AND SUBGROUPS 109
Sign Function. For α ∈ Z, we define


 1 if α > 0
sgn(α) = 0 if α = 0

−1 if α < 0
sgn is called the sign function.

i. Show that |α| > 0, i.e. is in N.
ii. Show that |α| = 0 if and only if α = 0.
iii. Show that |α| = |β| if and only if α = ±β.
iv. Show that |α + β| ≤ |α| + |β|.
v. Show that |αβ| ≤ |α||β|.
vi. Show that sgn(αβ) = sgn(α) sgn(β).
Q
vii. Show that any α ∈ Z can be written as α = sgn(a) p pordp (α) where the
product ranges over all primes p > 1 and ordp (a) is the largest natural
number such pordp (α) divides |a|.
12.8 Divisibility and Subgroups

We should start by nothing that, in terms of divisibility, an integer n and its
additive inverse −n have no distinction what so ever: an integer divides n if
and only if it divides −n and n divides an integer if and only if −m divides this
integer. Oh! By divisibility we mean the following: An integer n divides an
integer m if there is an integer q such that m = nq.
We first extend Theorem 11.10.1 from N to Z:
Theorem 12.8.1 (Division) Let a, b ∈ Z with b > 0. Then there are unique
q, r ∈ Z such that a = bq + r and 0 ≤ r < b.
Proof: If a ∈ N, then this is just Theorem 11.10.1. Assume a < 0. Then
applying Theorem 11.10.1 to −a we have −a = qb + r where 0 ≤ r < b. If r = 0,
then a = (−q)b and we are done for the existence. Assume 0 < r < b. We have
a = −qb − r = (−q − 1)b + (b − r). Since 0 < b − r < b, this proves the existence.
Now we prove the uniqueness. Assume bq + r = bq1 + r1 where 0 ≤ r, r1 < b.
If q = q1 , then clearly r = r1 . Assume q 6= q1 . Without loss of generality q > q1 .
Then b < b(q − q1 ) = r1 − r ≤ b, a contradiction. ¤
A nonempty subset of Z closed under substraction is called a subgroup of
Z. Note that if n ∈ Z, then the set
nZ := {nz : z ∈ Z}
is a subgroup of Z. We show that these are all the subgroups of Z:
Theorem 12.8.2 (Subgroups of Z) A subgroup of Z is of the form nZ for

some unique n ∈ N.
Proof: Let H be a subgroup of Z. We first prove three easy properties of H.

1) Since H 6= ∅, there is an element h ∈ H. Hence 0 = h − h ∈ H.
2) Let h ∈ H. Then −h = 0 − h ∈ H because 0 ∈ H by above. Hence
−H = H.
3) Let h, k ∈ H. Then h + k = h − (−k) ∈ H. Thus H is also closed under
addition.
Now we classify all subgroups of Z.
If H = {0}, then H = 0Z. So we assume H 6= {0}. Since H = −H, then H
has positive elements. Let n be the smallest positive integer in H. Since n ∈ H,
−n is also in H. Since H is closed under addition, it follows that nZ ⊆ H. We
will show the reverse inclusion. Let h ∈ H. Using Theorem 12.8.1, h = nq + r
for some q ∈ Z and 0 ≤ r < n. But now, r = h − nq and since nq ∈ nZ ⊆ H,
we get r ∈ H. But, unless r = 0, this contradicts the fact that n was chosen to
be the least positive integer in H. Thus r = 0 and h = nq + r = nq ∈ nZ. ¤
We note that for n, m ∈ Z, nZ ⊆ mZ if and only if m|n, in other words the
inclusion order on the set of subgroups of Z is the divisibility order reversed.
Thus a maximal (and proper, i.e. 6= Z) subgroup of Z is of the form pZ for
some prime p ∈ N. It also follows that any ascending chain of subgroups of Z is
stationary, i.e. if H0 ⊆ H1 ⊆ H2 ⊆ . . . is an ascending chain of subgroups, then
there is an n0 such that for all n > n0 , Hn = Hn0 .
Greatest Common Divisor, Least Common Multiple. Let a and b be

two integers which are not both 0. We say that d is the greatest common
divisor of a and b if d is the largest natural number that divides both a and b.
Lemma 12.8.3 For any a, b ∈ Z, gcd(a, b) exists and there are x, y ∈ Z such
that ax + by = gcd(a, b).
Proof: Replacing a and b by |a| and |b|, we may assume that a ≥ 0 and b ≥ 0.
Existence. Since 1 divides both a and b and since any number that divides
both a and b can be at most max(a, b) > 0, the set of natural numbers that
divide both a and b is a finite nonempty set bounded by max(a, b). Therefore
there is a largest such number. This proves the existence of gcd(a, b). We let
d = gcd(a, b).
Second Part. We proceed by induction on max(a, b). If a = 1, then take
x = 1, y = 0. If b = 1, then take x = 0, y = 1. This takes care of the initial
step max(a, b). Assume max(a, b) > 1. If a = b, then d = a and we may take
x = 1, y = 0. Assume a 6= b. Without loss of generality, we may assume that
a > b. Note that the divisors of a and b are the same as the divisors of a − b
and b. Hence gcd(a − b, b) = gcd(a, b) = d. Since max(a − b, b) < a = max(a, b),
by induction there are two integers x and y 0 such that x(a − b) + y 0 b = d, i.e.
xa + (y 0 − x)b = d. Take y = y 0 − x. ¤
12.8. DIVISIBILITY AND SUBGROUPS 111
Let a and b be two nonzero integers. We say that m is the least common
multiple of a and b if m is the least natural number that is divisible by both a
and b.
Lemma 12.8.4 For any a, b ∈ Z\{0}, lcm(a, b) exists and ab = ± gcd(a, b) lcm(a, b).
Proof: Replacing a and b by |a| and |b| again, we may assume that a > 0 and
b > 0. Since a and b both divide ab, lcm(a, b) exists.
Let d = gcd(a, b) and m = lcm(a, b). Let a0 and b be such that a = da0 and
b = db0 . Then ab = d2 a0 b0 . We need to prove that m = da0 b0 .
Since da0 b0 = ab0 = a0 b, a and b both divide da0 b0 .
Let x be divisible by both a and b. Then x = au = bv for some u, v. We
have a0 du = au = x = bv = b0 dv and so a0 u = b0 v. Since a0 and b0 cannot have
a common divisor (otherwise d would be larger), b0 must divide u. (This last
fact needs a serious proof, that we have not undertaken yet. I should
put this somewhere else). Write u = cb0 . Now x = au = acb0 = a0 dcb0 and
so a0 b0 d divides x, in particular a0 b0 d ≤ x. This shows that a0 b0 d is the least
multiple of a and b, i.e. a0 b0 d = m.
There is another way to define gcd and lcm, which is more modern and better
in several ways. If (Hi )i is any family of subgroups of Z, then it is clear that
∩i Hi is also a subgroup of Z. In particular, for any n, m ∈ Z, nZ ∩ mZ = kZ
for some unique k ∈ N. Since nZ ∩ mZ is the largest subgroup contained in nZ
and in mZ, it is clear that k is the smallest natural numbers that is a multiple
of both n and m. We call k the least common multiple of n and m and we
let lcm(n, m) = k.
If (Hi )i is any family of subgroups of Z, then it is clear that
X
Hi = {hi1 + . . . + hik : k ∈ N and hij ∈ Hij for all j = 1, . . . , k}
i
is also a subgroup of Z. In particular, for any n, m ∈ Z, nZ + mZ = kZ for

some unique k ∈ N. Since nZ + mZ is the smallest subgroup containing both
nZ and mZ, it is clear that k is the smallest natural numbers that is divisible
by both n and m. We call k the greatest common divisor of n and m and
we let gcd(n, m) = k. One
Q also writes (n, m) = k sometimes.
Q
Writing n = sgn(n) p pordp (n) and m = sgn(m) p pordp (m) , where p ranges
over all the primes in N, we see that
Y
lcm(n, m) = pmax{ordp (n),ordp (m)}
p
and Y
gcd(n, m) = pmin{ordp (n),ordp (m)} .
p
In particular,
Q min{ordp (n),ordp (m)}+min{ordp (n),ordp (m)}
lcm(n, m) gcd(n, m) = p
Qp ordp (n)+ordp (m) Q ordp (n) Q ordp (m)
= pp = pp pp = |nm|.
Thus we have:
Proposition 12.8.5 lcm(n, m) gcd(n, m) = |nm|.
Two integers are said to be prime to each other if they are both divisible
only by 1 and −1. In other words n and m are prime to each other if and only
if gcd(n, m) = 1.
Theorem 12.8.6 Two integers a and b are prime to each other if and only if
there are integers x and y such that ax + by = 1.
Proof: Suppose there are integers x and y such that ax + by = 1. Then if an

integer divides both a and b, then this integer must divide ax + by = 1, so it
must be ±1.
Conversely, suppose that gcd(a, b) = 1. Then 1 ∈ Z = aZ + bZ. It follows
that there are integers x and y such that ax + by = 1. ¤
Exercises.
i. Show that gcd(gcd(a, b), b) = gcd(a, gcd(b, c)).
ii. Show that lcm(lcm(a, b), b) = lcm(a, lcm(b, c)).
12.9 Euclidean Algorithm

There is a famous algorithm to find the greatest common divisor of two integers,
called Euclidean Algorithm.
Let n and m be two given nonzero integers. We want to find their greatest
common divisor. Set r0 = n and r1 = m. Divide r0 by r1 to get r0 = qr1 + r2
where q ∈ Z and 0 ≤ r2 < r1 . Inductively, suppose we have found rk and rk+1
where 0 ≤ rk+1 < rk . Divide rk by rk+1 to get rk = qrk+1 + rk+2 where q ∈ Z
and 0 ≤ rk+2 < rk+1 . In this way we get a strictly decreasing sequence (rk )k .
At a finite stage n > 1, we must have rn = 0. We claim that rn−1 = gcd(n, m).
Since rk = qrk+1 + rk+2 , it is clear that gcd(rk , rk+1 ) = gcd(rk+1 , rk+2 ),
therefore
gcd(n, m) = gcd(r0 , r1 ) = gcd(r1 , r2 ) = . . . = gcd(rn−1 , rn )
= gcd(rn−1 , 0) = rn−1 .
12.10 Quotients
The reader must have seen how to work “modulo n”. Here we set the founda-
tions.
For x ∈ Z and A ⊆ Z, we let
n + A = {n + a : a ∈ A}
12.11. CHINESE REMAINDER THEOREM 113
and
nA = {na : a ∈ A}.
Let H be a subgroup of Z. Thus H = nZ for some n ∈ N. Define the
following relation on Z:
x ≡n y ⇐⇒ x − y ∈ nZ ⇐⇒ n divides x − y.
It is easy to check that this is an equivalence relation. Let Z/nZ = Z/ ≡n . For

x ∈ Z, we denote the equivalence class of x by x. Thus
x = {y ∈ Z : n divides x − y} = {y ∈ Z : x − y ∈ nZ}
= {y ∈ Z : y − x ∈ nZ} = {y ∈ Z : y ∈ x + nZ} = x + nZ.
Unfortunately, in the notation x, as an element of Z/nZ, i.e. as x + nZ, the

number n does not exist. Therefore the reader should know at any point of our
discussion what the integer n is.
Lemma 12.10.1 Given n, for every x ∈ Z there is a unique r = 0, 1, . . . , n − 1

such that the element x of Z/nZ is equal to r.
Proof: Follows immediately from Theorem 12.8.1. ¤
12.11 Chinese Remainder Theorem

12.12 Subgroups of Z ⊕ Z
Chapter 13
Rational Numbers
13.1 Rational Numbers

Prologue. In Z we can do addition, substraction and multiplication, but not
division. For example we cannot divide 2 by 3 (i.e. the equation 3x = 2 has no
solution in Z), because 2/3 is not in Z. To remedy this situation, we will extend
the structure Z to a larger set where not only we can do addition, substraction
and multiplication compatible with the similar operations in Z, but also division.
We would like to invent numbers that will stand for 2/3 e.g. If we attempt
to denote 2/3 by the pair (2, 3), then we should have (2, 3) = (4, 6) and as we
know this is not the case. We overcome this difficulty by noting that a/b = c/d
if and only if ad = bc. Accordingly, on the set Z × (Z \ {0}), we define the
following relation:
(x, y) ≡ (z, t) ⇐⇒ xt = yz.
We will show that this is an equivalence relation. Then we will let Q to denote
the set Z × (Z \ {0})/ ≡. Finally we will define the four operations as follows:
(x, y) + (z, t) = (xt + yz, yt)

(x, y) − (z, t) = (xt − yz, yt)
(x, y) × (z, t) = (xz, yt)
(x, y)/(z, t) = (xt, yz)
(For the last operation we need z 6= 0). After that, we will define an ordering
on Q. To finish, we will embed Z inside Q.
Lemma 13.1.1 The relation on Z × (Z \ {0})/ ≡ defined by (x, y) ≡ (z, t) ⇐⇒
xt = yz is an equivalence relation.
Proof: a. i. Reflexivity. Let (x, y) ∈ X. Then since xy = yx, we have
(x, y) ≡ (x, y).
ii. Symmetry. Let (x, y), (z, t) ∈ X be such that (x, y) ≡ (z, t). Hence
xt = yz. Therefore zy = tx, implying (z, t) ≡ (x, y).
115
116 CHAPTER 13. RATIONAL NUMBERS
iii. Transitivity. Let (x, y), (z, t), (u, v) ∈ X be such that (x, y) ≡ (z, t)
and (z, t) ≡ (u, v). Hence xt = yz and zv = tu. Multiplying these equalities
side by side, we get xtzv = yztu. Since t 6= 0, by simplifying we get xzv = yzu.
If z 6= 0, then we can simplify further to get xv = yu, hence (x, y) ≡ (u, v).
Assume z = 0. Then xt = yz = 0 and tu = zv = 0. Since t 6= 0, we get
x = u = 0, so that xv = 0 = yu and (x, y) ≡ (u, v) again. ¤
We let Q = Z × (Z \ {0})/ ≡. Now we show that the attempt to define the
four operations as above is successful:
Lemma 13.1.2 Let (x, y), (z, t), (x0 , y 0 ), (z 0 , y 0 ) ∈ Z×(Z\{0}). Assume (x, y) ≡
(x0 , y 0 ) and (z, t) ≡ (z 0 , t0 ). Then
i. (xt + yz, yt) ≡ (x0 t0 + y 0 z 0 , y 0 t0 ).
ii. (xt − yz, yt) ≡ (x0 t0 − y 0 z 0 , y 0 t0 ).
iii. (xz, yt) ≡ (x0 z 0 , y 0 t0 ).
iv. If z 6= 0, then z 0 6= 0 and (xt, yz) ≡ (x0 t0 , y 0 z 0 ).
Proof: i. Assume (x, y) ≡ (x0 , y 0 ) and (z, t) ≡ (z 0 , t0 ). Then xy 0 = yx0 and

zt0 = tz 0 . Multiplying the first one by tt0 and the second one by yy 0 we get
xy 0 tt0 = yx0 tt0 and zt0 yy 0 = tz 0 yy 0 . Adding these two side by side we get
xy 0 tt0 +zt0 yy 0 = yx0 tt0 +tz 0 yy 0 , and factoring, we get (xt+yz)y 0 t0 = yt(x0 t0 +y 0 z 0 ),
meaning (xt + yz, yt) ≡ (x0 t0 + y 0 z 0 , y 0 t0 ).
ii. Similar.
iii. Assume (x, y) ≡ (x0 , y 0 ) and (z, t) ≡ (z 0 , t0 ). Then xy 0 = yx0 and
zt = tz 0 . Multiplying these two side by side, we get xy 0 zt0 = yx0 tz 0 , i.e.
0
xzy 0 t0 = ytx0 z 0 , meaning (xz, yt) ≡ (x0 z 0 , y 0 t0 ).

iv. If z 0 = 0, then zt0 = z 0 t = 0. But t0 6= 0, so z = 0. For the second part,
we need to show that xty 0 z 0 = yzx0 t0 . Left as an exercise. ¤
It follows that we can define the four operations +, −, × and division on
the set Q. As usual, we let αβ denote α × β.
Next we attempt define the order on Q as follows: Let (a, b), (c, d) ∈ Q.
Clearly, (a, b) = (−a, −b), so that we may assume that b > 0 and d > 0. Now
we want to let
(a, b) < (c, d) ⇐⇒ ad < bc.
But for this definition to be valid, we need to prove the following:
Lemma 13.1.3 Let (x, y), (z, t), (x1 , y1 ), (z1 , y1 ) ∈ Z × (N \ {0}). Assume
(x, y) ≡ (x1 , y1 ), (z, t) ≡ (z1 , t1 ) and xt < yz. Then x1 t1 < y1 z1 .
Proof: By assumptions xy1 = x1 y, zt1 = z1 t, y, y1 , t, t1 > 0 and xt < yz.

From the inequalities we get xtt1 y1 < yzt1 y1 . This and the first two equalities
yield x1 ytt1 < yy1 z1 t. Since y > 0 and t > 0, we can simplify to get x1 t1 < y1 z1 .
¤
This shows that the relation < is well defined.
By abuse of language, we will let 0 = (0, 1) ∈ Q and 1 = (1, 1) ∈ Q.
13.1. RATIONAL NUMBERS 117
Theorem 13.1.4 The following hold:

A1. Associativity. For all α, β, γ ∈ Q, (α + β) + γ = α + (β + γ).
A2 Identity Element. For all α ∈ Q, α + 0 = 0 + α = α.
A3 Inverse Element. For all α ∈ Q, there is a unique β ∈ Q such that
α + β = β + α = 0, namely, if α = (a, b), then β = (−a, b).
A4. Commutativity. For all α, β ∈ Q, α + β = β + α.
If Q∗ = Q \ {0}, then we also have:
M1. Associativity. For all α, β, γ ∈ Q∗ , (αβ)γ = α(βγ).
M2. Identity Element. For all α ∈ Q∗ , α × 1 = 1 × α = α.
M3. Inverse Element. For all α ∈ Q∗ , there is a unique β ∈ Q∗ such
that αβ = βα = 1, namely, if α = (a, b), then β = 1/α = (b, a).
M4. Commutativity. For all α, β ∈ Q∗ , αβ = βα.
D. Distributivity. For all α, β, γ ∈ Q, α(β +γ) = αβ +αγ and (α+β)γ =
αγ + βγ.
O. Total Order. < is a total order on Q.
OA. For all α, β, γ ∈ Q, if α < β then α + γ < β + γ.
OM. For all α, β, γ ∈ Q, if α < β and 0 < γ, then αγ < βγ.
Proof: Left to the reader. ¤

α
We let α/β = β = α × 1/β and α−1 = 1/α if α 6= 0.
α+β
Lemma 13.1.5 (Density.) For all α, β ∈ Q, if a < b then α < 2 < β.
Finally we embed Z in Q.
Proposition 13.1.6 Let i : Z −→ Q be the map defined by i(x) = (x, 1). Then
i is a one to one map and for all x, y ∈ Z,
i. i(x + y) = i(x) + i(y).
ii. i(xy) = i(x)i(y).
iii. x < y if and only if i(x) < i(y).
Furthermore, with our earlier convention, we have i(0) = 0 and i(1) = 1.
Also, for any α ∈ Q, there are x ∈ Z, y ∈ N \ {0} such that α = i(x)/i(y).
If we make the assumption that (x, y) = 1, then x and y are unique.
Proof: Easy. ¤
Now we identify Z with the subset i(Z) of Q. With this identification
Exercises.
i. Show that Q has no last or first element (for the order).
ii. Show that there is no x ∈ Q such that x × x = 2.
iii. Show that if αβ = 0 for α, β ∈ Q then either α or β is zero.

118 CHAPTER 13. RATIONAL NUMBERS
13.2 Some Combinatorics

n choose m
Fermat’s little Theorem
Problem. Quotient Fields.
Problem. Subgroups of Q. A subgroup of Q is a nonempty subset of Q

closed under substraction.
Theorem 13.2.1 (Finitely Generated Subgroups of Q) For a/b, c/d ∈ Q,

ha/b, c/di = hgcd(ad, bc)/bdi. Thus a finitely generated subgroup of Q is in fact
generated by one element.
Theorem 13.2.2 (Classification of Subgroups of Q) Let G be a nontrivial

subgroup of Q+ . Then there are unique a ∈ N \ {0} and f : {primes of N} −→
N ∪ {∞} such that
G = Q(a, f ) = a{x/pn1 1 . . . pnk k : k ∈ N, pi prime and ni ≤ f (pi ) forall i}.

Chapter 14
Real Numbers
119
120 CHAPTER 14. REAL NUMBERS
Chapter 15
Well-Ordered Sets
15.1 Definitions and Examples

A totally ordered structure (X, <) is called a well-ordered set if any nonempty
subset of X has a least element.
Examples.
i. Easy Examples. By definition ∅ is a well-ordered set. We have seen

that ω is a well-ordered set with its “natural” order. Any subset of a well-
ordered set is a well-ordered set with the restricted order. In particular
each natural number n is a well ordered set. The sets Z, Q, Q≥0 (together
with the natural order on them) are not well-ordered structures.
ii. Translation. Let (X, <) be a well-ordered structure. Let f : X −→ Y

be a bijection from X onto Y . Then we can translate the well-order
structure of X onto Y via f as follows: Declare that, for y, y 0 ∈ Y ,
y < y 0 if f −1 (y) < f −1 (y 0 ). This gives a well-order structure on Y , which
essentially the one of X. We will see that the two well-ordered sets are
“isomorphic”.
iii. Union – Addition. Let (X, <1 ) and (Y, <2 ) be two well-ordered sets.
Replacing X by X × {0} and Y by Y × {1}, and translating the orders
of X and Y onto these sets via the obvious maps, we may assume that
X ∩ Y = ∅. Now consider the set Z = X ∪ Y . Define the relation < on Z
as follows: 
 z, z 0 ∈ X and z <1 z 0
0
z < z ⇐⇒ z, z 0 ∈ Y and z <2 z 0

z ∈ X and z 0 ∈ Y
This relation well-orders Z. We will say that the ordered set is the disjoint
union of the ordered sets (X, <1 ) and (Y, <2 ) (in this order!) We will also
say for a while that Z is the sum of X and Y and we will write Z = X +Y .
121
122 CHAPTER 15. WELL-ORDERED SETS
iv. Lexicographic Ordering. Let (X, <1 ) and (Y, <2 ) be two well-ordered
sets. Consider the following relation on the set X × Y
½
0 0 y <2 y 0
(x, y) < (x , y ) ⇐⇒
y = y 0 and x <1 x0
This defines a well-ordering called lexicographic ordering on the Carte-
sian product X × Y .
Note that the lexicographic ordering on X × 2 is exactly the order defined
in the previous item when X = Y , i.e. X × 2 = X + X.
A nonempty well-ordered set X has a minimal element, say x0 . If X 6= {x0 },
then X \ {x0 } has also a minimal element, say x1 . Thus x1 is the element ”right
after” x0 . In other words x0 < x1 and if x0 < x then x1 ≤ x.
Let X be a well-ordered set, and let x ∈ X. Suppose x is the last element of
X, i.e. suppose that {y ∈ X : x < y} is not empty. Then this set has a (unique)
least element. We will call this element the successor of x and we will denote
it S(x).
An initial segment of a well-ordered set X is a subset I such that for all
a ∈ I and x ∈ X, if x < a then x ∈ I.
Exercises. Throughout X and Y stand for well-ordered sets.

i. Show that Y has a last element if and only if X + Y has a last element.
ii. Show that X and Y have a last element if and only if X × Y has a last
element.
iii. Show that 2 × X is not X + X always.
iv. Let I be an initial segment of X. Show that either I = X or I = (−∞, a)
for some a ∈ X. Conclude that for any two initial segments of X, one is
a subset (hence an initial segment) of the other.
v. Find all initial segments of N.
vi. Find all initial segments of N + N.
vii. Let = be a set of initial segments of X. Show that ∪I∈= I is an initial
segment of X.
15.2 Transfinite Induction

The following theorem, whose proof is trivial (and shorter than its statement),
is the essence of well-ordered sets:
Theorem 15.2.1 Let X be a well-ordered set. Let A be a subset of X. Assume
that for any x ∈ X, if (−∞, x) ⊆ A then x ∈ A. Then A = X.
Proof: Assume X 6= A. Let x be the least element of X\A. Then (−∞, x) ⊆ A.
So, by hypothesis, x ∈ A, a contradiction. ¤
15.3. MORPHISMS 123
15.3 Morphisms
A morphism from a well-ordered structure (X, <1 ) into a well-ordered struc-
ture (Y, <2 ) is a map f : X −→ Y such that for all x, x0 ∈ X, x <1 x0 if and
only if f (x) <2 f (x0 ). Note that any morphism from a well-ordered set into
another must be a one to one map, because if f (x) = f (x0 ) we can neither have
x < x0 nor x0 < x. A morphism which is also onto is called an isomorphism
of well-ordered sets. An isomorphism from a well-ordered set into itself is called
an automorphism, but as we will soon see in this subsection a well-ordered
set has a unique automorphism, only the identity. It follows that there is at
most one isomorphism from a well-ordered set onto another, because if φ and ψ
are two such isomorphisms then ψ −1 ◦ φ is an automorphism, therefore by the
result that we have stated ψ −1 ◦ φ = Id, i.e. φ = ψ.
If two well-ordered sets X and Y are isomorphic, then we write X ' Y .
This is an equivalence relation on the class of all well-ordered sets.
Theorem 15.3.1 A well-ordered set has a unique automorphism, namely the

identity.
Proof: Let (X, <) be a well-ordered structure. Let φ : X −→ X be an au-

tomorphism. Let A = {a ∈ X : φ(a) = a}. We will use Theorem 15.2.1. Let
x ∈ X and assume that (−∞, x) ⊆ A. We proceed to show that x ∈ A. This
will prove the theorem. By assumption, if a ∈ X and a < x, then φ(a) = a. It
follows that (since φ is one to one), φ(x) ≥ x. Assume φ(x) > x. Let y ∈ X be
such that φ(y) = x. Then, since φ(y) = x < φ(x), y < x. Hence x = φ(y) = y,
a contradiction. ¤
Now we show that given any two well-ordered sets, there is an embedding
of one into another. We will show more, we will show that given any two well-
ordered sets, there is a unique isomorphism of one of the two onto an initial
segment of the other. For example, the set 2N of even natural numbers, as a
subset of N, is a well-ordered set and the map 2n 7→ n is an embedding of 2N
onto an initial segment of N (onto N in fact).
We first need the following result which is a generalization of Theorem 15.3.1,
whose proof is in fact very similar.
Lemma 15.3.2 Let X be a well-ordered set and A an initial segment of X.

Then the only embedding of A onto an initial segment of X is the identity map
on A.
Proof: Let B be an initial segment of X. It is easy to show that either A ⊆ B

or B ⊆ A. Assume that there is an embedding f of A onto B. We will show
that A = B and that f = IdA . If B ⊆ A, we may switch the roles of A and B
and replace f by f −1 if necessary to assume that A ⊆ B = X. Thus we assume
that there is an isomorphism from A onto X. We proceed to show that A = X.
Let A1 = {a ∈ A.f (a) = a}. If A1 = A then we are done. Otherwise, let x be
the least element of A \ A1 . If f (x) < x, then f (x) ∈ A1 , so f (f (x)) = f (x) and
124 CHAPTER 15. WELL-ORDERED SETS
f (x) = x, a contradiction. Thus f (x) > x. Let y ∈ A be such that f (y) = x.

Then f (y) = x < f (x) and so y < x, implying y ∈ A1 . But then x = f (y) = y,
a contradiction again. ¤
Theorem 15.3.3 Given any two well-ordered sets, there is a unique embedding
of one into an initial segment of the other.
Proof: Uniqueness follows from Lemma 15.3.2. We will show the existence.
Let (X, <) and (Y, <) be two well-ordered sets. Consider the set = of all
initial segments of X that embed into an initial segment of Y . If A ∈ =, let fA
be the unique embedding of A into Y . Note that if A, B ∈ =, then one of A
or B is a subset of the other and if A ⊆ B, then fB is an extension of fA . By
Example vii, page 38, f := ∪A∈= fA is a function from Z := ∪A∈= A into Y . It is
easy to show that Z is an initial segment of X and that f is an embedding of Z
onto an initial segment of Y . Thus Z is a maximal element of =. If Z = X then
there is nothing more to do. If f (Z) = Y then f −1 is an embedding of Y onto
the initial segment Z of X. Assume Z 6= X and f (Z) 6= Y . Let x and y be the
least elements of X \ Z and Y \ f (Z) respectively. We can extend f to Z ∪ {x}
by sending x to y, we then obtain an embedding of the initial segment Z ∪ {x}
onto the initial segment f (Z) ∪ {y}. Thus Z ∪ {x} ∈ =. But this contradicts
the fact that Z is a maximal element of =. ¤
Exercises.
i. Show that n + N ' N for all n ∈ N.
ii. Show that n × N ' N for all n ∈ N.

iii. Show that N × n 6' N for all 0 6= n ∈ N.
iv. Let n, m ∈ N. Show that N + n ' N + m if and only if n = m.
v. Let n, m ∈ N. Show that N × n ' N × m if and only if n = m.
vi. Show that for n, m ∈ N, the meaning of n + m as a well-ordered set
coincides with the old notion n + m that we have defined.
vii. Show that N × N 6' N.
viii. Show that A + (B + C) ' (A + B) + C for all well-ordered sets A, B, C.
ix. Show that A × (B × C) ' (A × B) × C for all well-ordered sets A, B, C.
Chapter 16
Ordinals
16.1 Definition
An ordinal is a well-ordered set X such that for all x ∈ X, (−∞, x) = x. Thus,
if α is an ordinal and x, y ∈ α, y < x if and only if y ∈ x. Therefore an ordinal
is well-ordered by the membership relation ∈. But this condition is not enough
to make a set an ordinal. (See examples below).
Note that in an ordinal α, if x ∈ α, then x = (−∞, x) ⊆ α, i.e. any element
of an ordinal is a subset of the ordinal.
Examples.
i. The set N is an ordinal. When considered as an ordinal, it is customary

to denote N by ω.
ii. An initial segment of an ordinal is an ordinal. Thus each natural number

is an ordinal. On the other hand {0, 1, 3} is not an ordinal, although it is
well ordered by the membership relation.
iii. If α is an ordinal, then α+ := α ∪ {α} is also an ordinal.
Let α 6= ∅ = 0 be an ordinal and x its least element. We then have: x =

(−∞, x) = ∅ = 0. Thus 0 is the least element of any ordinal 6= 0.
Assume α 6= {0} = 1. Let x be the least element of α \ {0}. Then x =
(−∞, x) = {0} = 1. Thus any ordinal that has at least two elements contain 0
and 1 as elements.
Theorem 16.1.1 If α β are ordinals then either α < β or α = β or β < α.
Proof: Assume not. Consider α ∩ β. This is an initial segment of both α and

β. Let a and b be the least elements of α \ (α ∩ β) and β \ (α ∩ β) respectively.
Then a = (−inf ty, a) = α ∩ β and similarly b = (−inf ty, b) = α ∩ β. Thus
a = b ∈ α ∩ β, a contradiction. ¤
125
126 CHAPTER 16. ORDINALS
Corollary 16.1.2 A subset of an ordinal is an ordinal if and only if it is an

initial segment of that ordinal.
Corollary 16.1.3 A proper subset of an ordinal is an ordinal if and only if it

is an element of that ordinal.
Corollary 16.1.4 Two isomorphic ordinals are equal.
Proof: Let α and β be two isomorphic ordinals. By Theorem 16.1.1, we may

assume that α < β. Let f : α −→ β be the isomorphism. By Theorem 15.3.3,
f = Idα . So α = β. ¤
Exercises.
i. Show that every element of an ordinal is an ordinal.
ii. Let α be an ordinal and β ∈ α. Show that either β + ∈ α or β + = α.
iii. Show that the union of a set of ordinals is an ordinal.
iv. An ordinal that has no last element is called a limit ordinal. Find the
set of limit ordinals of ω, ω + ω and ω × ω.
16.2 Axiom of Replacement

We will prove in the next subsection that any well-ordered set is isomorphic to
an ordinal. For this we need the following axiom (which is considered to be
dangerous by some prominent set theorists).
Axiom of Replacement. If X is set and φ(x, y) is a formula such that for
all x ∈ X there is a unique set y such that φ(x, y), then there is a set Y whose
elements are exactly those sets y such that φ(x, y) for some x ∈ X.
It is aesthetically best not to use the Axiom of Replacement if one can do
otherwise.
16.3 Classification of Well-Ordered Sets

Theorem 16.3.1 Every well-ordered set is isomorphic to a unique ordinal.
Proof: Uniqueness follows from Corollary 16.1.4. We will show the existence
of an isomorphism. Let X be a well-ordered set. Consider the set = of initial
segments of X which are isomorphic to an ordinal. Thus for each A ∈ =, there
is an ordinal αA and an isomorphism fA : A −→ αA . By Theorem 16.1.4, the
ordinal α is unique. By Theorem 15.3.3, the isomorphism fA is unique.
Let A, B ∈ =. Either A ⊆ B or B ⊆ A. Assume A ⊆ B. Also either
αA ⊆ αB or αB ⊆ αA . Assume αB ⊆ αA . Then fA−1 ◦ fB is an isomorphism
16.4. ADDITION OF ORDINALS 127
from B into an initial segment A. Thus A = B. Therefore we may assume that

αA ⊆ αB .
Suppose for a moment that {αA : A ∈ =} is a set. Then ∪A∈= αA is an
ordinal and we may consider the map f := ∪A∈= fA from the initial segment
Y := ∪A∈= A of X onto the ordinal α := ∪A∈= αA . It is easy to check that this
is an isomorphism. Thus Y ∈ = and is the unique maximal element of =. If
Y = X then we are done. Otherwise, consider the least element y of X \ Y and
extend f to an isomorphism from Y ∪ {y} onto the ordinal α+ , contradicting
the maximality of Y .
Thus, it remains to show that {αA : A ∈ =} is a set. This is a direct
consequence of the Axiom of Replacement. ¤
16.4 Addition of Ordinals

If α and β are ordinals, then the well-ordered set α + β is a well-ordered set,
and by Theorem 16.3.1 α + β is isomorphic to a unique ordinal. Changing our
previous notation, we will denote this unique ordinal by α + β.
16.5 Multiplication of Ordinals

If α and β are ordinals, then the well-ordered set α × β is a well-ordered set,
and by Theorem 16.3.1 α × β is isomorphic to a unique ordinal. Changing our
previous notation, we will denote this unique ordinal by α × β.
Paradox of Burali-Forti
128 CHAPTER 16. ORDINALS
Chapter 17
Cardinals
A cardinal is an ordinal that is not in bijection with one of its elements. For
example ω and each natural number is a cardinal.
17.1 Addition of Cardinals

17.2 Multiplication of Cardinals
17.3 Problems
i. Let X be a set. A set Σ of subsets of X is called a σ-algebra if Σ is closed
under finite or countable unions and intersections and complementation.
Show that a σ-algebra Σ is either finite or countable. (Solution: Assume
not. For x ∈ ∪Σ, let A(x) = ∩{U ∈ Σ : x ∈ U }. Then A(x) ∩ A(y) 6= ∅
if and only if A(x) = A(y). Every element of Σ is a countable union of
these disjoint A(x)’s).
17.4 Final Exam of Math 112, May 2003

You should do at least one of Ic, IId and III.
I. Automorphisms of (Q, <).
Ia) Let a, b ∈ Q, a > 0. Let fa,b : Q −→ Q be defined by fa,b (x) = ax + b.
Show that fa,b is an order preserving bijection.
Ib) Let a, b ∈ Q with a < b and k ∈ Z. Show that there is an order
preserving bijection between the rational intervals [k, k + 1] ∩ Q and [a, b] ∩ Q.
Ic) Show that there are uncountably many order preserving bijections of Q.
II. Automorphisms of (R, +, ×).
IIa) Show that any additive map f from Z into R is given by f (x) = rx for
some r ∈ R.
129
130 CHAPTER 17. CARDINALS
IIb) Show that any additive map f from Q into R is given by f (x) = rx for
some r ∈ R.
IIc) Show that the only additive and multiplicative map f from Q into Q is
IdQ .
IId) Show that the only additive and multiplicative map f from R into R
is IdR .
III. Uniqueness of the Real Number System. For i = 1, 2, let (Ri , +i , ×i , 0i , 1i )
be two structures satisfying the axioms of real numbers (ordered complete field).
Show that there is a unique bijection f : R1 −→ R2 such that
i. f (x +1 y) = f (x) +2 f (y)
ii. f (x ×1 y) = f (x) ×2 f (y).
Chapter 18
Axiom of Choice and Zorn’s

Lemma
Axiom of Choice. Any set that does not contain ∅ has a choice function.
18.1 Zorn’s Lemma

Theorem 18.1.1 Any nonempty inductive set has a maximal element.
Proof: Let A be a nonempty inductive set. Let X be the set of chains of A

ordered with inclusion. It is easy to show that a maximal element of X gives
rise to a maximal element of A.
We note the following:
i. X is a set of subsets of A.
ii. Any subset of an element of X is also in X.
iii. The union of a chain of elements from X is also in X.
Let f be a choice function for =(A) \ {∅}.
For x ∈ X, let x̂ = {a ∈ A : x ∪ {a} ∈ X}. Clearly x ⊆ x̂.
Define g : X −→ X by g(x) = x if x = x̂ and g(x) = x ∪ {f (x̂ \ x)}.
We will show that g has a fixed point.
We will say that a subset T of X is a tower, if
T1. ∅ ∈ T .
T2. If x ∈ T then g(x) ∈ T .
T3. If S ⊆ T and if S is a chain, then ∪S ∈ T .
Clearly X is a tower and the intersection of towers is a tower. Thus there is
a smallest tower, say T◦ .
We will show that T◦ is a chain. From this it follows easily that g(x◦ ) = x◦
if x◦ = ∪T◦ .
So let us show that T◦ is a chain.
We say that x ∈ T◦ is comparable if x is comparable with any element
of T . Clearly ∅ is comparable. Clearly the set of comparable elements of T◦
131
132 CHAPTER 18. AXIOM OF CHOICE AND ZORN’S LEMMA
form a chain. We will show that the set of comparable elements form a tower.
We already have noticed T1. T3 is easy as well. Let us show T2. Let x
be comparable. If y ⊆ x, since x and g(y) are comparable, g(y) ⊆ x. Let
U := {y ∈ T◦ : either y ⊆ x or g(x) ⊆ y. Then U is a tower. So U = T◦ . It
follows that g(x) is also is comparable. This proves the theorem. ¤
18.2 Some Consequences of Zorn’s Lemma

18.2.1 Finite and Infinite Sets
18.2.2 König’s Lemma
A tree is a poset T such that for every t ∈ T , {s : s < t} is a well-ordered
set. An element of a tree is called a node or a vertex. Thus in a tree there
are least elements and every vertex of a tree is either an end point or it has at
least one immediate successor. A branch is a linearly ordered subset of a tree.
A tree is called finitely branching if every node has finitely many immediate
successors.
Lemma 18.2.1 (König’s Lemma) Every infinite finitely branching tree with
a unique least element has an infinite branch.
Proof: Let T be such a tree. For n ∈ ω, consider the set T (n) of elements t of
T such that {s ∈ T : s < t} has cardinality n − 1. Since T is finitely branching,
T (n) is finite for all n. An initial subtree of a tree T is subset S of T such
that for s ∈ S, {t ∈ T : t < s} ⊆ S. Consider the set Z of all infinite initial
subtrees of T . Since T ∈ Z, Z 6= ∅. Order Z by reversing the inclusion order,
i.e. for S1 , S2 ∈ Z, define S1 < S2 if S2 ⊂ S1 . We claim that Z is an inductive
set. More precisely, we claim that if (Si )i∈I is a chain from Z, then ∩i∈I Si ∈ Z.
We first show that ∩i∈I Si is an initial subtree. Let s ∈ ∩i∈I Si and let t ∈ T
be such that t < s. Then t ∈ Si for all i ∈ I. Hence t ∈ ∩i∈I Si .
We next show that ∩i∈I Si is infinite. Assume otherwise. Then ∩i∈I Si ⊆
T (n) for some n ∈ ω. For each t ∈ T (n + 1), since t 6∈ ∩i∈I Si there is an i(t) ∈ I
such that t 6∈ Si(t) . Let i = max{i(t) : t ∈ T (n + 1)}. Then T (n + 1) ∩ S(i) = ∅.
But then S(i) ⊆ T (n) and S(i) is finite, a contradiction.
By Zorn’s Lemma Z has a maximal element, say S. Thus S is a minimal
infinite initial tree. We claim that S is a branch. Assume otherwise. Then S
has two incomparable elements t1 and t2 . Deleting one of the ti and elements
above it, we get an infinite proper initial subtree of S, a contradiction. ¤
18.2.3 Some Unexpected Consequences of Zorn’s Lemma

Chapter 19
Axioms of Set Theory –

ZFC
133
134 CHAPTER 19. AXIOMS OF SET THEORY – ZFC
Chapter 20
V =L
135
136 CHAPTER 20. V = L
Chapter 21
Continuum Hypothesis
137
138 CHAPTER 21. CONTINUUM HYPOTHESIS
Chapter 22
Banach-Tarski Paradox
Let S 2 be the unit ball in R3 . By partitioning S 2 and applying translations and

rotations, we will duplicate S 2 .
Let G be a group acting on a set X. A subset E of X is called G-
paradoxical if there are pairwise disjoint subsets A1 , . . . , An , B1 , . . . , Bm of
E and g1 , . . . , gn , h1 , . . . , hm ∈ G such that E = ∪i gi (Ai ) = ∪i hj (Bj ).
Proposition 22.0.2 S 1 is countably SO2 (R)-paradoxical.
Proof: Let G be the subgroup formed by the rational multiples of 2π. Let
G = {ρi : i ∈ N}. Let H be a set of representatives of SO2 (R)/G. Let
A = H(1, 0) and Ai = ρi (A). The family (Ai )i is a countable partition of S 1 .
Then (A2i )i and (A2i+1 ) are such that S 1 = ∪i g2i (A2i ) = ∪i g2i+1 (A2i+1 ) where
g2i takes A2i onto Ai and g2i+1 takes A2i+1 onto Ai . ¤
Let G act on a set X. Let A, B ⊆ X. We say that A and B are G-
equidecomposable if there are finite partitions of A = tni=1 Ai and B =
tni=1 Bn and g1 , . . . , gn ∈ G such that gi (Ai ) = Bi . This is an equivalence
relation.
Lemma 22.0.3 If A ∼G B and if A is G-paradoxical, then B is G-paradoxical.
Proof: Easy. ¤
Lemma 22.0.4 S 1 is SO2 (R)-equidecomposable with S 1 \ {one point}.
Proof: Let A = {ein : n ∈ N \ {0}}. We have S 1 \ {1} = (S 1 \ (A ∪ {1})) t A ∼

(S 1 \ A) t (A ∪ {1}) = S 1 . ¤
A group G is called paradoxical if G acting on itself by left multiplication
is G-paradoxical.
Lemma 22.0.5 The free group F on two generators is paradoxical.
139
140 CHAPTER 22. BANACH-TARSKI PARADOX
Proof: Let a be b be the generators of F . Let
B(σ) = {words starting with σ},
where σ = a, b, a−1 , b−1 . So
F = {1} t B(a) t B(b) t B(a−1 ) t B(b−1 ).
But also F = a−1 B(a) = a−1 B(a) u B(a) and F = b−1 B(b) = b−1 B(b) u B(b).
¤
Lemma 22.0.6 Let G be a paradoxical group acting freely on X. Then X is

G-paradoxical.
Proof: Let A1 , . . . , An , B1 , . . . , Bm ⊆ G be disjoint and g1 , . . . , gn , h1 , . . . , hm ∈

G such that G = ∪i gi Ai = ∪i hj Bj . Let M be a choice set in the set of G-orbits.
Now (gM )g∈G partitions X because the action is free. Let A0i = ∪g∈Ai g(M )
and Bj0 = ∪g∈Bj g(M ). These are all pairwise disjoint. We have X = ∪i gi (A0i ) =
∪j hj (Bj0 ) because G = ∪i gi Ai = ∪i hj Bj . ¤
Corollary 22.0.7 If a group G contains a paradoxical subgroup H then G is

paradoxical.
Proof: Note that the action of H on G is free. So G is H-paradoxical by the

previous lemma. So G is G-paradoxical. ¤
Corollary 22.0.8 Any group containing a free subgroup of rank 2 is paradoxi-

cal.
Fact 22.0.9 SO3 (R) contains a free subgroup of rank 2. Thus SO3 (R) is para-
doxical.
Let G ≤ SO3 (R) be free on two generators. Let D be the set of points on
S 2 which are fixed by some g ∈ G \ {g}. Then D is countable. Let G act
on S 2 \ D. Thus S 2 \ D is paradoxical. We will show that S 2 \ D is SO3 (R)-
equidecomposable with S 2 . Choose an axis of rotation not passing through an
element of D. Consider the set of all rotations around this axis for which some
integer power of it sends an element of D to another point of D. This set is
countable. Pick a rotation ρ not in this set. Then ρn (D) ∩ ρm (D) = ∅ for
n 6= m. Let A = ∪∞ n 2 2
n=1 ρ (D) and B = S \ (A ∪ D). We have S \ D = A t B ∼
−1 2
B ∪ ρ (A) = S .
Thus S 2 is SO3 (R)-paradoxical.
Chapter 23
First Order Structures
141
142 CHAPTER 23. FIRST ORDER STRUCTURES
Chapter 24
Ultraproducts and
Ultrafilters
24.1 Nonstandard Models of N

24.2 Nonstandard Models of R
143
144 CHAPTER 24. ULTRAPRODUCTS AND ULTRAFILTERS
Chapter 25
Dimension Theory
145
146 CHAPTER 25. DIMENSION THEORY
Chapter 26
Exams
26.1 First Semester Midterm, November 2002

i. a) Given a set X, define ∪X as follows:
y ∈ ∪X if and only if there is an x ∈ X such that y ∈ x.
Show that ∪∅ = ∅. (3 pts.)

Proof: Set X = ∅ in the definition. Thus
y ∈ ∪∅ if and only if there is an x ∈ ∅ such that y ∈ x.
Since there is no x ∈ ∅, we see that ∪∅ = ∅.
b) Given a set X, define ∩1 X and ∩2 X as follows:
y ∈ ∩1 X if and only if y ∈ x for all x ∈ X

y ∈ ∩2 X if and only if y ∈ ∪X and y ∈ x for all x ∈ X
Is ∩1 X = ∩2 X for all X? Which definition do you prefer for ∩X and

why? (5 pts.)
Answer: Certainly ∩2 X ⊆ ∩1 X because ∩2 X = (∩1 X) ∩ (∪X) by defi-

nition.
The reverse inclusion holds only if X 6= ∅. Indeed, assume X 6= ∅ and let
y ∈ ∩1 X. To show that y ∈ ∩2 X, we need to show that y ∈ ∪X, i.e. that
y ∈ x for some x ∈ X. Since X 6= ∅ and y ∈ ∩1 X, this holds trivially, i.e.
y ∈ x for some x ∈ X.
The reverse inclusion does not hold if X = ∅. Indeed, ∩2 ∅ = (∩1 ∅)∩(∪∅) =
(∩1 ∅) ∩ ∅ = ∅. On the other hand,
y ∈ ∩1 ∅ if and only if y ∈ x for all x ∈ ∅
147
148 CHAPTER 26. EXAMS
and this condition holds for all y. Hence ∩1 ∅ is the whole universe and is
not even a set.
Thus we should prefer ∩2 X for the definition of ∩X since the outcome
would then be a set for any set X, even for X = ∅!
ii. Find a set X such that X ∩ ℘(X) 6= ∅. (3 pts.)

Answer. Since ∅ ∈ ℘(X) for any X, any set X that contains ∅ as element
would do, e.g. we may take X = {∅}.
iii. Let (Ai )i∈I be a family of sets.

T T
a) Show that i∈I ℘(Ai ) = ℘( i∈I Ai ). (3 pts.)
Proof:
T
X∈ i∈I ℘(Ai ) if and only if X ∈ ℘(Ai ) for all i ∈ I
if and only if X ⊆ATi for all i ∈ I
if and only if X ⊆ i∈I
T Ai
if and only if X ∈ ℘( i∈I Ai )
S S
b) What is the relationship between i∈I ℘(Ai ) and ℘( i∈I Ai )? (4 pts.)
S S
Answer: We have i∈I ℘(Ai ) ⊆ ℘( i∈I Ai ). Indeed,
S
X∈ i∈I ℘(Ai ) if and only if X ∈ ℘(Ai ) for some i ∈ I
if and only if X ⊆ Ai for some i ∈ I
implies X ⊆ ∪i∈I Ai
if and only if X ∈ ℘(∪i∈I Ai )
The reverse inclusion is false, in fact if oneSof the sets Ai does not contain
all the others, then ∪i∈I Ai ∈ ℘(∪i∈I Ai ) \ i∈I ℘(Ai ).
iv. What is the set (X × Y ) ∩ (Y × X)? (3 pts.)

Answer: Let (u, v) ∈ (X × Y ) ∩ (Y × X). Since (u, v) ∈ X × Y , u is an
element of X and v is an element of Y . Similarly, since (u, v) ∈ Y × X,
u is an element of Y and v is an element of X. Thus both u and v are
elements of X ∩ Y , i.e. (u, v) ∈ (X ∩ Y ) × (X ∩ Y ).
The reverse inclusion (X ∩ Y ) × (X ∩ Y ) ⊆ (X × Y ) ∩ (Y × X) is trivial.
v. Let α be a set such that x ⊆ α for all x ∈ α. Show that α ∪ {α} has the
same property. Give four examples of such sets. (4 pts.)
Proof: Let x ∈ α ∪ {α}. Then either x ∈ α or x ∈ {α}.
If x ∈ α, then by assumption x ⊆ α. Since α ⊆ α ∪ {α}, in this case we
get x ⊆ α ∪ {α}.
If x ∈ {α}, then x = α, and so x = α ⊆ α ∪ {α}.
26.1. FIRST SEMESTER MIDTERM, NOVEMBER 2002 149
Since ∅ has the property stated, starting from ∅, we can get as many
examples as we wish to, here are the first four:
∅=0
0 ∪ {0} = 1
1 ∪ {1} = 2
2 ∪ {2} = 3
vi. Let Γ be a graph such that for any vertices α, α1 , β, β1 , if α 6= α1 and

β 6= β1 , then there is a φ ∈ Aut(Γ) such that φ(α) = β and φ(α1 ) = β1 .
What can you say about Γ? (5 pts.)
Answer: Then either the graph is the complete graph (all possible edges
exist) or the graph without edges at all. Indeed, otherwise we may find
α, α1 and β 6= β1 such that α and α1 are connected (hence α 6= α1 ) and β
and β1 are not connected, but then it is impossible to send the connected
pair (α, α1 ) to the nonconnected pair (β, β1 ).
vii. Let φ : R −→ R be a one to one map such that φ(x + y) = φ(x) + φ(y)
and φ(x2 ) = φ(x)2 for all x, y ∈ R. Show that φ(xy) = φ(x)φ(y) for all
x, y ∈ R and φ(q) = q for all q ∈ Q. (10 pts.)
Proof: For any x, y ∈ R, we have φ(x)2 + 2φ(x)φ(y) + φ(y)2 = (φ(x) +
φ(y))2 = φ(x + y)2 = φ((x + y)2 ) = φ(x2 + 2xy + y 2 ) = φ(x2 ) + 2φ(xy) +
φ(y 2 ) = φ(x)2 +2φ(xy)+φ(y)2 and so 2φ(x)φ(y) = 2φ(xy) and simplifying,
we get φ(x)φ(y) = φ(xy). This proves the first part.
Since φ(0) = φ(0 + 0) = φ(0) + φ(0), we must have φ(0) = 0.
Since φ(1) = φ(1 · 1) = φ(1)φ(1), we must have φ(1) = 0 or φ(1) = 1. But
the first case is forbidden because φ is one to one and φ(0) = 0 already.
Hence φ(1) = 1.
Now, it follows easily by induction that φ(n) = n for all n ∈ N because
φ(n+1) = φ(n)+φ(1) = φ(n)+1 = n+1 (the last equality is the inductive
hypothesis).
Also, for n ∈ N, we have 0 = φ(0) = φ(n + (−n)) = φ(n) + φ(−n) and so
φ(−n) = −φ(n) = −n. Thus φ(n) = n for all n ∈ Z.
Now if q ∈ Q, then q = n/m some n, m ∈ Z and m 6= 0. Then we have
n = φ(n) = φ(mn/m) = φ(m)φ(n/m) = mφ(n/m) and so φ(n/m) =
n/m, i.e. φ(q) = q.
viii. Given a set X, define ℘n (X) as follows by induction on n: ℘0 (X) = X

and ℘n+1 (X) = ℘(℘n (X)).
a) Is there a natural number n such that for any set X, {{∅}, {{X}}} ∈
℘n (X)? (8 pts.)
Answer: For n ≥ 4 note the equivalence of the following propositions:
{{∅}, {{X}}} ∈ ℘n (X)

{{∅}, {{X}}} ⊆ ℘n−1 (X)
{∅}, {{X}} ∈ ℘n−1 (X)
{∅}, {{X}} ⊆ ℘n−2 (X)
∅, {X} ∈ ℘n−2 (X)
∅ ∈ ℘n−2 (X) and {X} ⊆ ℘n−3 (X)
{X} ⊆ ℘n−3 (X)
{X} ⊆ ℘n−3 (X)
X ∈ ℘n−3 (X)
X ⊆ ℘n−4 (X)
If n = 4, the last condition holds for all X.

Does it hold for n = 5, i.e. do we have X ⊆ ℘(X) for all X? For this
condition to hold, we need any element of X to be a subset of X, and
this does not always hold. For any i ≥ 1, one can find a set X such that
X 6⊆ ℘i (X). Thus the condition does not hold for any n ≥ 5 (details are
left as an exercise).
For n = 0, 1, 2, 3, find examples of X such that {{∅}, {{X}}} 6∈ ℘n (X).
b) Show that ℘(℘n (X)) = ℘n (℘(X)) for all sets X and all natural numbers
n. (8 pts.)
Proof: We proceed by induction on n. The condition certainly holds for
n = 0. Assume it holds for n. We have ℘(℘n+1 (X)) = ℘(℘n (℘(X))) =
℘n+1 (℘(X)).
c) Show that ℘n (℘m (X)) = ℘m (℘n (X)) for all sets X and all natural
numbers n and m. (8 pts.)
Proof: We proceed by induction on m. The condition certainly holds for
m = 0. By part (b) it also holds for m = 1. Assume it holds for m. We
have ℘n (℘m+1 (X)) = ℘n (℘(℘m (X))) = ℘(℘n (℘(X)) = ℘n+1 (℘m (X)).
ix. Define a partial order ≺ on N\{0, 1} by x ≺ y if and only if x2 |y. Describe

all the automorphisms of this poset. (5 pts.)
Answer: The minimal elements of this ordered set (call it Γ) are the
square free numbers. Thus any automorphism of Γ should send the square
free numbers onto the square free numbers. But the prime numbers have
a privilege. Indeed if p is a prime number, then p has an immediate
successor (namely p2 ) that has only one predecessor, namely p. Thus any
automorphism should be multiplicative and be given by a permutation of
primes.
x. Let X be a set. Let Γ be the set of subsets of X with two elements. On Γ

define the relation αRβ if and only if α ∩ β = ∅. Then Γ becomes a graph
with this relation.
26.1. FIRST SEMESTER MIDTERM, NOVEMBER 2002 151
a) Calculate Aut(Γ) when |X| = 4. (3 pts.)

Answer: The graph Γ is just six vertices joined two by two. A group
isomorphic to (Z/2Z)3 preserves the edges. And Sym(3) permutes the
edges. Thus the group has 8 × 3! = 48 elements.
More formally, one can prove this as follows. Let the vertices be 1, 2, 3,
4, 5, 6 and the edges be v1 = (1, 4), v2 = (2, 5) and v3 = (3, 6). We can
embed Sym(3) in Aut(Γ) ≤ Sym(6) via
Id3 7→ Id6
(12) 7→ (12)(45)
(13) 7 → (13)(46)
(23) 7 → (23)(56)
(123) 7 → (123)(456)
(132) 7 → (132)(465)
For any φ ∈ Aut(Γ) there is an element α in the image of Sym(3) such

that α−1 φ preserves the three edges v1 = (1, 4), v2 = (2, 5) and v3 = (3, 6).
Thus α−1 φ ∈ Sym{1, 4} × Sym{2, 5} × Sym{3, 6} ' (Z/2Z)3 . It follows
that Aut(Γ) ' (Z/2Z)3 o Sym(3) (to be explained next year).
b) Draw the graph Γ when X = {1, 2, 3, 4, 5}. (3 pts.)
There are ten points. Draw two pentagons one inside the other. Label the
outside points as {1, 2}, {3, 4}, {5, 1}, {2, 3}, {4, 5}. Complete the graph.
c) Show that Sym(5) imbeds in Aut(Γ) naturally. (You have to show that
each element σ of Sym(5) gives rise to an automorphism σ̃ of Γ in such
a way that the map σ 7→ σ̃ is an injection from Sym(5) into Aut(Γ) and
that σ^
1 ◦ σ2 = σ
f1 ◦ σ
f2 ). (8 pts.)
d) Show that Aut(Γ) ' Sym(5). (12 pts.)
Proof of (c) and (d): Clearly any element of σ ∈ Sym(5) gives rise to an
automorphism σ̃ of Γ via σ̃{a, b} = {σ(a), σ(b)}. The fact that this map
preserves the incidence relation is clear. This map is one to one because
if σ̃ = τ̃ , then for all distinct a, b, c, we have
{σ(b)} = {σ(a), σ(b)} ∩ {σ(b), σ(c)} = σ̃{a, b} ∩ σ̃{b, c}
= τ̃ {a, b} ∩ τ̃ {b, c} = {τ (a), τ (b)} ∩ {τ (b), τ (c)} = {τ (b)},
and hence σ(b) = τ (b).

Let φ ∈ Aut(Γ). We will compose φ by elements of Sym(5) to obtain the
identity map. There is an σ ∈ Sym(5) such that φ{1, 2} = σ̃{1, 2} and
φ{3, 4} = σ̃{3, 4}. Thus, replacing φ by σ −1 φ, we may assume that φ
fixes the vertices {1, 2} and {3, 4}. Now φ must preserve or exchange the
vertices {3, 5} and {4, 5}. By applying the element (34) of Sym(5) we may
assume that these two vertices are fixed as well. Now φ must preserve or
exchange the vertices {1, 3} and {2, 3}. By applying the element (12) of
Sym(5) we may assume that these two vertices are fixed as well. Now all
the vertices must be fixed.
26.2 First Semester Final, January 2003

Axiom of Regularity (AR) states that every nonempty set A has an element
x such that A ∩ x = ∅. Throughout, we will accept the axiom of regularity.
Let X be a set and < be a total order on X. We say that (X, <) is a well-
ordered set (or that < well-orders X) if every nonempty subset of X contains
a minimal element for that order, i.e. if for every nonempty subset A of X, there
is an m ∈ A such that m ≤ a for all a in A. Clearly, given a well-ordered set
A, such an m is unique. In particular every nonempty well-ordered set has a
unique minimal element.
Note that ∅ is a well-ordered set.
Of course, subsets of a well-ordered set X inherit the well-order of X. If
(X, <) is an ordered set and x ∈ X, we define i(x) = {y ∈ X : y < x} (the
initial segment of x). If X is a set, we set X + = X ∪ {X}. An ordinal is a
well-ordered set such that β = i(β) for all β ∈ α.
i. Show that no set x is a member of itself. (3 pts.)
ii. Show that there are no sets x and y such that x ∈ y and y ∈ x. (3 pts.)
iii. Show that X ⊂ X + . (2 pts.)
iv. Assume X is a well-ordered set. Order X + by extending the order of X

and stating that X is larger than its elements (i.e. put the element X to
the very end of X). Show that X + is also a well-ordered set. (3 pts.)
v. (Transfinite Induction) Let (X, <) be a well-ordered set and let A ⊆ X

be such that for all x ∈ X, if i(x) ⊆ A, then x ∈ A. Show that A = X. (5
pts.)
vi. Let X and Y be two well-ordered sets. Let A = (X × {0}) ∪ (Y × {1}).

Order A as follows:
(x1 , 0) < (x2 , 0) for all x1 and x2 in X and x1 < x2 .
(y1 , 1) < (y2 , 1) for all y1 and y2 inY and y1 < y2 .
(x, 0) < (y, 1) for all x ∈ X and y ∈ Y .
Show that the above relation well-orders A. (4 pts.)
vii. Show that ∅ is an ordinal. (2 pts.)
viii. Show that an ordinal α is a set well-ordered by the relation ∈, i.e. by the
relation < defined as follows: For all β, γ ∈ α, β < γ if and only if β ∈ γ.
Show that the converse of this statement is false. (3 pts.)
ix. Show that if α 6= ∅ is an ordinal, then ∅ ∈ α and ∅ is the least element of

α. (7 pts.)
x. Show that if α is an ordinal and β ∈ α, then β ⊂ α. (2 pts.)

26.3. FIRST SEMESTER, RESIT, JANUARY 2004 153
xi. Show that every element of an ordinal is an ordinal. (2 pts.)

xii. Show that if α is an ordinal, then α+ is also an ordinal. (2 pts.)
xiii. Let α be an ordinal and β ∈ α. Show that either β + ∈ α or β + = α. (8
pts.)
xiv. In Question vi, take X = ω and Y = 1 = {0}. Show that the well-
ordered set A obtained there is isomorphic to the ordinal ω + , i.e. there is
an order-preserving bijection from A onto ω + . (4 pts.)
xv. In Exercise vi, take X = 1 = {0} and Y = ω. Show that the well-ordered
set A obtained there is isomorphic to ω, i.e. there is an order-preserving
bijection from A onto ω. (4 pts.)
xvi. Let α β be ordinals. Show that either α < β or α = β or β < α. (20 pts.)
xvii. Show that the union of a set of ordinals is an ordinal. (3 pts.)
xviii. Let α and β be two ordinals. Let f : α −→ β be a strictly increasing
function. Show that if f is onto, then α = β and f is the identity map.
(20 pts.)
Axiom of Replacement says the following: If X is set and φ(x, y) is a

formula such that for all x ∈ X there is a unique set y such that φ(x, y),
then there is a set Y whose elements are exactly those sets y such that
φ(x, y) for some x ∈ X. You need this axiom for the next exercise.
xix. Show that every well-ordered set is isomorphic to an ordinal. (20 pts.)
26.3 First Semester, Resit, January 2004

Important Note. Write either in English or in Turkish, but in any event make
full sentences. Use proper punctuation. Do not use symbols such as ⇔, ⇒, ∃, ∀.
For each use of these symbols I will take away 1 pt. Explain all your answers,
but grades will be taken away for unnecessary text. Any unexplained answer
will get 0 point, whether the answer is correct or not.
I. Transitive Relations. Let X be a set. A binary relation on X is just
a subset of X × X.
A binary relation R on X is called transitive if, for all x, y, z ∈ X, (x, y) ∈
R and (y, z) ∈ R implies (x, z) ∈ R.
i. Which of the following a transitive binary relation on any set X? Explain.

(0 or 4 pts.)
i. X × X.
ii. ∅.
iii. {(x, x) : x ∈ X}.
iv. {(x, y) ∈ X 2 : x 6= y}.

Solution. The first three of them are always transitive. In the first one
any two elements are related, so it is clearly transitive. The second one
is transitive because no two elements are related. The third one is just
the equality relation. Last one is not transitive because if x 6= y are two
distinct elements of X, then x and y are related and y and x are related,
but x and x are not related.
ii. Which of the following a transitive binary relation on N? Explain. (0 or

4 pts.)
i. {(x, y) ∈ N2 : 5 divides x − y}.
ii. {(x, y) ∈ N2 : 5 divides x + y}.
iii. {(x, y) ∈ N2 : 5 > x − y}.
iv. {(x, y) ∈ N2 : 12 < x − y}.
Solution. Only the first one is transitive.
iii. Show that the intersection of a set of transitive relations on X is a tran-

sitive relation on X. (4 pts.)
Solution. Let (Ri )i be a family of transitive relations. Let (x, y), (y, z) ∈
∩i Ri . Then (x, y), (y, z) ∈ Ri for all i. Since each Ri is transitive, this
implies that (x, z) ∈ Ri for all i. Hence (x, z) ∈ ∩i Ri . This shows that
∩i Ri is a transitive relation.
iv. Show that for any binary relation R on X the intersection Rt of all the
transitive relations that contain R is the unique smallest transitive relation
on X that contains R. (10 pts.)
Rt is called the transitive closure of R
Solution. By above, the intersection Rt of all the transitive relations that
contain R is the a transitive relation on X that contains R. Since it is
the intersection of all the transitive relations on X that contains R, it can
only be the smallest one.
v. Show that if R and S are two binary relations then (R ∩ S)t ⊆ Rt ∩ S t .

(8 pts.)
Solution. Since R ⊆ Rt and S ⊆ S t , R ∩ S ⊆ Rt ∩ S t . Since Rt ∩ S t
is transitive by the above question, this implies that (R ∩ S)t ⊆ Rt ∩ S t
(because (R ∩ S)t is the smallest transitive relation that contains R ∩ S).
vi. Let R be a binary relation. Show that the subset
{S := (x, y) ∈ X 2 : ∃x = y1 , y2 , . . . , yn = y ∈ X such that
(yi , yi+1 ) ∈ R for all i = 1, . . . , n − 1} is a transitive relation that

contains R. Conclude that Rt = S. (10 pts.)
Solution. Let (x, y) and (y, z) be in S. By the very definition of S,

there are x = y1 , y2 , . . . , yn = y ∈ X and y = z1 , z2 , . . . , zm = z ∈ X
such that (zj , zj+1 ) ∈ R and (zj , zj+1 ) ∈ R for all i = 1, . . . , n − 1} and
j = 1, . . . , m − 1. Setting ti = yi and tn+j = zj for all i = 1, . . . , n − 1}
and j = 1, . . . , m − 1}, we see that x = t1 , (tk , tk+1 ) ∈ R and tn+m = z.
Thus (x, z) ∈ S. This shows that S is transitive.
vii. Show that in general (R ∩ S)t 6= Rt ∩ S t . (5 pts.)

Solution. Let X = {a, b, c, d}. Let R = {(a, b), (b, d)} and R = {(a, c), (c, d)}.
Then Rt = {(a, b), (b, d), (a, d)} and R = {(a, c), (c, d), (a, d)}. So Rt ∩
S t = {(a, d)} and R ∩ S = ∅.
II. Partial Orders. A binary relation < on a set X is called a partial

order on X if (writing x < y instead of (x, y) ∈ <),
PO1. Irreflexivity. For every x ∈ X, x 6< x.
and
PO2. Transitivity. For every x, y, z ∈ X, if x < y and y < z then

x < z.
We write x ≤ y if either x < y or x = y.

Let (X, <) be a partially ordered set and A ⊆ X. An element u ∈ X is
called an upper bound of A if a ≤ u for all a ∈ A. An element v ∈ X is
called a least upper bound of A if i) v is an upper bound for A and ii)
for any upper bound u of A, if u ≤ v then u = v.
viii. Give an example of a partially ordered set (X, <) and a subset A of X
which
i. has a least upper bound which is not in A.
ii. has exactly two least upper bounds.
iii. does not have a least upper bound.
iv. has a least upper bound which is in A. (4 pts.)
Solution: i. Take X = R with the usual order and let A = (0, 1). Then
lub(A) = 1 6∈ A.
ii. Let X = {a, b, c} and set a < b and a < c. Take also A = X. Then
both b and c are least upper bounds of A.
iii. Let X = R or N with the usual order and take A = N.
iv. Take X = R with the usual order and let A = (0, 1]. Then lub(A) =
1 ∈ A.
ix. Let (X, <) be a partially ordered set and A a subset of X. Suppose that
A has a least upper bound which is in A. Show that this is the only upper
bound of A. (2 pts.)
Proof: Let a ∈ A be a least upper bound of A. Let b be another upper
bound of A. Then, since a ∈ A, a ≤ b. Therefore, a being an upper bound
and b a least upper bound, a = b.
x. Let (X, <) be a partially ordered set. Show that any element of X is an
upper bound of ∅. (2 pts.)
Proof: Certainly any element of X is greater than or equal to any element
of ∅. ¤
xi. Let (X, <) be a partially ordered set. What can you say about (X, <) if
∅ has a least upper bound? (2 pts.)
Solution: Then the partially ordered set X must have a unique least
element.
xii. Let U be a set and let X = ℘(U ). Order X by strict inclusion. Show that
this is a partial order on X. (2 pts.) Show that any subset of X has a
unique least upper bound. (5 pts.)
Proof: Let a, b, c ∈ X be such that a ⊂ b ⊂ c. Then a ⊂ c. Since a 6⊂ a,
this shows that X is a partial order. Let A ⊂ X. Then I claim that ∪A,
i.e. ∪a∈A a is the unique least upper bound of A. Since a ⊆ ∪a∈A a for all
a ∈ A, it is clear that ∪a∈A a is an upper bound of A. Assume b is another
upper bound of A. Then a ⊆ b for all a ∈ A; hence ∪a∈A a ⊆ b. This
shows that ∪a∈A a is the least upper bound of A. ¤
xiii. Let (X, <) be a partial order. Suppose that for any a, b ∈ X, the set
{a, b} has a unique least upper bound. Let a ∨ b denote this least upper
bound.
i. Give an infinite example of such a partially ordered set. (2 pts.)
ii. Prove or disprove: (a ∨ b) ∨ c = a ∨ (b ∨ c) for all a, b, c ∈ X. (10 pts.)
Solution: i. N with the natural order is such an example. Here lub(x, y) =
max(x, y).
ii. Proof: Clearly a ≤ a ∨ (b ∨ c). Also b ≤ b ∨ c ≤ a ∨ (b ∨ c), thus
b ≤ a ∨ (b ∨ c) as well. These show that a ∨ b ≤ a ∨ (b ∨ c).
On the other hand c ≤ b ∨ c ≤ a ∨ (b ∨ c). With the result of the previous
paragraph we get (a ∨ b) ∨ c ≤ a ∨ (b ∨ c).
Similarly one can also show that a ∨ (b ∨ c) ≤ (a ∨ b) ∨ c. Therefore
(a ∨ b) ∨ c = a ∨ (b ∨ c).
III. Total Orders. If in addition to PO1 and PO2 stated above,
O3 For every x, y ∈ X, either x < y or x = y or y < x,

then the partial order is called a total order.
xiv. Show that in a totally ordered set (X, <) if a subset A of X has a least
upper bound then this least upper bound is the only upper bound of A.
(4 pts.)
Proof: Let A ⊆ X have a least upper bound, say b. Let c be another
least upper bound. Then either b ≤ c or c ≤ b, and any of these imply
b = c.
IV. Well-Ordered Sets. We say that a totally ordered set (X, <) is a
well-ordered set (or that < well-orders X) if every nonempty subset of
X contains a minimal element for that order, i.e. if for every nonempty
subset A of X, there is an m ∈ A such that m ≤ a for all a in A. (Note
that the element m must be in A).
xv. Give an example of a finite and an infinite well-ordered set. (2 pts.)

Solution. N is an infinite well ordered set. Any finite subset of N is a
well-ordered set with the restricted (and natural) order.
xvi. Let X = N × {0} ∪ N × {1}. On X define the relation < as follows: For
all x, y ∈ N,
(x, 0) < (y, 0) if and only if x<y

(x, 1) < (y, 1) if and only if x<y
(x, 0) < (y, 1) always
i. Is (X, <) a totally ordered set? (2 pts.)

ii. Is (X, <) a well-ordered set? (2 pts.)
Proof: i. Yes. This is the order obtained by putting a copy of N (namely
N × {1}) to the “very end” of a copy of N (namely N × {0}).
ii. Yes. Let ∅ 6= A ⊆ X.
If A∩(N×{0} 6= ∅, let n = lub{n ∈ N : (n, 0) ∈ A}. Then (n, 0) = lub(A).
If A∩(N×{0} = ∅, let n = lub{n ∈ N : (n, 1) ∈ A}. Then (n, 1) = lub(A).
xvii. Is the set {1/n : n ∈ N\{0}} together with the natural order a well-ordered
set? (2 pts.)
Answer: No, because the set itself does not have a least element.
xviii. Is the set {1/n : n ∈ N \ {0}} ∪ {0} together with the natural order a
well-ordered set? (2 pts.)
Answer: No. Let A = {1/n : n ∈ N \ {0}}. Then A does not have a least
element: glb(A) = 0 6∈ A.
xix. Find an infinite well-ordered set with a maximal element. (4 pts.)

Solution: Let X = N ∪ {a} where a is a new element. Define an order on
X as follows: Take the natural order on N and declare that a is greater
than any element of N. Then this is a well-order and a is its maximal
element.
xx. Show that in a well-ordered set the minimal element of any nonempty
subset is unique. (2 pts.)
Proof: Let A be a nonempty subset of a well-ordered set X. Let a and b
be two least elements of A. Either a < b or b < a or a = b, The first case
contradicts the fact that b is a minimal element of A because a ∈ A. The
second case gives a contradiction in a similar way.
xxi. Show that every nonempty well-ordered set has a unique minimal element.
(2 pts.)
Proof: Follows from above taking A to be the well-ordered set itself.
xxii. Let (X, <) be a well-ordered set. Show that all the elements of X except
possibly one of them satisfies the following property: “There exists a y
such that x < y and for all z if x < z then y < z”. (5 pts.) Show that
such a y, when exists, is unique (3 pts.)
Proof: Let x ∈ X. Assume x is not the last element of X. Then
{y ∈ X : x < y} is a nonempty subset of X and as such has a minimal
element, say y, and by one of the questions above this minimal element y
is unique.
26.4 First Semester, February 2003

Let X be a set. A set of subsets of X - whose elements are called open subsets
of X - is called a topology on X, if
T1. ∅ and X are open subsets of X.
T2. The intersection of two open subsets of X is an open subset of X.
T3. The union of any collection of open subsets of X is an open subset of
X.
More formally, a subset τ of ℘(X) is called a topology on X, if
T1. ∅ ∈ τ and X ∈ τ.
T2. If U, V ∈ τ, then U ∩ V ∈ τ.
T2. If σ ⊆ τ, then ∪ σ ∈ τ.
i. Show that {∅, X} is a topology on X. (2 pts.)
ii. Show that ℘(X) is a topology on X. (2 pts.)
iii. Let A be a subset of X. Show that {∅, A, X} is a topology on X. (2 pts.)

26.4. FIRST SEMESTER, FEBRUARY 2003 159
iv. Let A and B be two subsets of X. Find a finite topology on X that contains
A and B. (3 pts.)
v. Let A, B and C be three subsets of X. Find a finite topology on X that

contains A and B. What is the maximum size of the smallest such topol-
ogy? (4 pts.)
vi. Let τ be a topology on X. Let A ⊆ X. Show that there is a unique largest

open subset of A. We denote this subset by A◦ . (6 pts.)
Show that for all subsets A and B of X,
a) If A ⊆ B then A◦ ⊆ B ◦ . (4 pts.)
b) (A ∩ B)◦ ⊆ A◦ ∩ B ◦ . (5 pts.)
c) (A ∪ B)◦ = A◦ ∪ B ◦ . (7 pts.)
vii. Show that if σ and τ are topologies on X, then σ ∩ τ is also a topology

on X. (5 pts.)
viii. Show that if T is any set of topologies on X, then ∩T is also a topology

on X. (3 pts.)
ix. Show that if σ ⊆ ℘(X) is a set of subsets of X, then the intersection hσi
of all the topologies that contain σ is the smallest topology on X that
contains σ. In other words, if A ∈ σ then A is open in the topology hσi,
and hσi is the smallest topology on X with this property. This topology
is called the topology generated by σ. (10 pts.)
x. Let σ be the set of singleton subsets of X. Find hσi. (3 pts.)
xi. Let n ∈ N. Let σ be the set of subsets of X of cardinality n. Find hσi. (5

pts.)
xii. Assume σ is the set of cofinite subsets of X. Find hσi. (4 pts.)
xiii. Let σ ⊆ ℘(X). Consider the set [σ] of subsets A of X with the following
property:
For any x ∈ A there are finitely many S1 , . . . , Sn ∈ σ such that

x ∈ S1 ∩ . . . ∩ Sn ⊆ U .
a) Show that [σ] is a topology on X. (5 pts.)

b) Show that hσi ⊆ [σ]. (5 pts.)
c) Show that hσi = [σ]. (5 pts.)
Thus U ∈ hσi if and only if for all x ∈ hσi there are finitely many
S1 , . . . , Sn ∈ σ such that x ∈ S1 ∩ . . . ∩ Sn ⊆ U . (15 pts.)
xiv. Let σ be the set of all open intervals of R. Consider the set R with the
topology hσi.
a) Show that a subset A of R is open (in this topology) if and only if for
all a ∈ A there is an ² ∈ R>0 such that (a − ², a + ²) ⊆ A. (7 pts.)
b) Show that no finite and nonempty subset of R is open (in this topology).
(2 pts.)
c) Show that [0, 1) is not open. (3 pts.)
d) Show that a cofinite subset is open. (3 pts.)
d) Is Q open? (3 pts.)
xv. Let σ1 be the set of all intervals of R of the form [a, b) of R for a ≤ b ∈ R.
a. Show that hσi ⊂ hσ1 i. (5 pts.)
b. Can a singleton set be open in this topology? (3 pts.)
c. Show that [0, 1] is not open in this topology. (3 pts.)
xvi. Let σ2 be the set of all closed and bounded intervals of R. Show that
hσ1 i ⊂ hσ2 i where σ1 is as in number xv. (5 pts.)
26.5 First Semester Final and Its Correction,

January 2004
You may assume that you know all the basic arithmetic properties of (Z, +, ×, 0, 1)
and (N, +, ×, 0, 1).
i. Let X = Z × (Z \ {0}). Define the relation ≡ on X by
(x, y) ≡ (z, t) ⇔ xt = yz
for every (x, y), (z, t) ∈ X.

a) Show that this is an equivalence relation on X.
b) Find the equivalence classes of (0, 1) and of (3, 3).
c) Show that if (x, y) ≡ (x0 , y 0 ) and (z, t) ≡ (z 0 , t0 ) then (xt + yz, yt) ≡
(x0 t0 + y 0 z 0 , y 0 t0 ).
d) Show that if (x, y) ≡ (x0 , y 0 ) and (z, t) ≡ (z 0 , t0 ) then (xz, yt) ≡ (x0 z 0 , y 0 t0 ).
Proof: a. i. Reflexivity. Let (x, y) ∈ X. Then since xy = yx, we have
(x, y) ≡ (x, y).
ii. Symmetry. Let (x, y), (z, t) ∈ X be such that (x, y) ≡ (z, t). Hence
xt = yz. Therefore zy = tx, implying (z, t) ≡ (x, y).
iii. Transitivity. Let (x, y), (z, t), (u, v) ∈ X be such that (x, y) ≡ (z, t)
and (z, t) ≡ (u, v). Hence xt = yz and zv = tu. Multiplying these
26.5. FIRST SEMESTER FINAL AND ITS CORRECTION, JANUARY 2004161
equalities side by side, we get xtzv = yztu. Since t 6= 0, by simplifying

we get xzv = yzu. If z 6= 0, then we can simplify further to get xv = yu,
hence (x, y) ≡ (u, v).
Assume z = 0. Then xt = yz = 0 and tu = zv = 0. Since t 6= 0, we get
x = u = 0, so that xv = 0 = yu and (x, y) ≡ (u, v) again.
b. (0, 1) := {(x, y) ∈ X : (x, y) ≡ (0, 1)} = {(x, y) ∈ X : x = 0} =

{(0, y) : y ∈ Z \ {0}}.
(3, 3) := {(x, y) ∈ X : (x, y) ≡ (3, 3)} = {(x, y) ∈ X : 3x = 3y} = {(x, x) :
x ∈ Z \ {0}}.
c) Assume (x, y) ≡ (x0 , y 0 ) and (z, t) ≡ (z 0 , t0 ). Then xy 0 = yx0 and

zt0 = tz 0 . Multiplying the first one by tt0 and the second one by yy 0 we
get xy 0 tt0 = yx0 tt0 and zt0 yy 0 = tz 0 yy 0 . Adding these two side by side we
get xy 0 tt0 + zt0 yy 0 = yx0 tt0 + tz 0 yy 0 , and factoring, we get (xt + yz)y 0 t0 =
yt(x0 t0 + y 0 z 0 ), meaning (xt + yz, yt) ≡ (x0 t0 + y 0 z 0 , y 0 t0 ).
d) Assume (x, y) ≡ (x0 , y 0 ) and (z, t) ≡ (z 0 , t0 ). Then xy 0 = yx0 and

zt0 = tz 0 . Multiplying these two side by side, we get xy 0 zt0 = yx0 tz 0 , i.e.
xzy 0 t0 = ytx0 z 0 , meaning (xz, yt) ≡ (x0 z 0 , y 0 t0 ).
ii. Find a graph which has only three automorphisms.

Solution. Consider the graph whose points are
{a, a0 , a00 , a000 , b, b0 , b00 , b000 , c, c0 , c00 , c000 }
and whose vertices are
aa0 , aa00 , a00 a000 , bb0 , bb00 , b00 b000 , cc0 , cc00 , c00 c000 , ab, bc, ca, a0 b00 , b0 c00 , c0 a00 .
It works!
iii. Let a and b be two integers which are not both 0. We say that d is the
greatest common divisor of a and b if d is the largest natural number
that divides both a and b. Show that for any a, b ∈ Z, gcd(a, b) exists and
that there are x, y ∈ Z such that ax + by = gcd(a, b).
Proof: Replacing a and b by |a| and |b|, we may assume that a ≥ 0 and
b ≥ 0.
Existence. Since 1 divides both a and b and since any number that
divides both a and b can be at most max(a, b) > 0, the set of natural
numbers that divide both a and b is a finite nonempty set bounded by
max(a, b). Therefore there is a largest such number. This proves the
existence of gcd(a, b). We let d = gcd(a, b).
Second Part. We proceed by induction on max(a, b). If a = 1, then take
x = 1, y = 0. If b = 1, then take x = 0, y = 1. This takes care of the
initial step max(a, b). Assume max(a, b) > 1. If a = b, then d = a and

we may take x = 1, y = 0. Assume a 6= b. Without loss of generality, we
may assume that a > b. Note that the divisors of a and b are the same
as the divisors of a − b and b. Hence gcd(a − b, b) = gcd(a, b) = d. Since
max(a − b, b) < a = max(a, b), by induction there are two integers x and
y 0 such that x(a − b) + y 0 b = d, i.e. xa + (y 0 − x)b = d. Take y = y 0 − x.
iv. Let a and b be two nonzero integers. We say that m is the least common
multiple of a and b if m is the least natural number that is divisible by
both a and b. We let m = lcm(a, b). Show that for any a, b ∈ Z \ {0},
lcm(a, b) exists and that ab = ± gcd(a, b) lcm(a, b).
Proof: Replacing a and b by |a| and |b| again, we may assume that a > 0
and b > 0. Since a and b both divide ab, lcm(a, b) exists.
Let d = gcd(a, b) and m = lcm(a, b). Let a0 and b be such that a = da0
and b = db0 . Then ab = d2 a0 b0 . We need to prove that m = da0 b0 .
Since da0 b0 = ab0 = a0 b, a and b both divide da0 b0 .
Let x be divisible by both a and b. Then x = au = bv for some u, v. We
have a0 du = au = x = bv = b0 dv and so a0 u = b0 v. Since a0 and b0 cannot
have a common divisor (otherwise d would be larger), b0 must divide u.
(This last fact needs a serious proof, that we have not undertaken yet. I
shouldn’t have asked this question at this stage). Write u = cb0 . Now
x = au = acb0 = a0 dcb0 and so a0 b0 d divides x, in particular a0 b0 d ≤ x.
This shows that a0 b0 d is the least multiple of a and b, i.e. a0 b0 d = m.
v. Find formulas for the sums
12 + 22 + . . . + n2
and
13 + 23 + . . . + n3 ,
and prove your result.
Proof: We claim that
n(n + 1)(2n + 1)
12 + 22 + . . . + n2 = .
6
We proceed by induction on n. For n = 1, it is easy to check the validity
of the formula. Assume the statement holds for n. To prove it for n + 1,
we compute:
n(n+1)(2n+1)
12 + 22 + . . . + n2 + (n + 1)2 = 6 + (n + 1)2
(n+1)(n(2n+1)+6(n+1))
= 6
(n+1)(2n2 +7n+6))
= 6
(n+1)(n+2)(2n+3)
= 6
n0 (n0 +1)(2n0 +1)
= 6
26.6. SECOND SEMESTER, MIDTERM, MAY 2003 163
where n0 = n + 1. This proves the equality by induction.

We claim that
n2 (n + 1)2
13 + 23 + . . . + n3 = .
4
We proceed by induction on n. For n = 1, it is easy to check the validity
of the formula. Assume the statement holds for n. To prove it for n + 1,
we compute:
n2 (n+1)2
13 + 23 + . . . + n3 + (n + 1)3 = 4 + (n + 1)3
(n+1)2 (n2 +4n+4)
= 4
(n+1)2 (n+2)2
= 4
m2 (m+1)2
= 4
where m = n + 1. This proves the equality by induction.
vi. Recall that a natural number p 6= 0, 1 is called prime if whenever p divides

a product xy of two natural numbers x and y then p divides either x or y.
A natural number p 6= 0, 1 is called irreducible if whenever p = xy for
two natural numbers x and y then either x or y is 1. Show that a natural
number is prime if and only if it is irreducible.
Proof: Let p be prime. Assume that a|p. Then p = ab for some b.
It follows that p divides ab. Thus p divides either a or b. Assume –
without loss of generality – that p divides a. Then px = a some x. Hence
p = ab = pxb. Since p 6= 0, it follows that xb = 1. Thus b = 1, and so
a = p.
Let now p be an irreducible. We will prove that p is a prime. Let p divide
xy. We will show that p divides either x or y. We proceed by induction on
p+x+y. Dividing x and y by p we get x = pq1 +x1 and y = pq2 +y1 where
x1 , y1 < p. Since xy = (pq1 +x1 )(pq2 +y1 ) = p(pq1 q2 +q1 y1 +q2 x1 )+x1 y1 ,
thus p divides x1 y1 . Assume x1 y1 6= 0. Thus p ≤ x1 y1 < p2 . It follows
that x1 y1 = rp for some r = 1, . . . , p − 1. If r = 1, then either p = x1
or p = y1 , a contradiction. Let q be an irreducible dividing r. Thus
q ≤ r < p. By induction q divides either x1 or y1 , say q divides x1 . Write
x1 = qx2 and r = qr0 . We have qx2 y1 = x1 y1 = rp = qr0 p and x2 y1 = r0 p.
By induction p divides either x2 or y1 , in which case it divides x or y
(respectively). Thus we may assume that x1 y1 = 0. Hence one of x1 or
y1 is 0, say x1 = 0. Then x = pq1 + x1 = pq1 and p divides x. ¤
26.6 Second Semester, Midterm, May 2003

Let X be a set. A filter on X is a set = of subsets of X that satisfies the
following properties:
i) If A ∈ = and A ⊆ B ⊆ X, then B ∈ =.
ii) If A and B are in =, then so is A ∩ B.
iii) ∅ 6∈ = and X ∈ =.
If A ⊆ X is a fixed nonempty subset of X, then the set of subsets =(A) of
X that contain A is a filter on X. Such a filter is called principal filter. If X
is infinite, then the set of cofinite subsets of X is a filter, called Fréchet filter.
A filter is called ultrafilter if it is a maximal filter.
We fix a set X.
i. Show that a principle filter =(A) on X is an ultrafilter if and only if A is

a singleton set.
Proof: It is clear that if ∅ 6= B ⊂ A, then =(A) ⊂ =(B). Thus =(A)
cannot be an ultrafilter unless A is a singleton set. Conversely let A =
{a} is a singleton set. Let = be a filter properly containing =(A). Let
Y ∈ = \ =(A). Then a 6∈ Y and so a ∈ Y c . Therefore Y c ∈ =(A) ⊂ =. It
follows that ∅ = Y ∩ Y c ∈ =, a contradiction.
ii. Show that the Fréchet filter (on an infinite set) is not contained in a
principal filter.
Proof: Let = be the Fréchet filter. Assume = ⊆ Im(A) for some A ⊆ X.
Let a ∈ A. Then X \ {a}, being a cofinite set, is in =, but is not in =.
iii. Show that the intersection of a set of filters is a filter.

Proof: All three conditions of the definition are trivially met.
iv. Show that if = is a set of subsets of X such that A1 ∩ A2 ∩ . . . ∩ An 6= ∅ for

any A1 , A2 , . . . , An ∈ =, then there is a filter that contains =. Describe
this filter in terms of =.
Proof: Let
h=i = {A ⊆ X : there are A1 , . . . , An ∈ = such that A1 , ∩ . . .∩An ⊆ A}.
It is easy to show that = ⊆ h=i and that h=i is a filter. You should also
note that h=i is the smallest filter that contains =.
v. Show that a filter = is an ultrafilter if and only if for any A ⊆ X, either

A or Ac is in =. Conclude that in an ultrafilter =, if A t B ∈ =, then one
of A or B in =.
Proof: Suppose first that = is a filter such that for any A ⊆ X, either A
or Ac is in =. Let =0 be a filter properly containing =. Let A ∈ =0 \ =.
Then Ac ∈ = ⊂ =0 . Thus ∅ = A ∩ Ac ∈ =0 , a contradiction.
Conversely, let = be an ultrafilter. Let A ⊆ X. Assume that neither A
nor Ac is in =. By question number iv, there are B and C in = such that
A ∩ B = ∅ and Ac ∩ C = ∅. Then A ∩ B ∩ C = ∅ and Ac ∩ B ∩ C = ∅,
implying that ∅B ∩ C ∈ =, a contradiction.
Assume now A t B is in the ultrafilter =. Then either A or Ac is in =. In
the second case B = Ac ∩ (A t B) ∈ =.
26.6. SECOND SEMESTER, MIDTERM, MAY 2003 165
vi. Conclude that any ultrafilter on X that contains a finite subset of X is

a principal filter. Deduce that every nonprincipal ultrafilter contains the
Fréchet filter.
Proof: If = is an ultrafilter that contains a finite subset, let A be a
subset of = of smallest size. Let a ∈ A}. By the second part of the
previous question, either {a} ∈ = or A \ {a} ∈ =. Hence A = {a}.
Now let = be a nonprincipal ultrafilter. Then it does not contain any finite
subset. By the previous question, it contains all cofinite subsets, hence
the Fréchet filter is a subset of =.
vii. (AC) Show that for any filter = on X, there is an ultrafilter on X that
contains =.
Proof: Let Z be the set of filters on X that contains =. Order Z by
inclusion. Since = ∈ Z, Z 6= ∅. It is clear that Z is an inductive set,
because if (=i )i∈I is a chain from Z, then one can show easily that ∪i∈I =i
is a filter that contains =. Hence by Zorn’s Lemma Z has a maximal
element. Clearly this maximal element of Z is an ultrafilter.
viii. (AC) Show that if X is infinite then there are nonprincipal ultrafilters on
X.
Proof: Apply the previous question to the Fréchet Filter to get an ultra-
filter = that contains the Fréchet filter. By ii, = is not principal.
ix. Let = be a filter on X. Let Ax (x ∈ X) be sets. Consider the product

Q
x∈X Ax := {f = (fx )x∈X
Q : fx ∈ Ax for all x ∈ X}
= {f : X −→ x∈X Ax : f (x) ∈ Ax for all x ∈ X}.
Q
On the set x∈XAx consider the relation defined by
f ≡ g ⇐⇒ {x ∈ X : f (x) = g(x)} ∈ =.
Show that this is an equivalence relation.

Proof: Reflexivity and symmetry are clear. Let us show transitivity. Let
f ≡ g and g ≡ h. Then {x ∈ X : f (x) = g(x)} ∩ {x ∈ X : g(x) = h(x)} ⊆
{x ∈ X : f (x) = h(x)} ∈ = because the set on the left hand side is in =.
Q
FromQnow on we fix a filter = on X and we let M = x∈X Ax / ≡. For
f ∈ x∈X Ax , we will let [f ] to denote its class in M .
x. Show that if Ax = A for all x ∈ X and if = is nonprincipal filter then the

map that sends an element a ∈ A to the class of the constant function a
is an injection from A into M .
Proof: Let a, b ∈ A. Assume [(a)x∈X ] = [(b)x∈X ]. Then {x ∈ X : a =
b} ∈ =, so is nonempty, so a = b.
xi. Suppose that each Ax is a set ordered by a relation <. On M define the
relation < by,
[(fx )x∈X ] < [(gx )x∈X ⇐⇒ {x ∈ X : fx < gx } ∈ =.
Show that this defines an order on M .

Proof: It is clear that [f ] 6< [f ] for any [f ] ∈ M . Assume [f ] < [g] and
[g] < [h]. Then {x ∈ X : fx < gx } ∈ = and {x ∈ X : gx < hx } ∈ =. So
their intersection is also in =. But the set {x ∈ X : fx < hx } contains this
intersection, so is also in =. Hence [f ] < [g].
xii. Show that if = is an ultrafilter and if each (Ax , <) is totally ordered then
so is (M, <).
Proof: Let [f ], [g] ∈ M . Assume [f ] 6= [g]. Then {x ∈ X : fx = gx } 6∈ =
and so (by Question v) {x ∈ X : fx 6= gx } ∈ =. Therefore, since the sets
{x ∈ X : fx < gx }
and
{x ∈ X : fx > gx }
partition the set {x ∈ X : fx 6= gx } which is in =, one of them should be
in = by Question v again, in which case either [f ] < [g] or [g] < [f ].
xiii. Suppose that each (Ax , <) has a largest element. Is it true that (M, <)
has a largest element?
Proof: Of course!
xiv. Suppose that each (Ax , <) is well-ordered. Is it true that (M, <) is well-
ordered?
Proof:
Index
(X, ?), 35 vn , 69
(n, m) = k, 111 (, 69
(x, y), 29, 82 ), 69
(x, y, z), 30 2+2=4, 91
=, 69 2 × 2 = 4, 93
A−1 , 83
X/ ≡, 50 1, 80
X × Y , 29 2,
√ 80
∅, 13, 77 2, 19
⇔ , 15 3, 80
⇒,
T 15 4, 80
Sx∈X x, 26
i∈I Ai , 28
absolute value, 108
∩X, 26 addition, 58, 90
∩, 25 addition of natural numbers, 90, 91
◦, 32, 41, 85 addition of ordinals, 127
∪, 27, 79, 81 antisymmetric relation, 48
≡, 69 arrival set of a function, 31
∃, 69 associativity, 36, 41
∀, 15 associativity for multiplication, 93
∈, 11, 69 associativity of addition, 91
≤, 54 associativity of functions, 33
¬, 69 Aut(Γ), 63
6∈
Q, 11 automorphism group, 63
Qi∈I X, 44 automorphism of a binary unirela-
Qi∈I Xi , 44 tional structure, 63
i∈I fi , 45 automorphism of well-ordered sets,
\, 25, 78
√ 123
2, 18 automorphisms of (Q, <), 129
⊂, 13 automorphisms of (R, +, ×), 129
⊆, 12, 77 Axiom of Regularity, 152
∧, 69 Axiom of Replacement, 126, 153
℘(x), 82
{x}, 11 back and forth argument, 67
{x1 , . . . , xn }, 11 bijection, 40
f −1 , 40 binary operation, 35
f1 × f2 , 45 binary relation, 47, 153
167
168 INDEX
branch, 132 finite set, 11, 96

finitely branching tree, 132
canonical surjection, 52 first coordinate, 29
cardinality, 96 f n , 32
Cartesian product, 29, 44, 82 for all, 15
characteristic function, 41 Func(X, Y ), 32
classification of well-ordered sets, 126 function, 31, 84
commutativity, 36
commutativity of addition, 91 gcd(n, m), 111
commutativity of multiplication, 94 germ, 52
complement of a set, 25 graph, 64, 84
complete graph, 64 greatest common divisor, 110, 111
composite of two functions, 32
composition, 32, 41, 85 identity element, 36, 41
conjunction, 69 identity function, 32, 85
connected vertices, 64 IdX , 32, 33
constant function, 32 if . . . then, 15
Continuum Hypothesis, 137 if and only if, 15
cycles, 42 image, 33
cyclic representation, 42 image restriction, 85
induced map, 52
De Morgan laws, 27 induction, 57
definable subsets, 78 inductive set, 87
∆, 29, 36 ∞, 20
diagonal, 83, 85 infinite set, 11, 96
difference, 25, 78 infinity, 20
disjoint sets, 25 initial segment, 122, 152
domain, 84 injection, 38
domain of a function, 31 integers, 18
dual of a binary unirelational struc- intersection, 25, 26, 78
ture, 63 interval, 20
inverse element, 41
element, 11, 76 inverse image, 33
emptyset, 13, 76, 77 inverse of a bijection, 40
equal sets, 12 irreducible numbers, 98
equality, 76 irreflexive relation, 48
equality of sets, 11 irreflexivity, 155
equality symbol, 69 isomorphic binary unirelational struc-
equivalence class, 50 tures, 63
equivalence relation, 48 isomorphism of a binary unirelational
equivalence relation generated by, 52 structure, 63
existential quantifier, 69 isomorphism of well-ordered sets, 123
exponentiation, 58
König’s Lemma, 132
factorial, 58
family, 23 lcm(n, m), 111
INDEX 169
least common multiple, 111 product of the functions, 45

least upper bound, 155 projection, 32
left distributivity, 93 property, 78
left identity element, 36 ℘(X), 13
left parentheses, 69
lexicographic ordering, 122 Q,√18
limit ordinal, 126 Z[ 2], 19
quotient set, 50
map, 31, 84
member, 11 R, 19
membership relation, 69 rational numbers, 18
minus, 78 real numbers, 18
morphism of well-ordered sets, 123 reflexive relation, 48
multiplication, 58 reflexivity, 49
multiplication of natural numbers, related vertices, 64
92, 93 relations, 47
multiplication of ordinals, 127 restriction, 85
restriction of a function, 32
N, 17, 87 right distributivity, 94
natural numbers, 17 right identity element, 36
negation, 69 right identity element for multipli-
node, 132 cation, 93
nZ, 109 right parentheses, 69
right simplification for addition, 91,
o(g), 44 92
ω, 87 Russell, Bertrand, 79
one-to-one correspondence, 40
one-to-one function, 38 second coordinate, 29
onto, 39 sequence, 44
open subset of Q, 22 set, 9, 11, 76
open subset of R, 22 set of arrival, 84
ordp , 109 set of subsets, 13
ordered pair, 82 sets of sets, 21
ordering of natural numbers, 94 sgn, 109
ordinal, 125, 152 σ-algebra, 129
sign function, 109
pair, 29 simplification for multiplication, 94
Paradox of Burali-Forti, 127 singleton set, 11, 80
parentheses, 69 subgroup, 109, 118
partial order, 53, 155 subset, 12, 77
partition, 50 subset defined by a property, 20
poset, 53 successor, 122
power set, 82 successor function, 58
prime number, 98 superset, 12
prime to each other, 112 surjection, 39
product, 35 S(x), 87, 122
170 INDEX
symmetric difference, 29, 36

symmetric group, 41
symmetric relation, 48
symmetry, 49
Sym(N), 43
Sym(n), 41, 43
Sym(X), 41
total order, 55, 157

transfinite induction, 122, 152
transitive closure, 154
transitive relation, 48, 153
transitivity, 49, 155
tree, 132
type of an element of Sym(n), 43
ultrafilters, 143
ultraproduct, 143
union, 27, 79, 81
union of two functions, 38
upper bound, 155
variables, 69
vertex, 132
well-ordered set, 88, 121, 152, 157
X c , 25
Z, 18
Z,√18
Z[ 2], 19
zero, 80

Set Theory Lecture Notes

Uploaded by

Copyright:

Available Formats

Set Theory Lecture Notes

Uploaded by

Document Information

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

Set Theory Lecture Notes

Uploaded by

Copyright:

Available Formats

Foundations of Mathematics I

February 12, 2004

I Naive Set Theory 7

2 Operations with Sets 25

7 Some Automorphism Groups 63

II Axiomatic Set Theory 73

12.8 Divisibility and Subgroups . . . . . . . . . . . . . . . . . . . . . . 109

13 Rational Numbers 115

14 Real Numbers 119

15 Well-Ordered Sets 121

18 Axiom of Choice and Zorn’s Lemma 131

19 Axioms of Set Theory – ZFC 133

21 Continuum Hypothesis 137

22 Banach-Tarski Paradox 139

23 First Order Structures 141

24 Ultraproducts and Ultrafilters 143

25 Dimension Theory 145

Naive Set Theory

Basic Concepts and

1.1 Sets, Subsets and Emptyset

The set {0, 1, 2, 3, . . .} is infinite.

Theorem 1.1.1 A set with no elements is a subset of every set.

Corollary 1.1.2 There is at most one set without elements.

Theorem 1.1.3 If |X| = n is finite then |℘(X)| = 2n .

after this theorem.

Proof: Assume |X| = n. To form a subset of X, we must decide which elements

i. Let X = {0, 1}. List the elements of {℘(A) : A ∈ ℘(X)}.

ii. Let X = ∅. List the elements of {℘(A) : A ∈ ℘(X)}.

iii. Show that ∅ 6= {∅}.

iv. Write the 16 elements of ℘(℘({0, 1})).

v. Show that ℘(∅) = {∅}.

vi. Show that ℘(℘(∅)) = {∅, {∅}}.

vii. Find ℘(℘(℘(∅))).

viii. Find a set X with an element x such that x ∈ X and x ⊆ X.

ix. Can we have x ⊆ {x}? Answer: According to the definition, x ⊆ {x}

x. Suppose x = {x}. Find ℘(x). Is x ⊆ ℘(x)? Find ℘(℘(x)). Is x ⊆ ℘(℘(x))?

xi. Show that X = ∅ satisfies the following formula: “for every x ∈ X, x is a

xiii. Show that X ⊆ Y if and only if ℘(X) ⊆ ℘(Y ).

xiv. Show that if X ⊆ Y then ℘(X) ∈ ℘(℘(Y )).

xvi. Show that for any set X, {∅} ∈ ℘(℘(X)).

xvii. Show that for any set X, {∅} ∈ ℘(℘(℘(℘(℘(X))))).

xviii. Show that for any set X, {X} ∈ ℘(℘(X)).

xix. Show that for any set X, {{∅}, {{X}}} ∈ ℘(℘(℘(℘(X)))).

xx. Show that {℘(A) : A ⊆ X} ∈ ℘(℘(℘(X))).

1.2 Notes on Formalism

either there is an element in x which is not in y

∃z((z ∈ x ∧ z 6∈ y) ∨ (∃z(z ∈ y ∧ z 6∈ x)).

What about the proposition that x ⊂ y, how is it formalized? By definition,

1.3 Number Sets

whose elements are natural numbers is denoted by the symbol N. Thus

For example 5 ∈ N, but −3 6∈ N, 2/3 6∈ N.

Z = {. . . , −3, −2, −1, 0, 1, 2, 3, . . .}.

We have −3, 5 ∈ Z, but 3/5 6∈ Z. Clearly N ⊂ Z. We can add and multiply

Q = {a/b : a, b ∈ Z and b 6= 0}.

Lemma 1.3.1 There is no q ∈ Q such that q 2 = 2.

Proof: Assume not. Let q ∈ Q be such that q 2 = 2. Let a, b ∈ Z be such that

is closed under division (by a nonzero element). We leave this as an exercise to

Exercises and Examples.

ii. Show that there is no smallest rational number > 0.

1.4 Subsets Defined By a Property