Chapter 2

Basic Set Theory

A set is a Many that allows itself to
be thought of as a One.
- Georg Cantor
This chapter introduces set theory, mathematical induction, and formalizes the notion of mathematical
functions. The material is mostly elementary. For
those of you new to abstract mathematics elementary
does not mean simple (though much of the material
is fairly simple). Rather, elementary means that the
material requires very little previous education to understand it. Elementary material can be quite challenging and some of the material in this chapter, if
not exactly rocket science, may require that you adjust you point of view to understand it. The single
most powerful technique in mathematics is to adjust
your point of view until the problem you are trying
to solve becomes simple.
Another point at which this material may diverge
from your previous experience is that it will require
proof. In standard introductory classes in algebra,
trigonometry, and calculus there is currently very little emphasis on the discipline of proof. Proof is, however, the central tool of mathematics. This text is
for a course that is a students formal introduction to
tools and methods of proof.

2.1

Set Theory

can be written:
{2n : n is an integer}
The opening and closing curly braces denote a set, 2n
specifies the members of the set, the colon says such
that or where and everything following the colon
are conditions that explain or refine the membership.
All correct mathematics can be spoken in English.
The set definition above is spoken The set of twice
n where n is an integer.
The only problem with this definition is that we
do not yet have a formal definition of the integers.
The integers are the set of whole numbers, both positive and negative: {0, 1, 2, 3, . . .}. We now introduce the operations used to manipulate sets, using
the opportunity to practice curly brace notation.
Definition 2.1 The empty set is a set containing
no objects. It is written as a pair of curly braces with
nothing inside {} or by using the symbol .
As we shall see, the empty set is a handy object.
It is also quite strange. The set of all humans that
weigh at least eight tons, for example, is the empty
set. Sets whose definition contains a contradiction or
impossibility are often empty.
Definition 2.2 The set membership symbol is
used to say that an object is a member of a set. It
has a partner symbol
/ which is used to say an object
is not in a set.

A set is a collection of distinct objects. This means

that {1, 2, 3} is a set but {1, 1, 3} is not because 1
appears twice in the second collection. The second Definition 2.3 We say two sets are equal if they
collection is called a multiset. Sets are often specified have exactly the same members.
with curly brace notation. The set of even integers
23

24

CHAPTER 2. BASIC SET THEORY

S T = {1, 3, 5},

Example 2.1 If
S = {1, 2, 3}
then 3 S and 4
/ S. The set membership symbol
is often used in defining operations that manipulate
sets. The set
T = {2, 3, 1}
is equal to S because they have the same members: 1,
2, and 3. While we usually list the members of a set
in a standard order (if one is available) there is no
requirement to do so and sets are indifferent to the
order in which their members are listed.
Definition 2.4 The cardinality of a set is its size.
For a finite set, the cardinality of a set is the number
of members it contains. In symbolic notation the size
of a set S is written |S|. We will deal with the idea
of the cardinality of an infinite set later.
Example 2.2 Set cardinality
For the set S = {1, 2, 3} we show cardinality by writing |S| = 3

S U = {2, 3, 5}, and

T U = {3, 4, 5}
Definition 2.6 If A and B are sets and A B =
then we say that A and B are disjoint, or disjoint
sets.
Definition 2.7 The union of two sets S and T is
the collection of all objects that are in either set. It
is written S T . Using curly brace notion
S T = {x : (x S) or (x T )}
The symbol or is another Boolean operation, one that
is true if either of the propositions it joins are true.
Its symbolic equivalent is which lets us re-write the
definition of union as:
S T = {x : (x S) (x T )}

Example 2.4 Unions of sets.

Suppose S = {1, 2, 3}, T = {1, 3, 5}, and U =
{2, 3, 4, 5}.
We now move on to a number of operations on sets. Then:
You are already familiar with several operations on S T = {1, 2, 3, 5},
numbers such as addition, multiplication, and negation.
S U = {1, 2, 3, 4, 5}, and
Definition 2.5 The intersection of two sets S and
T is the collection of all objects that are in both sets.
It is written S T . Using curly brace notation
S T = {x : (x S) and (x T )}
The symbol and in the above definition is an example of a Boolean or logical operation. It is only
true when both the propositions it joins are also true.
It has a symbolic equivalent . This lets us write the
formal definition of intersection more compactly:
S T = {x : (x S) (x T )}
Example 2.3 Intersections of sets
Suppose S = {1, 2, 3, 5},
T = {1, 3, 4, 5}, and U = {2, 3, 4, 5}.
Then:

T U = {1, 2, 3, 4, 5}
When performing set theoretic computations, you
should declare the domain in which you are working.
In set theory this is done by declaring a universal set.
Definition 2.8 The universal set, at least for a
given collection of set theoretic computations, is the
set of all possible objects.
If we declare our universal set to be the integers then
{ 21 , 32 } is not a well defined set because the objects
used to define it are not members of the universal
set. The symbols { 21 , 23 } do define a set if a universal
set that includes 12 and 23 is chosen. The problem
arises from the fact that neither of these numbers are
integers. The universal set is commonly written U.
Now that we have the idea of declaring a universal
set we can define another operation on sets.

25

2.1. SET THEORY

2.1.1

Venn Diagrams

A Venn diagram is a way of depicting the relationship

between sets. Each set is shown as a circle and circles
overlap if the sets intersect.
Example 2.5 The following are Venn diagrams for
the intersection and union of two sets. The shaded
parts of the diagrams are the intersections and unions
respectively.

notation for not is . There is not much savings in

space as the definition of compliment becomes
S c = {x : (x S)}
Example 2.6 Set Compliments
(i) Let the universal set be the integers. Then the
compliment of the even integers is the odd integers.
(ii) Let the universal set be {1, 2, 3, 4, 5}, then the
compliment of S = {1, 2, 3} is S c = {4, 5} while
the compliment of T = {1, 3, 5} is T c = {2, 4}.
(iii) Let the universal set be the letters {a, e, i, o, u, y}.
Then {y}c = {a, e, i, o, u}.
The Venn diagram for Ac is

AB

Ac

AB

Notice that the rectangle containing the diagram is

labeled with a U representing the universal set.
Definition 2.9 The compliment of a set S is the
collection of objects in the universal set that are not
in S. The compliment is written S c . In curly brace
notation
S c = {x : (x U) (x
/ S)}
or more compactly as
S c = {x : x
/ S}

We now have enough set-theory operators to use them

to define more operators quickly. We will continue to
give English and symbolic definitions.
Definition 2.10 The difference of two sets S and
T is the collection of objects in S that are not in T .
The difference is written S T . In curly brace notation
S T = {x : x (S (T c ))},

or alternately

S T = {x : (x S) (x
/ T )}

Notice how intersection and complementation can be

used together to create the difference operation and
that the definition can be rephrased to use Boolean
operations. There is a set of rules that reduces the
There is also a Boolean symbol associated with the number of parenthesis required. These are called opcomplementation operation: the not operation. The erator precedence rules.
however it should be apparent that the compliment of
a set always depends on which universal set is chosen.

26

CHAPTER 2. BASIC SET THEORY

(i) Other things being equal, operations are per- Another important tool for working with sets is the
formed left-to-right.
ability to compare them. We have already defined
what it means for two sets to be equal, and so by
(ii) Operations between parenthesis are done first, implication what it means for them to be unequal.
starting with the innermost of nested parenthe- We now define another comparator for sets.
sis.
Definition 2.12 For two sets S and T we say that
(iii) All complimentations are computed next.
S is a subset of T if each element of S is also an
(iv) All intersections are done next.
element of T . In formal notation S T if for all
x S we have x T .
(v) All unions are performed next.
If S T then we also say T contains S which
(vi) Tests of set membership and computations,
can
be
written T S. If S T and S 6= T then we
equality or inequality are performed last.
write S T and we say S is a proper subset of T .
Special operations like the set difference or the
symmetric difference, defined below, are not included Example 2.9 Subsets
in the precedence rules and thus always use paren- If A = {a, b, c} then A has eight different subsets:
thesis.

{a}
{b}
{c}
Example 2.7 Operator precedence
Since complementation is done before intersection
the symbolic definition of the difference of sets can be
rewritten:
S T = {x : x S T c }

If we were to take the set operations

A B Cc

and put in the parenthesis we would get

(A (B (C c )))
Definition 2.11 The symmetric difference of
two sets S and T is the set of objects that are in one
and only one of the sets. The symmetric difference is
written ST . In curly brace notation:
ST = {(S T ) (T S)}
Example 2.8 Symmetric differences
Let S be the set of non-negative multiples of two that
are no more than twenty four. Let T be the nonnegative multiples of three that are no more than
twenty four. Then
ST = {2, 3, 4, 8, 9, 10, 14, 15, 16, 20, 21, 22}

{a, b}

{a, c}

{b, c}

{a, b, c}

Notice that A A and in fact each set is a subset of

itself. The empty set is a subset of every set.
We are now ready to prove our first proposition.
Some new notation is required and we must introduce an important piece of mathematical culture. If
we say A if and only if B then we mean that either
A and B are both true or they are both false in any
given circumstance. For example: an integer x is
even if and only if it is a multiple of 2. The phrase
if and only if is used to establish logical equivalence. Mathematically, A if and only if B is a way
of stating that A and B are simply different ways
of saying the same thing. The phrase if and only
if is abbreviated iff and is represented symbolically
as the double arrow . Proving an iff statement is
done by independently demonstrating that each may
be deduced from the other.
Proposition 2.1 Two sets are equal if and only if
each is a subset of the other. In symbolic notation:
(A = B) (A B) (B A)

Proof:
Another way to think about this is that we need numbers that are positive multiples of 2 or 3 (but not both) Let the two sets in question be A and B. Begin by
assuming that A = B. We know that every set is
that are no more than 24.

27

2.1. SET THEORY

a subset of itself so A A. Since A = B we may
substitute into this expression on the left and obtain
B A. Similarly we may substitute on the right and
obtain A B. We have thus demonstrated that if
A = B then A and B are both subsets of each other,
giving us the first half of the iff.
Assume now that A B and B A. Then
the definition of subset tells us that any element of
A is an element of B. Similarly any element of B
is an element of A. This means that A and B have
the same elements which satisfies the definition of set
equality. We deduce A = B and we have the second
half of the iff. 2

where it is false. It is thus possible for a false mathematical statement to be true most of the time. In
the next chapter we will develop the theory of prime
numbers. For now we will assume the reader has a
modest familiarity with the primes. The statement
Prime numbers are odd is false once, because 2 is a
prime number. All the other prime numbers are odd.
The statement is a false one. This very strict definition of what makes a statement true is a convention
in mathematics. We call 2 a counter example. It is
thus necessary to find only one counter-example to
demonstrate a statement is (mathematically) false.

A note on mathematical grammar: the symbol 2 indicates the end of a proof. On a paper turned in by a
student it is usually taken to mean I think the proof
ends here. Any proof should have a 2 to indicate its
end. The student should also note the lack of calculations in the above proof. If a proof cannot be read
back in (sometimes overly formal) English then it is
probably incorrect. Mathematical symbols should be
used for the sake of brevity or clarity, not to obscure
meaning.

Example 2.10 Disproof by counter example

Prove that the statement A B = A B is false.

Proposition 2.2 De Morgans Laws Suppose

that S and T are sets. DeMorgans Laws state that
(i) (S T )c = S c T c , and
(ii) (S T )c = S c T c .
Proof:
c

Let A = {1, 2} and B = {3, 4}. Then A B =

while A B = {1, 2, 3, 4}. The sets A and B form a
counter-example to the statement.

Problems
Problem 2.1 Which of the following are sets? Assume that a proper universal set has been chosen and
answer by listing the names of the collections of objects that are sets. Warning: at least one of these
items has an answer that, while likely, is not 100%
certain.
(i) A = {2, 3, 5, 7, 11, 13, 19}
(ii) B = {A, E, I, O, U }

Let x (S T ) ; then x is not a member of S or

T . Since x is not a member of S we see that x (iii) C = { x : x < 0}

S c . Similarly x T c . Since x is a member of both
these sets we see that x S c T c and we see that (iv) D = {1, 2, A, 5, B, Q, 1, V }
(S T )c S c T c . Let y S c T c . Then the
(v) E is the list of first names of people in the 1972
definition of intersection tells us that y S c and
phone book in Lawrence Kansas in the order
y T c . This in turn lets us deduce that y is not a
they appear in the book. There were more than
member of S T , since it is not in either set, and
35,000 people in Lawrence that year.
so we see that y (S T )c . This demonstrates that
S c T c (S T )c . Applying Proposition 2.1 we get (vi) F is a list of the weight, to the nearest kilogram,
that (S T )c = S c T c and we have proven part (i).
of all people that were in Canada at any time in
The proof of part (ii) is left as an exercise. 2
2007.
In order to prove a mathematical statement you must
prove it is always true. In order to disprove a mathe- (vii) G is a list of all weights, to the nearest kilogram,
that at least one person in Canada had in 2007.
matical statement you need only find a single instance

28

CHAPTER 2. BASIC SET THEORY

Problem 2.2 Suppose that we have the set U =

{n : 0 n < 100} of whole numbers as our
universal set. Let P be the prime numbers in U ,
let E be the even numbers in U , and let F =
{1, 2, 3, 5, 8, 13, 21, 34, 55, 89}. Describe the following
sets either by listing them or with a careful English
sentence.
(i) E c ,
(ii) P F ,
(iii) P E,
(iv) F E F E c , and
c

(v) F F .

Problem 2.5 Find an example of an infinite set that

has a finite complement, be sure to state the universal
set.
Problem 2.6 Find an example of an infinite set that
has an infinite complement, be sure to state the universal set.
Problem 2.7 Add parenthesis to each of the following expressions that enforce the operator precedence
rules as in Example 2.7. Notice that the first three describe sets while the last returns a logical value (true
of false).
(i) A B C D
(ii) A B C D

Problem 2.3 Suppose that we take the universal set (iii) Ac B c C

U to be the integers. Let S be the even integers, let
T be the integers that can be obtained by tripling any (iv) A B = A C
one integer and adding one to it, and let V be the set
Problem 2.8 Give the Venn diagrams for the folof numbers that are whole multiples of both two and
lowing sets.
three.
(i) A B (ii) B A (iii) Ac B
(i) Write S, T , and V using symbolic notation.
(iv) AB (v) (AB)c (vi) Ac B c
(ii) Compute S T , S V and T V and give symbolic representations that do not use the symbols
U
S, T , or V on the right hand side of the equals
sign.
Problem 2.4 Compute the cardinality of the following sets. You may use other texts or the internet.

2
7

(i) Two digit positive odd integers.

(ii) Elements present in a sucrose molecule.

(iii) Isotopes of hydrogen that are not radioactive.
(iv) Planets orbiting the same star as the planet you
are standing on that have moons. Assume that
Pluto is a minor planet.

3
1

0
4
C

(v) Elements with seven electrons in their valence

Problem 2.9 Examine the Venn diagram above.
shell. Remember that Ununoctium was discovered in 2002 so be sure to use a relatively recent Notice that every combination of sets has a unique
number in common. Construct a similar collection
reference.
of four sets.
(vi) Subsets of S = {a, b, c, d} with cardinality 2.
Problem 2.10 Read Problem 2.9. Can a system of
(vii) Prime numbers whose base-ten digits sum to ten. sets of this sort be constructed for any number of sets?
Be careful, some have three digits.
Explain your reasoning.

29

2.2. MATHEMATICAL INDUCTION

Problem 2.11 Suppose we take the universal set to
be the set of non-negative integers. Let E be the
set of even numbers, O be the set of odd numbers
and F = {0, 1, 2, 3, 5, 8, 13, 21, 34, 89, 144, ...} be the
set of Fibonacci numbers. The Fibonacci sequence is
0, 1, 1, 2, 3, 5, 8, . . . in which the next term is obtained
by adding the previous two.
(i) Prove that the intersection of F with E and O
are both infinite.

2.2

Mathematical Induction

Mathematical induction is a technique used in proving mathematical assertions. The basic idea of induction is that we prove that a statement is true in one
case and then also prove that if it is true in a given
case it is true in the next case. This then permits the
cases for which the statement is true to cascade from
the initial true case. We will start with the mathematical foundations of induction.

(ii) Make a Venn diagram for the sets E, F , and O, We assume that the reader is familiar with the symand explain why this is a Mickey-Mouse problem. bols <, >, and . From this point on we will
denote the set of integers by the symbol Z. The
Problem 2.12 A binary operation is commuta- non-negative integers are called the natural numbers.
tive if x y = y x. An example of a commuta- The symbol for the set of natural numbers is N. Any
tive operation is multiplication. Subtraction is non- mathematical system rests on a foundation of axioms.
commutative. Determine, with proof, if union, inter- Axioms are things that we simply assume to be true.
section, set difference, and symmetric difference are We will assume the truth of the following principle,
commutative.
adopting it as an axiom.
Problem 2.13 An identity for an operation is The well-ordering principle: Every non-empty

an object i so that, for all objects x, i x = x i = set of natural numbers contains a smallest
element.
x. Find, with proof, identities for the operations set
union and set intersection.
The well ordering principle is an axiom that
agrees with the common sense of most people familProblem 2.14 Prove part (ii) of Proposition 2.2.
iar with the natural numbers. An empty set does
not contain a smallest member because it contains
Problem 2.15 Prove that
no members at all. As soon as we have a set of natA (B C) = (A B) C
ural numbers with some members then we can order
those members in the usual fashion. Having ordered
Problem 2.16 Prove that
them, one will be smallest. This intuition agreeing
A (B C) = (A B) C
with this latter claim depends strongly on the fact
the integers are whole numbers spaced out in inProblem 2.17 Prove that
crements of one. To see why this is important consider the smallest positive distance. If such a distance
A(BC) = (AB)C
existed, we could cut it in half to obtain a smaller
Problem 2.18 Disprove that
distance - the quantity contradicts its own existence.
The well-ordering principle can be used to prove the
A(B C) = (AB) C
correctness of induction.
Problem 2.19 Consider the set S = {1, 2, 3, 4}. For
each k = 0, 1, . . . , 4 how many k element subsets does Theorem 2.1 Mathematical Induction I Suppose that P (n) is a proposition that it either true or
S have?
false for any given natural numbers n. If
Problem 2.20 Suppose we have a set S with n 0
elements. Find a formula for the number of different (i) P (0) is true and,
subsets of S that have k elements.
(ii) when P (n) is true so is P (n + 1)
Problem 2.21 For finite sets S and T , prove
Then we may deduce that P (n) is true for any natural
|S T | = |S| + |T | |S T |
number.

30

CHAPTER 2. BASIC SET THEORY

Proof:

induction that the proposition is true for all natural

numbers. 2

Assume that (i) and (ii) are both true statements. Let S be the set of all natural numbers for
which P (n) is false. If S is empty then we are done,
so assume that S is not empty. Then, by the well
ordering principle, S has a least member m. By (i)
above m 6= 0 and so m 1 is a natural number. Since
m is the smallest member of S it follows that P (m1)
is true. But this means, by (ii) above, that P (m) is
true. We have a contradiction and so our assumption
that S 6= must be wrong. We deduce S is empty
and that as a consequence P (n) is true for all n N.
2

The set of all subsets of a given set is itself an important object and so has a name.

The technique used in the above proof is called proof

by contradiction. We start by assuming the logical
opposite of what we want to prove, in this case that
there is some m for which P (m) is false, and from
that assumption we derive an impossibility. If an assumption can be used to demonstrate an impossibility
then it is false and its logical opposite is true.
A nice problem on which to demonstrate mathematical induction is counting how many subsets a finite
set has.
Proposition 2.3 Subset counting. A set S with
n elements has 2n subsets.
Proof:
First we check that the proposition is true when
n = 0. The empty set has exactly one subset: itself. Since 20 = 1 the proposition is true for n = 0.
We now assume the proposition is true for some n.
Suppose that S is a set with n + 1 members and that
x S. Then S {x} (the set difference of S and a set
{x} containing only x) is a set of n elements and so,
by the assumption, has 2n subsets. Now every subset
of S either contains x or it fails to. Every subset of S
that does not contain x is a subset of S {x} and so
there are 2n such subsets of S. Every subset of S that
contains x may be obtained in exactly one way from
one that does not by taking the union with {x}. This
means that the number of subsets of S containing or
failing to contain x are equal. This means there are
2n subsets of S containing x. The total number of
subsets of S is thus 2n + 2n = 2n+1 . So if we assume
the proposition is true for n we can demonstrate that
it is also true for n + 1. It follows by mathematical

Definition 2.13 The set of all subsets of a set S

is called the powerset of S. The notation for the
powerset of S is P(S).
This definition permits us to rephrase Proposition 2.3
as follows: the power set of a set of n elements has
size 2n .
Theorem 2.1 lets us prove propositions that are true
on the natural numbers, starting at zero. A small
modification of induction can be used to prove statements that are true only for those n k for any
integer k. All that is needed is to use induction on
a proposition Q(n k) where Q(n k) is logically
equivalent to P (n). If Q(n k) is true for n k 0
then P (n) is true for n k and we have the modified
induction. The practical difference is that we start
with k instead of zero.
Example 2.11 Prove that n2 2n for all n 2.
Notice that 22 = 4 = 2 2 so the proposition is true
when n = 2. We next assume that P (n) is true for
some n and we compute:
n2
n2 + 2n + 1
(n + 1)2
(n + 1)2
(n + 1)2
(n + 1)2

2n
2n + 2n + 1
2n + 2n + 1
2n + 1 + 1
2n + 2
2(n + 1)

To move from the third step to the fourth step we

use the fact that 2n > 1 when n 2. The last step
is P (n + 1), which means we have deduced P (n + 1)
from P (n). Using the modified form of induction we
have proved that n2 2n for all n 2.
It is possible to formalize the procedure for using
mathematical induction into a three-part process.
Once we have a proposition P (n),

31

2.2. MATHEMATICAL INDUCTION

(i) First demonstrate a base case by directly demonstrating P (k),
(ii) Next make the induction hypothesis that P (n) is
true for some n,
(iii) Finally, starting with the assumption that P (n)
is true, demonstrate P (n + 1).
These steps permit us to deduce that P (n) is true for
all n k.
Example 2.12 Using induction, prove
1+ 2 + + n =

1
n(n + 1)
2

In this case P (n) is the statement

up lists of numbers. If we wished to sum some formula f (i) over a range from a to b, that is to say
a i b, then we write :
b
X

f (i)

i=a

On the other hand if S is a set of numbers and we

want to add up f (s) for all s S we write:
X
f (s)
sS

The result proved in Example 2.12 may be stated in

the following form using sigma notation.
n
X
i=1

i=

1
n(n + 1)
2

Proposition 2.4 Suppose that c is a constant and

that f (i) and g(i) are formulas. Then
Pb
Pb
Pb
(i)
i=a g(i)
i=a f (i) +
i=a (f (i) + g(i)) =
Pb
Pb
Pb
Base case: 1 = 21 1(1 + 1), so P (1) is true. Induc(ii)
i=a g(i)
i=a f (i)
i=a (f (i) g(i)) =
tion hypothesis: for some n,
Pb
Pb
(iii)
i=a f (i).
i=a c f (i) = c
1
1 + 2 + + n = n(n + 1)
Proof:
2
1
1 + 2 + + n = n(n + 1)
2

Compute:
1 + 2 + + (n + 1) =
=
=
=
=
=

1 + 2 + + n + (n + 1)
1
n(n + 1) + (n + 1)
2
1
(n(n + 1) + 2(n + 1))
2


1 2
n + 3n + 2
2
1
(n + 1)(n + 2)
2
1
(n + 1)((n + 1) + 1)
2

and so we have shown that if P (n) is true then so is

P (n + 1). We have thus proven that P (n) is true for
all n 1 by mathematical induction.

Part (i) and (ii) are both simply the associative

law for addition: a+(b+c) = (a+b)+c applied many
times. Part (iii) is a similar multiple application of
the distributive law ca + cb = c(a + b). 2
The sigma notation lets us work with indefinitely long
(and even infinite) sums. There are other similar notations. If A1 , A2 , . . . , An are sets then the intersection or union of all these sets can be written:
n
\

Ai

i=1
n
[

Ai

i=1

Similarly if f (i) is a formula on the integers then

n
Y

i=1

f (i)

We now introduce sigma notation which makes problems like the one worked in Example
P 2.12 easier to is the notation for computing the product f (1) f (2)
state and manipulate. The symbol
is used to add f (n). This notation is called Pi notation.

32

CHAPTER 2. BASIC SET THEORY

Definition
P 2.14 When we solve an expression
Pinvolving
to obtain a formula that does not use
or
. . . as in Example 2.12 then we say we have found a
closed form for the
Pn expression. Example 2.12 finds
a closed form for i=1 i.

Problem 2.24 Suppose that X Y with |Y | = n

and |X| = m. Compute the number of subsets of Y
that contain X.

Problem 2.25 Compute the following sums.

P20
i=1 i,
At this point we introduce a famous mathematical (i)
P
30
sequence in order to create an arena for practicing (ii)
i=10 i, and
proofs by induction.
P21
(iii)
i=20 i.
Definition 2.15 The Fibonacci numbers are defined as follows. f1 = f2 = 1 and, for n 3, Problem 2.26 Using mathematical induction, prove
the following formulas.
fn = fn1 + fn2 .
Pn
Example 2.13 The Fibonacci numbers with four or (i)
i=1 1 = n,
Pn 2 n(n+1)(2n+1)
fewer digits are: f1 = 1, f2 = 1, f3 = 2, f4 = 3,
, and
i=1 i =
6
f5 = 5, f6 = 8, f7 = 13, f8 = 21, f9 = 34, f10 = 55, (ii)
Pn 3 n2 (n+1)2
f11 = 89, f12 = 144, f13 = 233, f14 = 377, f15 =
.
(iii)
i=1 i =
4
610, f16 = 987, f17 = 1597, f18 = 2584, f19 = 4181,
and f20 = 6765.
Problem 2.27 If f (i) and g(i) are formulas and c
Example 2.14 Prove that the Fibonacci number f3n
is even.
Solution:

and d are constants prove that

b
X
i=a

(c f (i) + d g(i)) = c

b
X
i=a

f (i) + d

b
X

g(i)

i=a

Notice that f3 = 2 and so the proposition is true

when n = 1. Assume that the proposition is true for
some n 1. Then:
f3(n+1)

f3n+3

(2.1)

=
=

f3n+2 + f3n+1
f3n+1 + f3n + f3n+1

(2.2)
(2.3)

(2.4) Problem 2.28 Suppose you want to break an n m

chocolate bar, like the 6 4 example shown above,
but this suffices because f3n is even by the induction into pieces corresponding to the small squares shown.
hypothesis while 2f3n+1 is also even. The sum is thus What is the minimum number of breaks you can
even and so f3(n+1) is even. If follows by induction make? Prove your answer is correct.
that f3n is even for all n. 2
Problem 2.29 Prove by induction that the sum of
the first n odd numbers equals n2 .
=

2 f3n+1 + f3n

Problems
Problem 2.22 Suppose that S = {a, b, c}. Compute
and list explicitly the members of the powerset, P(S).
Problem 2.23 Prove that for a finite set X that
|X| |P(X)|

Problem 2.30 Compute the sum of the first n positive even numbers.
Problem 2.31 Find a closed form for
n
X
i=1

i2 + 3i + 5

33

2.3. FUNCTIONS

Problem 2.32 Let f (n, 3) be the number of subsets Problem 2.41 Consider the statement All cars are
of {1, 2, . . . , n} of size 3. Using induction, prove that the same color. and the following proof .
f (n, 3) = 61 n(n 1)(n 2).
Proof:
Problem 2.33 Suppose
that we
have
sets
X1 , X2 , . . . , Xn and Y1 , Y2 , . . . , Yn such that Xi Yi .
Prove that the intersection of all the Xi is a subset
of the intersection of all the Yi :
n
\

i=1

Xi

n
\

Yi

i=1

Problem 2.34 Suppose that S1 , S2 , . . . Sn are sets.

Prove the following generalization of DeMorgans
laws:
Tn
Sn
c
(i) ( i=1 Si ) = i=1 Sic , and
Tn
Sn
c
(ii) ( i=1 Si ) = i=1 Sic .
Problem 2.35 Prove by induction that the Fibonacci number f4n is a multiple of 3.

Problem 2.36 Prove that if r is a real number r 6= 1

and r 6= 0 then
n
X
i=0

ri =

1 rn+1
1r

Problem 2.37 Prove by induction that the Fibonacci number f5n is a multiple of 5.
Problem 2.38 Prove by induction that the Fibonacci number fn has the value
!n
!n

1+ 5
1 5
5
5

fn =
5
2
5
2
Problem 2.39 Prove that for sufficiently large n the
Fibonacci number fn is the integer closest to
!n

5 1+ 5
5
2
and compute the exact value of f30 . Show your work
(i.e. dont look the result up on the net).
Problem 2.40 Prove that n(n1)(n2)(n3)
is a
24
whole number for any whole number n.

We will prove for n 1 that for any set of n

cars all the cars in the set have the same color.
Base Case: n=1 If there is only one car then
clearly there is only one color the car can be.
Inductive Hypothesis: Assume that for any set of
n cars there is only one color.
Inductive step: Look at any set of n + 1 cars.
Number them: 1, 2, 3, . . . , n, n + 1. Consider the
sets {1, 2, 3, . . . , n} and {2, 3, 4, ..., n + 1}. Each
is a set of only n cars, therefore for each set there
is only one color. But the nth car is in both sets
so the color of the cars in the first set must be the
same as the color of the cars in the second set.
Therefore there must be only one color among all
n + 1 cars.
The proof follows by induction. 2
What are the problems with this proof ?

2.3

Functions

In this section we will define functions and extend

much of our ability to work with sets to infinite sets.
There are a number of different types of functions and
so this section contains a great deal of terminology.
Recall that two finite sets are the same size if they
contain the same number of elements. It is possible
to make this idea formal by using functions and, once
the notion is formally defined, it can be applied to
infinite sets.
Definition 2.16 An ordered pair is a collection of
two elements with the added property that one element comes first and one element comes second. The
set containing only x and y (for x 6= y) is written
{x, y}. The ordered pair containing x and y with x
first is written (x, y). Notice that while {x, x} is not
a well defined set, (x, x) is a well defined ordered pair
because the two copies of x are different by virtue of
coming first and second.

34

CHAPTER 2. BASIC SET THEORY

The reason for defining ordered pairs at this point of ordered pairs {(r2 , r) : r 0}. This function is
is that it permits us to make an important formal well defined because each non-negative real number is
the square of some positive real number.
definition that pervades the rest of mathematics.
Definition 2.17 A function f with domain S and
range T is a set of ordered pairs (s, t) with first element from S and second element from T that has the
property that every element of S appears exactly once
as the first element in some ordered pair. We write
f : S T for such a function.
Example 2.15 Suppose that A = {a, b, c} and B =
{0, 1} then
f = {(a, 0), (b, 1), (c, 0)}
is a function from A to B. The function f : A B
can also be specified by saying f (a) = 0, f (b) = 1 and
f (c) = 0.
The set of ordered pairs {(a, 0), (b, 1)} is not a function from A to B because c is not the first coordinate of any ordered pair. The set of ordered pairs
{(a, 0), (a, 1), (b, 0), (c, 0)} is not a function from A
to B because a appears as the first coordinate of two
different ordered pairs.
In calculus you may have learned the vertical line rule
that states that the graph of a function may not intersect a vertical line at more than one point. This
corresponds to requiring that each point in the domain of the function appear in only one ordered pair.
In set theory, all functions are required to state their
domain and range when they are defined. In calculus
functions had a domain that was a subset of the real
numbers and you were sometimes required to identify
the subset.
Example 2.16 This example contrasts the way
functions were treated in a typical calculus course with
the way we treat them in set theory.
Calculus: find the domain of the function

f (x) = x
Since we know that the square root function exists
only for non-negative real numbers the domain is {x :
x 0}.

Set theory: the function f = x from the nonnegative real numbers to the real numbers is the set

The major contrasts between functions in calculus

and functions in set theory are:
(i) The domain of functions in calculus are often
specified only by implication (you have to know
how all the functions used work) and are almost
always a subset of the real numbers. The domain
in set theory must be explicitly specified and may
be any set at all.
(ii) Functions in calculus typically had graphs that
you could draw and look at. Geometric intuition
driven by the graphs plays a major role in our
understanding of functions. Functions in set theory are seldom graphed and often dont have a
graph.
A point of similarity between calculus and set theory is that the range of the function is not explicitly
specified. When we have a function f : S T then
the range of f is a subset of T .
Definition 2.18 If f is a function then we denote
the domain of f by dom(f ) and the range of f by
rng(f )
Example 2.17 Suppose that f (n) : N N is defined by f (n) = 2n. Then the domain and range of f
are the integers: dom(f ) = rng(f ) = N. If we specify
the ordered pairs of f we get
f = {(n, 2n) : n N}
There are actually two definitions of range that are
used in mathematics. The definition we are using, the
set from which second coordinates of ordered pairs in
a function are drawn, is also the definition typically
using in computer science. The other definition is the
set of second coordinates that actually appear in ordered pairs. This set, which we will define formally
later, is the image of the function. To make matters
even worse the set we are calling the range of a function is also called the co-domain. We include these
confusing terminological notes for students that may
try and look up supplemental material.

35

2.3. FUNCTIONS
Definition 2.19 Let X, Y, and Z be sets. The composition of two functions f : X Y and g : Y Z
is a function h : X Z for which h(x) = g(f (x))
for all x X. We write g f for the composition of
g with f .
The definition of the composition of two functions
requires a little checking to make sure it makes sense.
Since every point must appear as a first coordinate of
an ordered pair in a function, every result of applying
f to an element of X is an element of Y to which g can
be applied. This means that h is a well-defined set of
ordered pairs. Notice that the order of composition is
important - if the sets X, Y , and Z are distinct there
is only one order in which composition even makes
sense.

We use the symbol R for the real numbers. We also

assume familiarity with interval notation for contiguous subsets of the reals. For real numbers a b
(a, b)
(a, b]
[a, b)
[a, b]

is
is
is
is

{x : a < x < b}
{x : a < x b}
{x : a x < b}
{x : a x b}

Example 2.20 The function f : Z Z given by

f (n) = 5 n is a surjection. If we set m = 5 n then
n = 5 m. This means that if we want to find some
n so that f (n) is, for example, 8, then 5 8 = 3
and we see that f (3) = 8. This demonstrates that
all m have some n so that f (n) = m, showing that
all m appear as the second coordinate of an ordered
pair in f .

Example 2.18 Suppose that f : N N is given by

x2
f (n) = 2n while g : N N is given by g(n) = n + 4. The function g : R R given by g(x) = 1+x2 is not
a surjection because 1 < g(x) < 1 for all x R.
Then
(g f )(n) = 2n + 4
Definition 2.22 A function that is both surjective
while
and injective is said to be bijective. Bijective func(f g)(n) = 2(n + 4) = 2n + 8
tions are called bijections.
We now start a series of definitions that divide functions into a number of classes. We will arrive at a
point where we can determine if the mapping of a
function is reversible, if there is a function that exactly reverses the action of a given function.
Definition 2.20 A function f : S T is injective
or one-to-one if no element of T (no second coordinate) appears in more than one ordered pair. Such a
function is called an injection.
Example 2.19 The function f : N N given by
f (n) = 2n is an injection. The ordered pairs of f are
(n, 2n) and so any number that appears as a second
coordinate does so once.
The function g : Z Z given by g(n) = n2 is not
an injection. To see this notice that g contains the
ordered pairs (1, 1) and (1, 1) so that 1 appears twice
as the second coordinate of an ordered pair.
Definition 2.21 A function f : S T is surjective or onto if every element of T appears in an
ordered pair. Surjective functions are called surjections.

Example 2.21 The function f : Z Z given by

f (n) = n is a bijection. All of its ordered pairs have
the same first and second coordinate. This function
is called the identity function.
15

10

-5

-10

-15
-3

-2

-1

The function g : R R given by g(x) = x3 4x is

not a bijection. It is not too hard to show that it is a
surjection, but it fails to be an injection. The portion
of the graph shown above demonstrates that g(x) takes
on the same value more than once. This means that

36

CHAPTER 2. BASIC SET THEORY

some numbers appear twice as second coordinates of Definition 2.24 The inverse of a function f : S
ordered pairs in g. We can use the graph because g is T is a function g : T S so that for all x S,
a function from the real numbers to the real numbers. g(f (x)) = x and for all y T , f (g(y)) = y.
For a function f : S T to be a bijection every
element of S appears in an ordered pair as the first
member of an ordered pair and every element of T
appears in an ordered pair as the second member of
an ordered pair. Another way to view a bijection is as
a matching of the elements of S and T so that every
element of S is paired with an element of T . For
finite sets this is clearly only possible if the sets are
the same size and, in fact, this is the formal definition
of same size for sets.
Definition 2.23 Two sets S and T are defined to be
the same size or to have equal cardinality if there
is a bijection f : S T .
Example 2.22 The sets A = {a, b, c} and Z =
{1, 2, 3} are the same size. This is obvious because
they have the same number of elements, |A| = |Z| = 3
but we can construct an explicit bijection

If a function f has an inverse we use the notation

f 1 for that inverse. Since an exponent of 1 also
means reciprocal in some circumstances this can be
a bit confusing. The notational confusion is resolved
by considering context. So long as we keep firmly in
mind that functions are sets of ordered pairs it is easy
to prove the proposition/definition that follows after
the next example.
Example 2.23 If E is the set of even integers then
the bijection f (n) = 2n from Z to E has the inverse
f 1 : E Z given by g(2n) = n. Notice that defining the rule for g as depending on the argument 2n
seamlessly incorporates the fact that the domain of g
is the even integers.

f = {(a, 3), (b, 1), (c, 2)}

with each member of A appearing once as a first coordinate and each member of B appearing once as a
second coordinate. This bijection is a witness that A
and B are the same size.
Let E be the set of even integers. Then the function
g:ZE
in which g(n) = 2n is a bijection. Notice that each
integers can be put into g and that each even integer
has exactly one integer that can be doubled to make it.
The existence of g is a witness that the set of integers
and the set of even integers are the same size. This
may seem a bit bizarre because the set Z E is the
infinite set of odd integers. In fact one hallmark of an
infinite set is that it can be the same size as a proper
subset. This also means we now have an equality set
for sizes of infinite sets. We will do a good deal more
with this in Chapter 3.
Bijections have another nice property: they can be
unambiguously reversed.

x
If g(x) = x1
, shown above with its asymptotes
x = 1 and y = 1 then f is a function from the
set H = R {1} to itself. The function was chosen to have asymptotes at equal x and y values; this
is a bit unusual. The function g is a bijection. Notice that the graph intersects any horizontal or vertical line in at most one point. Every value except
x = 1 may be put into g meaning that g is a function
on H. Since the vertical asymptote goes off to in
both directions, all values in H come out of g. This
demonstrates g is a bijection. This means that it has
an inverse which we now compute using a standard

37

2.3. FUNCTIONS
technique from calculus classes.
y

y(x 1) =
xy y =
xy x =

x(y 1) =
x =
which tells us that g 1 (x) =
function is its own inverse.

x
x1
x
x
y
y
y
y1
x
x1

so g = g 1 : the

Proposition 2.5 A function has an inverse if and

only if it is a bijection.
Proof:
Suppose that f : S T is a bijection. Then if
g : T S has ordered pairs that are the exact reverse
of those given by f it is obvious that for all x S,
g(f (x)) = x, likewise that for all y T , f (g(y)) = y.
We have that bijections posses inverses. It remains
to show that non-bijections do not have inverses.
If f : S T is not a bijection then either it is
not a surjection or it is not an injection. If f is not a
surjection then there is some t T that appears in no
ordered pair of f . This means that no matter what
g(t) is, f (g(t)) 6= t and we fail to have an inverse.
If, on the other hand, f : S T is a surjection
but fails to be an injection then for some distinct
a, b S we have that f (a) = t = f (b). For g : T S
to be an inverse of f we would need g(t) = a and
g(t) = b, forcing t to appear as the first coordinate
of two ordered pairs in g and so rendering g a nonfunction. We thus have that non-bijections do not
have inverses. 2
The type of inverse we are discussing above is a twosided inverse. The functions f and f 1 are mutually
inverses of one another. It is possible to find a function that is a one-way inverse of a function so that
f (g(x)) = x but g(f (x)) is not even defined. These
are called one-sided inverses.

B is proven by first assuming A and deducing B and

then separately assuming B and deducing A. The
formal symbol for A iff B is A B. Likewise we
have symbols for the ability to deduce B given A,
A B and vice-versa B A. These symbols are
spoken A implies B and B implies A respectively.
Proposition 2.6 Suppose that X, Y , and Z are
sets. If f : X Y and g : Y Z are bijections
then so is g f : X Z.
Proof: this proof is left as an exercise.
Definition 2.25 Suppose that f : A B is a function. The image of A in B is the subset of B made
of elements that appear as the second element of ordered pairs in f . Colloquially the image of f is the set
of elements of B hit by f . We use the notation Im(f )
for images. In other words Im(f ) = {f (a) : a A}.
Example 2.24 If f : N N is given by the rule
f (n) = 3n then the set T = {0, 3, 6, . . .} of natural
numbers that are multiples of three is the image of f .
Notation: Im(f ) = T .
If g : R R given by g(x) = x2 then
Im(g) = {y : y 0, y R}
There is a name for the set of all ordered pairs drawn
from two sets.
Definition 2.26 If A and B are sets then the set of
all ordered pairs with the first element from A and
the second from B is called the Cartesian Product
of A and B.
The notation for the Cartesian product of A and
B is A B. using curly brace notation:
A B = {(a, b) : a A, b B}
Example 2.25 If A = {1, 2} and B = {x, y} then

Note on mathematical grammar: Recall that when

A B = {(1, x), (1, y), (2, x), (2, y)}
two notions, such as bijection and has an inverse
are equivalent we use the phrase if and only if (abbreviated iff) to phrase a proposition declaring that The Cartesian plane is an example of a Cartesian
the notions are equivalent. A proposition that A iff product of the real numbers with themselves: R R.

38

2.3.1

CHAPTER 2. BASIC SET THEORY

Permutations

n
n!

0
1

1
1

2
2

3
6

4
24

5
120

6
720

7
5040

In this section we will look at a very useful sort of

function, bijections of finite sets.
Proposition 2.7 The number of permutations of a
finite set with n elements is n!.
Definition 2.27 A permutation is a bijection of a
finite set with itself. Likewise a bijection of a finite Proof: this proof is left as an exercise.
set X with itself is called a permutation of X.
Notice that one implication of Proposition 2.6 is that
Example 2.26 Let A = {a, b, c} then the possible the composition of two permutations is a permutapermutations of A consist of the following six func- tion. This means that the set of permutations of a
set is closed under functional composition.
tions:
{(a,a)(b,b)(c,c)}

{(a,a)(b,c)(c,b)}

{(a,b)(b,a)(c,c)}

{(a,b)(b,c)(c,a)}

{(a,c)(b,a)(c,b)}

{(a,c)(b,b)(c,a)}

Notice that the number of permutations of three objects does not depend on the identity of those objects.
In fact there are always six permutations of any set of
three objects. We now define a handy function that
uses a rather odd notation. The method of showing permutations in Example 2.26, explicit listing of
ordered pairs, is a bit cumbersome.
Definition 2.28 Assume that we have agreed on an
order, e.g. a,b,c, for the members of a set X =
{a, b, c}. Then one-line notation for a permutation
f consists of listing the first coordinate of the ordered
pairs in the agreed on order. The table in Example
2.26 would become:
abc
bac
cab

Definition 2.30 A fixed point of a function f :

S S is any x S such that f (x) = x. We say that
f fixes x.

Problems
Problem 2.42 Suppose for finite sets A and B that
f : A B is an injective function. Prove that
|B| |A|
Problem 2.43 Suppose that for finite sets A and B
that f : A B is a surjective function. Prove that
|A| |B|.
Problem 2.44 Using functions from the integers to
the integers give an example of
(i) A function that is an injection but not a surjection.
(ii) A function that is a surjection but not an injection.

acb
bca
cba

in one line notation. Notice the saving of space.

(iii) A function that is neither an injection nor a surjection.

Definition 2.29 The factorial of a natural number (iv) A bijection that is not the identity function.
n is the product
Problem 2.45 For each of the following functions
n
Y
from the real numbers to the real numbers say if the
i
n(n 1)(n 2) 3 2 1 =
function is surjective or injective. It may be neither.
i=1

with the convention that the factorial of 0 is 1. We

denote the factorial of n as n!, spoken n factorial.
Example 2.27 Here are the first few factorials:

(i) f (x) = x2 (ii) g(x) = x3


x
x0

(iii) h(x) =
x x < 0

39

2.3. FUNCTIONS

Interlude
The Collatz Conjecture
One of the most interesting features of mathematics is that it is possible to phrase problems
in a few lines that turn out to be incredibly hard. The Collatz conjecture was first posed
in 1937 by Lothar Collatz. Define the function f from the natural numbers to the natural
numbers with the rule

3n + 1 n odd
f (n) =
n
n even
2
Collatz conjecture is that if you apply f repeatedly to a positive integer then the resulting
sequence of numbers eventually arrives at one. If we start with 17, for example, the result
of repeatedly applying f is:
f (17) = 52, f (52) = 26, f (26) = 13, f (13) = 40, f (40) = 20, f (20) = 10,
f (10) = 5, f (5) = 16, f (16) = 8, f (8) = 4, f (4) = 2, f (2) = 1
The sequences of numbers generated by repeatedly applying f to a natural number are
called hailstone sequences with the collapse of the value when a large power of 2 appears
being analogous to the impact of a hailstone. If we start with the number 27 then 111 steps
are required to reach one and the largest intermediate number is 9232. This quite irregular
behavior of the sequence is not at all apparent in the original phrasing of the problem.
The Collatz conjecture has been checked for numbers up to 5 261 (about 5.764 1018 )
by using a variety of computational tricks. It has not, however, been proven or disproven.
The very simple statement of the problem causes mathematicians to underestimate the
difficulty of the problem. At one point a mathematician suggested that the problem might
have been developed by the Russians as a way to slow American mathematical research.
This was after several of his colleagues spent months working on the problem without
obtaining results.
A simple (but incorrect) argument suggests that hailstone sequences ought to grow indefinitely. Half of all numbers are odd, half are even. The function f slightly more than triples
odd numbers and divides even numbers in half. Thus, on average, f increases the value of
numbers. The problem is this: half of all even numbers are multiples of four and so are
divided in half twice. One-quarter of all even numbers are multiples of eight and so get
divided in half three times, and so on. The net effect of factors that are powers of two is
to defeat the simple argument that f grows on average.

40

CHAPTER 2. BASIC SET THEORY

Problem 2.46 True or false (and explain): The Problem 2.59 Suppose that X and Y are finite sets
function f (x) = x1
x+1 is a bijection from the real num- and that |X| = |Y | = n. Prove that there are n!
bijections of X with Y .
bers to the real numbers.
Problem 2.47 Find a function that is an injection
of the integers into the even integers that does not
appear in any of the examples in this chapter.
Problem 2.48 Suppose that B A and that there
exists a bijection f : A B. What may be reasonably
deduced about the set A?

Problem 2.60 Suppose that X and Y are sets with

|X| = n, |Y | = m. Count the number of functions
from X to Y .
Problem 2.61 Suppose that X and Y are sets with
|X| = n, |Y | = m for m > n. Count the number of
injections of X into Y .

Problem 2.49 Suppose that A and B are finite sets.

Prove that |A B| = |A| |B|.

Problem 2.62 For a finite set S with a subset T

prove that the permutations of S that have all members of T as fixed points form a set that is closed
under functional composition.

Problem 2.50 Suppose that we define h : N N as

follows. If n is even then h(n) = n/2 but if n is odd
then h(n) = 3n + 1. Determine if h is a (i) surjection
or (ii) injection.

Problem 2.63 Compute the number of permutations of a set S with n members that fix at least m < n
points.

Problem 2.51 Prove proposition 2.6.

Problem 2.52 Prove or disprove: the composition
of injections is an injection.
Problem 2.53 Prove or disprove: the composition
of surjections is a surjection.
Problem 2.54 Prove proposition 2.7.
Problem 2.55 List all permutations of
X = {1, 2, 3, 4}
using one-line notation.
Problem 2.56 Suppose that X is a set and that f ,
g, and h are permutations of X. Prove that the equation f g = h has a solution g for any given permutations f and h.
Problem 2.57 Examine the permutation f of Q =
{a, b, c, d, e} which is bcaed in one line notation. If
we create the series f, f f, f (f f ), . . . does the
identity function, abcde, ever appear in the series?
If so, what is its first appearance? If not, why not?

Problem 2.64 Using any technique at all, estimate

the fraction of permutations of an n-element set that
have no fixed points. This problem is intended as an
exploration.
Problem 2.65 Let X be a finite set with |X| = n.
Let C = X X. How many subsets of C have the
property that every element of X appears once as a
first coordinate of some ordered pair and once as a
second coordinate of some ordered pair?
P
Problem
Q 2.66 An alternate version of Sigma ( )
and Pi ( ) notation works by using a set as an index.
So if S = {1, 3, 5, 7} then
X
Y
s = 16 and
s = 105
sS

sS

Given all the material so far, give and defend reasonable values for the sum and product of an empty
set.
Problem 2.67 Suppose that f : [0, 1] [0, 1] for
1 < < is given by

( + 1)x
,
x + 1
prove that f is a bijection.
f (x) =

Problem 2.68 Find, to five decimals accuracy:

Problem 2.58 If f is a permutation of a finite set,
Ln(200!)
prove that the sequence f, f f, f (f f ), . . . must
contain repeated elements.
Explain how you obtained the answer.

2.4. + 1

2.4

41

+1

the representation of the next natural number. This

permits us to conclude that the set of all natural numWe conclude the chapter with a brief section that bers
demonstrates a strange thing that can be accom{0, 1, 2, . . .}
plished with set notation. We choose to represent
fits the definition of a natural number. Which natthe natural numbers 0, 1, 2, . . . by sets that contain
ural number is it? It is easy to see, in the minimal
the number of elements counted by the corresponding
set representation, that for natural numbers m and
natural number. We also choose to do so as simply as
n, m < n implies that the representation of m is a
possible, using only curly braces and commas. Given
subset of the representation of n. Every finite natural
this the numbers and their corresponding sets are:
number is a subset of the set of all natural numbers
and so we conclude that {0, 1, 2, . . .} is an infinite
0 : {}
natural number. The set notation thus permits us to
1 : {{}} = {0}
construct an infinite number.
2 : {{},{{}}}={0,1}
3 : {{},{{}},{{},{{}}}}={0,1,2}
The set consisting of the representations of all finite
4 : {{},{{}},{{},{{}}},{{},{{}},{{},{{}}}}}
natural numbers is an infinite natural number. The
={0,1,2,3}
number has been given the name , the lower-case
omega. In addition to being a letter omega tradiThe trick for the above representation is this. Zero is tionally also means the last. The number comes
represented by the empty set. One is represented by after all the finite natural numbers. If we now apply
the set of the only thing we have constructed - zero, Proposition 2.8 we see that
represented as the empty set. Similarly the represen {} = + 1
tation of two is the set of the representation of zero
and one (the empty set and the set of the empty set). This means that we can add one to an infinite numThis representation is incredibly inefficient but it uses ber. Is the resulting number + 1 a different number
a very small number of symbols. This representation from ? It turns out the answer is yes, because
also has a useful property. As always, we will start the representations of these numbers are different as
with a definition.
sets. The representation of contains no infinite sets
while the representation of + 1 contains one.
Definition 2.31 The minimal set representation of the natural numbers is constructed as follows:
(i) Let 0 be represented by the empty set.
(ii) For n > 0 let n be represented by the set
{0, 1, . . . , n 1}.
The shorthand {0, 1} for {{}, {{}}} is called the simplified notation for the minimal set representation.
We now give the useful property of the minimal set
representation.
Proposition 2.8 n + 1 = n {n}
Proof:
This follows directly from Definition 2.31 by considering the set difference of the representations of n and
n 1. 2
The definition says that any set of the representations
of consecutive natural numbers, starting at zero, is

Problems

Problem 2.69 Find the representation for 5 using

the curly-brace-and-comma notation.
Problem 2.70 Give the minimal set representation
of + 2 using the simplified notation.
Problem 2.71 Suppose that n > m are natural
numbers and that S is the minimal set representation of n while T is the minimal set representation of
m. Is the representation of n m a member of the
set difference S T ?
Problem 2.72 Give a formula, as a function of n,
for the number of times that the symbol { appears in
the representation of n.
Problem 2.73 Prove or disprove: there are an infinite number of distinct infinite numbers.

42

CHAPTER 2. BASIC SET THEORY

Interlude
Russells Paradox
Bertrand Arthur William Russell, 3rd Earl
Russell, OM, FRS (18 May 1872-2 February 1970), commonly known as simply
Bertrand Russell, was a British philosopher, logician, mathematician, historian,
religious skeptic, social reformer, socialist
and pacifist. Although he spent the majority of his life in England, he was born in
Wales, where he also died.
Let Q be the set of all sets that do not contain themselves as a member. Consider the
question:Does Q contain itself? If the answer to this question is no then Q, by definition
must contain itself. If, however, Q contains itself then it is by definition unable to contain
itself. This rather annoying contradiction, constructed by Russell, had a rather amusing
side effect.
Friedrich Frege had just finished the second of a three volume set of works called the Basic
Laws of Arithmetic that was supposed to remove all intuition from mathematics and place
it on a purely logical basis. Russell wrote Frege, explaining his paradox. Frege added
an appendix to his second volume that attempted to avoid Russells paradox. The third
volume was never published.
It is possible to resolve Russells paradox by being much more careful about what objects
may be defined to be sets; the category of all sets that do not contain themselves gives rise to
no contradiction (it does give rise to an entire field of mathematics, category theory). The
key to resolving the paradox from a set theoretic perspective is that one cannot assume
that, for every property, there is a set of all things satisfying that property. This is a
reason why it is important that a set is properly defined. Another consequence of Russells
paradox is a warning that self-referential statements are both potentially interesting and
fairly dangerous, at least on the intellectual plane.
The original phrasing of Russells paradox was in terms of normal and abnormal sets. A
set is normal if it fails to contain itself and abnormal otherwise. Consider the set of all
normal sets. If this set is abnormal, it contains itself but by definition the set contains only
normal sets and hence it is itself normal. The normality of this set forces the set to contain
itself, which makes it abnormal. This is simply a rephrasing of the original contradiction.
Puzzle: what does the circuit below have to do with Russells paradox and what use is it?