You are on page 1of 40

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/285819281

Real Number System

Chapter · April 2016

CITATIONS READS

0 12,045

1 author:

Mukta Bhandari
Chowan University
3 PUBLICATIONS   0 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Mukta Bhandari on 06 December 2015.

The user has requested enhancement of the downloaded file.


2

Real Number System

The real number system has strong claims to a central position in mathematics. It is
the point of departure for the vast field of mathematics called "Analysis"[? ]. The Real
Numbers system provides foundation with model and techniques for modern mathematics.
It is the beginning of the beginning of mathematics.This chapter will deal with the
constructive and the axiomatic approach of Real Number System. Anybody interested on
axiomatic approach of Real Numbers System can see the book [? ].

2.1 The Number Line

copyright material
Let L be the line extended indefinitely to both directions from the point "0".The point
labeled "0" is called the origin of the line. The marks 1,2,3,4,5 etc to the right of "0" are
positive integers and -1,-2,-3,-4,-5 etc marked on the left of "0" are negative integers. A
line "L" calliberated this way is called a number line. Every real number corresponds to
a point on the number line and vice versa.

−5 −4 −3 −2 −1 0 1 2 3 4 5 L

Figure 2.1: Number Line

Note that the integers depicted on the number line L above are in increasing order
in value from left to right. The direction from left to right in a number line is called a
positive direction. The set of all real numbers, each of which is represented by a point on
the number line, is denoted by R.

2.1.1 Rational Numbers:


The whole numbers
... − 4, −3, −2, −1, 0, 1, 2, 3, 4, ....
are called integers. The set of integers is denoted by Z. That is
Z = {... − 4, −3, −2, −1, 0, 1, 2, 3, 4, ....}.

27
The whole numbers 1, 2, 3, 4, ... are called the positive integers. These are also know as
Natural numbers and their set is denoted by N or Z+ . That is

N = Z+ = {1, 2, 3, 4, ...}.

The whole numbers −1, −2, −3, −4, ... are called negative integers and their set is denoted
by Z− . That is,
Z− = {−1, −2, −3, −4, ...}.
The integer 0 (zero) is neither positive nor negative. This is also called a neutral number.
Thus, [
Z = Z+ Z− ∪ {0}.
There are numbers other than integers which fill many other (how many?) points on
the number line L. See the following figure:

0 .1 .2 .3 .4 1/2 .6 .7.75.8 .9 1

Figure 2.2: Some Rationals in a Number Line

copyright material
Note that the set of integers Z is not closed under division. If we divide one integer by
other nonzero integer then sometime we get an integer but most of the time we end up
getting a number which is not an integer. For example, 6 ÷ 3 = 2 which is an integer but
6 ÷ 12 = 12 = .5 which is not an integer. We can extract all the integers by this process
and many others which are not integers. Such numbers are called rationals. The set of all
rational numbers is denoted by Q. Thus
p
 
Q= : p, q ∈ Z, q 6= 0 .
q
Note that Z ⊂ Q. That is Z is a proper subset of Q. Also note that the set of all
rational numbers Q is closed under addition, subtraction, multiplication, and division
(provided that it is not divided by 0). This means that adding, subtracting, multiplying,
or dividing (by non-zero integer)any two rational numbers results into a rational number.

2.1.2 Irrational Numbers:


The fact that there are elements in R which are not in Q is not immediately obvious.The

ancient Greek Society of Pythagoreans in the 6th century B.C. proved that the 2, the
diagonal
√ length of a unit square, can not be written as a ratio of integers. In other
words 2 can not be rational number.This also means that square of no rational number
can be 2. The rational numbers are not enough to fill all the dots on the number line.
Indeed,there are uncountably many dots left after the rational numbers occupy the dots
on the number line. In order to motivate this we see the following example:

28

Example 41. 2 is not a rational number.


Proof. Assume, on the contrary, that 2 is a rational number. Then

√ m
2=
n

where m, n ∈ Z, n 6= 0, and m and n have no common factor. This implies that

m2 = 22 .

Then m2 is even and hence m is even. That is m = 2 for some k ∈ Z. This in turn implies
that n2 = 22 forcing n2 and hence n to be even. Thus both m and n√are even. This
contradicts that m and n have no common factor. This forces that 2 can not be a
rational number. This proves our assertion.

copyright material
Remark 42. The above example
√ proves that

2 is not a rational number. This does
not prove the existence of 2 as a real number. We need√ to explore more properties of
real number system in order to prove the existence of 2 as real number. This will be
done in section 2.7.1.

Definition 43. Any number which √ is not rational is called an irrational number.From
the above example, we see that 2 is an irrational number. The other standard examples
of irrational numbers are e, π. There are uncountably many unknown irrational numbers.
We will prove this fact later. There is no standard notation for the set of irrational
numbers, but the notations Q̄, R\Q, or R−Q where bar, minus sign, or backlash indicates
the set complement of the set of rational number Q over the set of real numbers R.


Locating 2 in a Number Line:


Next we see how to locate 2 in a number line. Consider the following figure:

29
D 1 C

√ g
2
1 1

0 1
A B E


Figure 2.3: 2 is an irrational number
copyright material

In the figure above, ABCD is a√unit square. By Pythagorean formula, it follows that
the length of the diagonal
√ AC = 2. Now we draw an arc with center at A (located at
the origin o)and radius√ 2 that meets the number line at E and the unit square at C.
Note that AE = AC = 2 ( AE = AC being the √ radii of the same circle). Thus the point
on the number line represents the the number 2 which is not a rational.

Similarly, 3 can be located in the number line. This is left as an exercise for some
curious reader.

This actually completes the real number system. Thus R = Q Q̄ = Q (R \ Q). Note
S S

that the set of real numbers R is closed under all four basic arithmetic operations except
for the case dividing by zero. In other words, adding, subtracting, multiplying, and
dividing (except by 0) always yields a real number in return. We will present the algebraic
and order properties of R in the next section. Before jumping to that section we present
the following mindmap of the real number system.

30
Irrational
Num-
bers

Real Num-
ber System

Non-
Rational integers
Num-
bers

Pos. Integers
Inte-
ger Neg.
Inte-
copyright
Zeromaterial gers

Figure 2.4: Real Number System

In order to understand more about the real number system we next discuss its various
properties including algebraic, order, and completeness.

Theorem 44. Let r ∈ Q, r 6= 0 and let x be an irrational number. Then,

(i) rx is an irrational number.

(ii) r + x is an irrational number.

m
Proof. Let r = n where m, n ∈ Z, n 6= 0. We prove both part (i) and part (ii) by contra-
diction.
a
(i) Assume that rx is a rational number. That is rx = b where a, b ∈ Z, b 6= 0. Then,

m a
rx = x=
n b
which implies that

31
m a an
x= . =
n b bm
where an, bm ∈ Z, bm =
6 0. This means that x must be a rational and this contradicts
the hypothesis that x is an irrational. This contradiction implies that rx must be
an irrational.
a
(ii) Assume, if possible, that r + x is a rational number. That is r + x = b where
a, b ∈ Z, b 6= 0. Then,

a m an − bm
x= − =
b n bn
where an − bm, bn ∈ Z, bn 6= 0. This means that x must be an irrational which
contradicts the hypothesis. Therefore r + x must be an irrational.

2.2 Mathematical Induction


Mathematical induction is one of the powerful methods of proving the validity of a
mathematical statement for all natural numbers. It is a specialized form of deductive
reasoning used to prove a fact about all the elements of an infinite set by performing a

copyright material
finite number of steps. It is a form of direct proof and is done in two steps. The first
step, also know as base step, is to prove the validity of the statement for the first natural
number. The next step, which is known as the inductive step, is to prove that the validity
of the statement for one natural number induces the validity of the statement to the next
natural number. Then the statement becomes true for all natural numbers by applying
the induction on the base step. Note that the set of natural numbers is

N := {1, 2, 3, . . . } .
One of the special property of the set of natural numbers N, called the Well-Ordering
Principle of N, is required to establish the Principle of Mathematical Induction.

Definition 45. A set of real numbers S is called well-ordered if every nonempty subset
of S has a least element.

Example 46. (i) Any finite set of real numbers is well ordered.

(ii) The well known sets Z, R, Q are not well ordered. Given any element x in these set
we can always find a smeller element x − 1 in the same set.

(iii) Some sets which do have least element are not well-ordered. Consider the set
S = [0, 2]. This set S has the smallest element 0. However, this set is not well
ordered because one of its nonempty subset (0, 2] doesn’t have a least element. For
x
any x ∈ (0, 2] there exists a smaller element ∈ (0, 2].
2
32
Theorem 47 (Well-Ordering Property of N). Every nonempty subset of N has a
least element. In other words, if S is a nonempty subset of N, then there exists m ∈ N
such that m ≤ n for every n ∈ N.

Here is the principle of mathematical induction.

Theorem 48 (Principle of Mathematical Induction:). Let P (n) represents a state-


ment about n ∈ N. Suppose that

1. P (1) is true.

2. If P (k) is true, then so is P (k + 1) for every k ∈ N.

Then P (n) is true for all n ∈ N.

Proof. We prove this by contradiction. Assume that P (1) is true. We want to prove that
P (n) is true for all n ∈ N. Assume, on the contrary, that P (n) is not true for some n ∈ N.
Let n0 be the least such n that P (n0 ) is false. Note that n0 = 6 1 because P (1) is true.
So, n0 is an integer greater than 1. That is there exists a natural number n1 such that
n0 = 1 + n1 . Since n1 < n0 and n1 is the least integer for which P (n0 ) is false, it follows
that P (n1 ) must be true. But then by the induction hypothesis, P (n0 ) = P (n1 + 1) must
be true. This contradicts to our assumption. This implies that it is not possible to have
P (n) false for any n ∈ N. Hence, P (n) is true for all n ∈ N.

copyright material
It is quite natural that a statement be false for the first finite number of natural
number and true for the rest. For example, the statement

P (n) : n! < nn , n ∈ N.
This statement is not true for n = 1 but true for the rest. So, a true statement is

P (n) : n! < nn , n ≥ 2
In order to prove such an statement, we use a modified version of Principle of Mathematical
Induction.
Theorem 49 (Principle of Mathematical Induction (Modified Version)). Let
k ∈ N and let P (n) be a statement for n ≥ k. Suppose that

1. P (k) is true.

2. P (n) is true implies P (n + 1) is true for all n ≥ k.

Then the statement P (n) is true for every n ≥ k.

Proof. The proof of this theorem is left as an exercise.

33
Step 1 of this principle is called the base step, and the assumption in step 2 is called
the induction hypothesis. Step 2 is called the bridge step.
The following examples show how Principle of Mathematical Induction can be applied
to prove an statement for all or all but first finite number of natural numbers.

Example 50. Prove that

1 + 3 + 4 + · · · + (2n − 1) = n2 for every n ∈ N.

Proof. Let
P (n) : 1 + 2 + 3 + · · · + (2n − 1) = n2
denote the statement for n ∈ N. For n = 1, P (1) = 1 = 12 . So P (1) is true. Next suppose
P (k) is true. That is

P (k) : 1 + 2 + 3 + · · · + (2k − 1) = k 2 .
We want to prove that P (k + 1) is true. For n = k + 1,

P (k + 1) : 1 + 2 + 3 + . . . (2k − 1) + (2k + 1) = k 2 + (2k + 1) = (k + 1)2 .

This proves that if p(k) is true then P (k + 1) is true for every k ∈ N. Hence by the
principle of Mathematical Induction, P (n) is true for every n ∈ N. That is,

1 + 3 + 4 + · · · + (2n − 1) = n2
copyright material
for every n ∈ N.

Example 51. Prove that 2n ≥ 2n + 1 for every integer n ≥ 3.

Proof. Let P (n) denote the statement 2n > 2n + 1, n ∈ N. Observe that P (n) is not true
for n = 1, 2 because P (1) : 2 > 2(1) + 1 = 3 and P (2) : 4 = 22 > 2(2) + 1 = 5 are both false.
For n = 3, 8 = 23 > 2(3) + 1 = 7 is true. That is P (3) is true. Next assume that P (k) is
true for n = k. This means that

P (k) : 2k > 2k + 1

is true. We want to establish P (k + 1) is true based on P (k) is true. Now for n = k + 1,

2k+1 = 2.2k > 2(2k + 1) = 2k + 2(k + 1) > 2k + 3 = 2(k + 1) + 1.


Here we applied P (k) and the inequality k + 1 ≥ 2. Thus P (k + 1) is true whenever P (k)
is true. This completes the induction and therefore P (n) : 2n > 2n + 1 is true for all n ≥ 3.

Example 52 (Euler’s Prime Generating Polynomial:). Euler’s Prime Generating


polynomial P (n) = n2 + n + 41 takes on prime values for most of the integers, yet it fails to
be true for all n ∈ N. It is easy to see that P (n) is a prime for 1 ≤ n ≤ 39. But P (40) = 412
which is not a prime.

34
2.3 Finite, Countable and Uncountable Sets
This section is about counting the elements of a set and categorizing it as a finite or infinite
set. Sometimes our counting process terminates after all the elements being exhausted
and sometime the elements are never exhausted and process of counting never terminates.
If the counting process terminates after finite steps then the set is known as a finite set
and it is known as an infinite set if the counting process never terminates. The counting
elements of an infinite set is not as easy and obvious as it sounds to be.

Definition 53. (i) A nonempty set S is said to be finite if there exists a bijection
between the set S and the set {1, 2, 3, . . . , n} for some positive integer n ∈ N. In
this case, the set S is said to have n elements. We say that an empty set ∅ is a
finite set with 0 elements.

(ii) A set is said to be infinite if it is not finite.

A finite set can not be in bijection with any of its proper subset whereas an infinite
set is in bijection with some of its proper subset. This is the main difference between a
finite set and an infinite set.

Example 54. (i) The set S = {3, 5, 7, 8, 11} is a finite set with 5 elements. We can
define a bijection between the sets {1, 2, 3, 4, 5} and the set S. For example, a
bijection f : {1, 2, 3, 4, 5} → S can be defined as f (1) = 3, f (2) = 5, f (3) = 7, f (4) =
8, and f (5) = 11.
copyright material
(ii) The set of positive even integers {2, 4, 6, . . . } is not finite because it can not be in
bijection with any finite subset of the set of natural numbers N. So, this set is infinite.

Definition 55. A set S is said to be countably infinite if there exists a bijection


between the set of natural numbers N = {1, 2, 3, . . . } and the set S. If f is such a bijection
then S can be expressed as {f (1), f (2), f (3), . . . } or {a1 , a2 , a3 , . . . } where ak = f (k) for
k = 1, 2, 3, . . . . A countably infinite set is also known as denumerable.

Example 56. (i) The set of natural numbers N = {1, 2, 3, . . . } is countably infinite. This
is because there exists a bijection f : N → N defined by f (n) = n for every n ∈ N.

(ii) The set of positive integers Z+ , the set of negative integers Z− are uncountably
infinite sets.

Definition 57 (Countable and Uncountable Sets). A set S is said to be countable if


it is either finite or countably infinite. A set which is not countable is called uncountable.

We state the following theorem without a proof.

Theorem 58. Every subset of a countable set is countable.

35
Proof. Let S ⊂ T where T is countable. We want to prove that S is countable. If S is
finite, then it is countable as every finite set is countable. So, suppose S is an infinite
subset of the countable set T . T being countable, there exists a bijection

f :N→T
so that

T = {f (1), f (2), f (3), . . . } = {t1 , t2 , t3 , . . . } .


Next, we define a function

g:N→N
such that

g(1) = min {m : tm ∈ S}

g(2) = min {m : tm ∈ S, m > k(1)}

.. ..
. .
Having chosen g(1) < g(2) < · · · < g(n − 1), we can choose g(n) such that

g(n) = min {m : tm ∈ S, m > g(n − 1)}


copyright material
Consider the composition function f ◦ g : N → S. Note that f ◦ g(n) = f (g(n)) = f (m) =
tm ∈ S where m is smallest such index with m > g(n − 1) by construction. Also note that
g preserves the order. That is m > n ⇒ g(m) > g(n). This means that g is one-to-one.
Note that the function f ◦ g is one to one because

f [g(m)] = f [g(n)] ⇒ g(m) = g(n) ⇒ m = n.


This prove that S is countable.

Theorem 59. Every infinite set has a countably infinite subset.

Proof. Let T be an infinite set not necessarily countable. Otherwise, we are done because
T can be considered as a countable subset of itself. Now we construct a set S as follows:
Select an element in T and name it t1 . We can do this because T has infinitely many points.
Next, we select another element in T different from t1 , and name it t2 . Having chosen
t1 , t2 , t3 , . . . tn−1 , all mutually different, we can select an element in T \ {t1 , t2 , t3 , . . . tn−1 }
and name it tn . Again, this can be done because T \ {t1 , t2 , t3 , . . . tn−1 } 6= ∅ as T is infinite.
Thus, by induction, we can select a point in T and name it tn for every n ∈ N. Let
S = {tn : n ∈ N} where each tn is specially selected element of T . By construction, S
is infinite, S ⊂ T and S is countably infinite because there exists a bijection f : N → S
defined by f (n) = tn for every n ∈ N.

36
The next theorem establishes the equivalent conditions of countability of a set.

Theorem 60. The following conditions are equivalent.

(a) S is countably infinite

(b) There exists a subset A of N such that a map f : A → S is surjective (onto).

(c) There exists a subset B of N and a map g : S → B that is injective (one-to-one)

Proof. (i) (a) ⇒ (b) and (a) ⇒ (c) both follow from the definition of the countability of
the set S. So, we will prove (b) ⇒ (a) and (c) ⇒ (a) in order to complete the proof.

(ii) Here we prove (c) ⇒ (a). Assume that (c) holds. That is, there exists a subset B of
N such that the map
g:S→B
is injective. Then g(S) ⊆ B ⊆ N, and

g : S → g(S)
is bijective. Then S ≈ g(S). Note that g(S) is countable being a subset of the
countable set N. Therefore S is countably infinite.

(iii) Finally, we prove (c) ⇒ (b). Assume that (c) holds. That is, there exists A ⊆ N suct
that a map
copyright material
f :A→S
is surjective. This means that for every s ∈ S, f −1 (s) ⊆ A is a nonempty subset of N.
Each of these subsets has a minimal element, say ns , by Well-Ordering Principle of
N. Let n o
T = ns ∈ N : ns is a minimal element of f −1 (s), s ∈ S .
Then the function f˜ = f T : T → S is one-to-one and is already onto. So, f˜ is a
bijection between T and S. Note that T is countable being a subset of countable set
N. Therefore, S is countably infinite.

Remark 61. The results of the theorem still holds if we replace subsets A and B of N
in the statement by N itself.

Theorem 62. N × N is countably infinite where N is the set of natural numbers.

Proof. Let a function


g : N×N → N
be defined by
g(m, n) = 2m 3n where m, n ∈ N.

37
Then this function is one-to-one. For, if we select (m, n), (m0 , n0 ) ∈ N × N such that
g(m, n) = g(m0 , n0 ) then
0 0
2m .3n = 2m .3n .
This implies that
0 0
2m−m = 3n −n .
This is only possible if m − m0 = 0 and n0 − n = 0. That is, if m = m0 and n = n0 . Therefore
(m, n) = (m0 , n0 ). This proves that the map g is one-to-one. Therefor, by the theorem 60,
the set N × N is countably infinite.

Corollary 1. A finite product of the set of natural numbers N is countable. In general,


a finite product of countable sets is countable.

Theorem 63. Let A = {Ai : i ∈ N} be a family of countable collection of countable sets.


S
Then A = i∈N Ai is countable.

Proof. Let Ai = {ai1 , ai2 , ai3 , . . . } for every i ∈ N. Then


[
A= Ak = {ai,j : (i, j) ∈ N × N} .
copyright material
k∈N

Consider the map

f : A → N×N
defined by
f (ai,j ) = (i, j) for every (i, j) ∈ N × N.
This map is clearly onto. Therefore, by the theorem 60, it follows that the set A is
countably infinite.

Corollary 2. The set of integers Z is countable.

Proof. The result immediately follows because

Z = Z+ ∪ {0} ∪ Z− .

Theorem 64. The set of rational numbers Q is countably infinite (denumerable).

38
( )
p
Proof. Note that the set of rational numbers Q = : p ∈ Z, q ∈ Z \ {0} . We know that
q
a finite product of countable sets is countable (see corollory 1). So, Z × (Z \ {0}) is
countable. Consider a map
F : Z × (Z \ {0}) → Q
defined by
p
F ((p, q)) =
q
for every (p, q) ∈ Z × (Z \ {0}) . This map is clearly onto. Therefore, by the theorem
60, it follows that the set of rational numbers Q is countably infinite (denumerable).

Example 65. The set of all polynomials with integer coefficients is countable.
Proof. Let Pk denote the set of all polynomials of degree k with integer coefficients. If P
denote the set of all polynomials of integer coefficients, then
[
P= Pk .
k∈N

Consider a map

fi : Ai → Zi+1
defined by
copyright i
material
f (p (x)) = (a , a , . . . , a , a )
i i i−1 1 0

where pi (x) = ai xi + ai−1 xi−1 + · · · + a1 x + a0 ∈ Ai for i = 1, 2, 3, . . . . Clearly fi is


injective. We know that a finite product of denumerable sets is denumerable. So Zi+1 , a
finite product of i + 1 copies of the countable set Z, is countably infinite. So, there exists
a bijection

gi+1 : Zi+1 → N.
For each i = 1, 2, 3, . . . , consider a map

hi : Ai → N
defined by

hi (p) = gi+1 ◦ fi (p), p ∈ Ai .


Then hi is injective being a composite of injective functions. This implies that Ai
is countable for every i = 1, 2, 3, . . . . Finally, A is a countable union of countable sets.
Therefore, A is countable.

Example 66. (i) The set of positive even integers or even natural numbers is denu-
merable. Let E = {2, 4, 6, . . . } denote the set of even natural numbers. We can define
a bijection f : N → E where f (n) = 2n for every n ∈ N.

39
(ii) The set O = {1, 3, 5, 7, . . . } of odd natural numbers is denumerable.

Uncountable Sets:
So far we have seen facts about countable sets with examples. There do exist countably
infinite sets like the set of rational numbers Q, the set of natural numbers N etc. By
definition, any set in bijection with a countable set is countable and it is uncountable
otherwise. Here we establish the fact that there do exist uncountable sets.
Definition 67. Any two sets A and B are said to be numerically equivalent if there
exists a bijection between these two sets. We then also say that the sets A and B have
the same cardinality and denote it by A ≈ B.

Definition 68. For any two sets A and B, we say that A is dominated by B, and write
A 4 B, if there exists an injective map f : A → B.

The binary relation 4 satisfies the following properties: For any sets A, B, and C:

(i) A 4 A. A function f : A → A defined by f (x) = x for every x ∈ A defines an injection.


Indeed, this is a bijection.

(ii) A 4 B, B 4 A ⇒ A ≈ B. This follows from the famouse Shroeder Bernstein Theorem.

(iii) A 4 B, B 4 C ⇒ A 4 C. This follows because a composite of injective functions is

copyright material
injective.

Theorem 69 (Cantor’s Theorem). Let A be any set. Then A 4 2A and A  2A .

Proof. Let A be any set. Then we can define a map f : A → 2A by f (a) = {a} , a ∈ A.
This map is clearly injective. For a, b ∈ A,

f (a) = f (b) ⇒ {a} = {b} ⇒ a = b.


Therefore A 4 2A follows. In order to prove that A  2A , now it suffices to prove that
there can not exist a bijection between the sets A and 2A . Suppose, on the contrary, that
there is a bijection g : A → 2A . Define a set

B = {x ∈ A : x ∈
/ g(x)}
This makes sense because g(x) is a subset of A for every x ∈ A. Also, B ⊆ A and therefore
B ∈ 2A . Note that the map g is surjective. So, there exists a ∈ A such that g(a) = B.
Now a ∈ A implies that either a ∈ B or a ∈ / B. Now,

a∈B⇒a∈
/ g(a) = B,
which is a contradiction. Again,

a∈
/ B ⇒ a ∈ f (b) = B

40
which is again a contradiction. Thus there is a contradiction whether a ∈ B or a ∈
/ B. The
source of this contradiction is our assumption of the existence of bijection g. So there
can not exist such a bijection. We already proved that there is an injection. Therefore
A 4 2A but A  2A . This completes the proof.

Definition 70 (Binary Sequence:). A sequence of 0’s and 1’s is called a binary


sequence. For example
(0, 1, 1, 1, 0, 0, 1, 0, . . . )
and,
(1, 1, 0, 1, 0, 0, 1, 0, . . . )
are examples of binary sequences. For any set A, a function χA : N → {0.1} represents
a binary sequence. For example, if A = {2, 4, 6, . . . }, a set of positive even integers, then
χA : N → {0, 1} represents the following binary sequence

(0, 1, 0, 1, 0, 1, . . . ).

Theorem 71. The set of all binary sequences is uncountable.

Proof. Let X denote the set of all binary sequences. Suppose, if possible, X is countable.
Then
copyright material
X = (x1 , x2 , x3 , . . . , xi , . . . )
where
xi = (xi1 , xi2 , xi3 , . . . , xii , . . . ) , i ∈ N
with each xij = 0 or 1. Now, we construct a binary sequence

y = {y1 , y2 , y3 , . . . , yi , . . . }
where

1, if xii = 0;
yi =
0, if xii = 1.

for i = 1, 2, 3, . . . . We observe that y 6= x1 for y1 6= x11 ; y 6= x2 for y2 =


6 x22 , and so on.
In general, y 6= xi because yi 6= xii for i = 1, 2, 3, . . . . Thus the binary sequence y, so
constructed, is not associated to any integer i ∈ N. This contradicts that X is countable.
Therefore, the set X, the set of all binary sequences, must be uncountable. This completes
the proof.

Theorem 72. X ≈ 2N where X is the set of all binary sequences.

41
Proof. Consider a map f : 2N → X defined by

f (A) = χA : N → {0, 1} , A ⊆ N.
f is one-to-one: For A, B ⊆ N,

f (A) = f (B) ⇒ χA = χB ⇒ A = B.
This imples that f is one-to-one.
f is onto: Let x ∈ X be a binary sequence. Let

A = {α ∈ N : xα = 1} .
Then, f (A) = χA = x. For example, if x = (1, 1, 1, 0, 0, . . . ) then we select A = {1, 2, 3}
so that f (A) = χA = (1, 1, 1, . . . ) = x.
Therefore f is a bijection and hence X ≈ 2N .

Corollary 3. 2N is uncountable.

Proof. The proof follows from the theorem 72.

Now we have an example of uncountable set. So, there do exist an uncountable set 2N
where N is the set of natural number which is countable.
copyright material
Theorem 73. The open interval (0, 1) is uncountable.

Proof. We prove this by contradiction. Suppose, if possible, (0, 1) is countably infinite.


Then there exists a bijection f : N → (0, 1). Note that each an = f (n) ∈ (0, 1) has a decimal
expansion

an = 0.an1 an2 an3 . . .


where anj ∈ {0, 1, 2, 3, . . . , 8, 9}. The decimal expansion of an is unique and non-
terminating without repeat if it is irrational. If an is a rational, then its decimal expansion
may terminate or repeat. In case it terminates after finite number of digits after decimal,
then we use 0’s for the rest. We now list the decimal expansion of the numbers in (0, 1) in
the following arrays:

a1 = f (1) = 0.a11 a12 a13 . . . ,


a2 = f (2) = 0.a21 a22 a23 . . . ,
a3 = f (3) = 0.a31 a32 a33 . . . ,
..
.

Now we construct a number b = 0.b1 b2 b3 . . . in (0, 1) such that

42

1, if aii = 3;
bi =
3, 6 3.
if aii =

The choice of digits in the decimal expansion of b makes sure that it has no repeating
9’s making sure that it is not an alternate expansion of a rational number. Note that
b=6 a1 because b1 6= a11 ; b 6= a2 because b2 6= a22 and so on. In general, b 6= ai because
bi 6= aii for every i ∈ N. Thus, we have constructed a number b ∈ (0, 1) which is not in the
list. This contradicts to our assumption that f is a bijection. So there can not exist such
a bijection and therefore (0, 1) must be uncountable.

Theorem 74. Let S ⊆ X and S is uncountable. Then X is also uncountable.

Proof. Suppose, on the contrary, that X is countable. Clearly, X is denumerable. We


know, from the theorem 58, that a subset of a countable set is countable. So, S is
countable. This contradicts to the hypothesis that S is uncountable. This means that X
must be uncountable.

copyright material
Theorem 75. The set of real numbers R is uncountable.

Proof. Note that (0, 1) j R where (0, 1) is uncountable (see theorem 73).Therefore R is
uncountable which follows from the theorem 74.

Remark 76. (i) Any subset of R which contains (0, 1) is uncountable. For example,
[0, 1), (−1, 1), (−2, 3) etc are all uncountable.

(ii) Subset of an uncountable set is not necessarily uncountable. For example, Q ⊆ R


where R is uncountable, but Q is countable. Note that subset of a countable set is
countable.

2.4 The Algebraic Properties of R


Here we first introduce the algebraic properties These properties are also called the field
properties in abstract algebra. Algebraic properties are based on the properties of addition
and multiplication. We use the signs ” + ” and ”.” as a usual addition and multiplication of
real numbers. That is given any two real numbers a, b ∈ R their addition and multiplication
are respectively denoted by a + b and a.b.

43
2.4.1 Algebraic Properties of R :
There are two binary operations addition and multiplication on R denoted by + and .
respectively. These two operations satisfy the following algebraic properties which are
also known as the field axioms:

(i) Commutative Property of Addition: a + b = b + a for every a, b ∈ R.

(ii) Associative Property of Addition: (a + b) + c = a + (b + c) for every a, b, c ∈ R.

(iii) Existence of Additive Identity zero: There exists 0 ∈ R, called zero or additive
identity, such that a + 0 = a for every a ∈ R.

(iv) Existence of Additive Inverse: For every a ∈ R there exists −a ∈ R such that
a + (−a) = 0

(v) Commutative Property of Muktiplication: a.b = b.a for every a, b ∈ R.

(vi) Associative Property of Multiplication: (a.b).c = a.(b.c) for every a, b, c ∈ R.

(vii) Existence of Multiplicative Identity: There exists 1 ∈ R such that a.1 = a for
every a ∈ R.

(viii) Existence of Multiplicative Inverse: For every for every a ∈ R, a 6= 0 there


exists a1 ∈ R such that a. a1 = 1.

(ix) Distributive Property of Multiplication over Addition: a.(b + c) = a.b + a.c


copyright material
for every a, b, c ∈ R.

Note that the properties (i) to (iv) are related to addition, next four properties (v) to
(viii) are related to multiplication. The last property (ix) connects the two operations.
These are exactly the field axioms in abstract algebra. So the set of real numbers R is a
field with respect to the two binary operations addition and multiplication.

Example 77. Here we give some examples of sets which are fields or not a field.
(i) The set of real numbers R is a field.

(ii) The set of rational numbers Q is a field.

(iii) The set of integers Z is not a field.

(iv) The set of natural numbers N is not a field.

2.5 The Order Properties of R:


In addition to satisfying algebraic properties, the set of real numbers R also satisfy the
order properties. The order properties are central in understanding the notion of positivity
of a real number and inequalities.

44
2.5.1 The Order Properties of R:
There exists a nonempty subset P of R satisfying the following properties:

(i) If a, b ∈ P then a + b ∈ P.

(ii) If a, b ∈ P then a.b ∈ P

(iii) If a ∈ R then exactly one of the following holds: a ∈ P, a = 0, −a ∈ P. This is called a


Trichotomy Property because this divides R into three distinct types of numbers:
positive, zero, and negative.

The set P is called the set of positive real numbers. From the order property we
deduct the following properties about positive real numbers:

(i) Property i implies that the sum of two positive real numbers is positive.

(ii) Property ii implies that the product of two positive real numbers is positive.

(iii) Property iii implies that for any real number a ∈ R one and only one of the following
holds:
a is positive, a = 0, −a is positive.

Definition 78. Here we define positive and negative real numbers:

copyright material
(i) Positive Real Number: If a ∈ P then we write a > 0. Such a number a is then
called positive or strictly positive real number.

(ii) Nonnegative Real Number: If a ∈ P ∪ {0} then we write a ≥ 0. Such a number


a is then called a nonnegative real number.

(iii) Negative Real Number: If −a ∈ P then we write a < 0. Such a number a is


then called a negative or strictly negative real number.

(iv) Nonpositive Real Number If −a ∈ P ∪ {0} then we write a ≤ 0. Such a number


a is then called a nonpositive real number.

The following definition is going to introduce the notion of inequality between two
real numbers using the set P of positive real numbers.

Definition 79. Let a, b ∈ R.

(a) a > b or b < a if a − b ∈ P.

(b) a ≥ b or b ≤ a if a − b ∈ P.

A consequence of Trichotomy property and the definition 79 is that for any pair of
real numbers a, b ∈ R one and only of the following inequalities hold:

a > b, a = b, b>a (2.1)

45
Next, we establish some basic rules of inequalities This rules has already been used in
earlier mathematics course, for example, in college algebra.

Theorem 80. Let a, b, c be any real numbers in R.

(a) a < b and b < c ⇒ a < c.

(b) a < b ⇒ a + c < b + c for every c ∈ R.

(c) a > b and c > 0 ⇒ ac > bc.

(d) a > b and c < 0 ⇒ ac < bc.

Proof. The proofs are all based on the definition 79 and the Order Properties in 2.5.

Proof of part (a) :

a < b and b < c ⇒ b − a, c − b ∈ P This further implies that c − a = (b − a) + (c − b) ∈ P.


Therefore, a < c.

Proof of part (b):

(b + c) − (a + c) = b − a ∈ P because a < b. This implies that a + c < b + c. This proves


part (b).

Proof of part (c). copyright material


Note that a − b, c ∈ P because a > b, c > 0 as given. Then ac − bc = (a − b)c ∈ P and
therefore ac > bc.

Proof of part (d).

Note that a − b, −c ∈ P because a > b, c < 0 as given. Then bc − ac = (b − a)c =


(a − b)(−c) ∈ P and therefore ac < bc or bc > ac.

Theorem 81. (a) a2 > 0 if a ∈ R and a 6= 0.

(b) 1>0.

(c) n > 0 for every n ∈ N.

Proof. Proof of part (a):


By Trichotomy Property 2.1 it follows that a > 0 or a < 0. If a > 0 then a ∈ P So,
a2 = a.a ∈ P and therefore a2 > 0.
If a < 0 then −a ∈ P. So, a2 = −a. − a ∈ P and therefore a2 > 0.
Part (b) follows from part (a) because 12 = 1. Proof of part (c) follows from induction.

46
Remark 82. Smallest positive real number doesn’t exist.

1
let a be any positive real number. Then a is a positive real number smaller than
2
1
a. That is 0 < a < a. This means that given any positive real number we can exhibit
2
a positive real number smaller than it. This implies that there is no smallest positive
number.
This results leads to establish the next theorem which has wide application in the
proofs of many results in real analysis.

Theorem 83. If a is a nonnegative real number with a <  for every  > 0, then a = 0.

Proof. Given that a ≥ 0, and a <  for every  > 0. We claim that a = 0. Suppose, if
1 1
possible, a > 0. Take  > 0 so that  = a. Then,  > 0 but  = a < a. This contradicts
2 2
the hypothesis. This proves that a = 0 must be true.

Remark 84. We can further strengthen this result by replacing a <  by a ≤  in its
copyright material
hypothesis. That is, if a ∈ R, 0 ≤ a ≤  for every  ≥ 0, then a = 0. The proof of this
result is left to the readers.

Next we we discuss about positivity of a real number. It is an important tool in


the study of inequalities with real numbers. Note that the product of two positive real
numbers is positive. However, the product being positive doesn’t necessarily mean that
each factor is positive. We see this in the following theorem.

Theorem 85. Let ab > 0. Then either both a and b are positive or both are negative.

Proof. By hypothesis ab > 0 which means that a 6= 0 and b 6= 0. Then by Trichotomy


Property 2.1, it follows that either a > 0 or a < 0 and similarly either b > 0 or b < 0. We
1
first prove that if a > 0, then b > 0. Note that a1 > 0 for a > 0, and so b = ab > 0.
a
Similarly, it can be proved that if a < 0, then b < 0.

Theorem 86. If ab < 0 then either a > 0 and b < 0 or a < 0 and b > 0.

Proof. The proof of this left as an exercise.

47
2.6 Some Useful Inequalities:

Theorem 87 (AM-GM Inequality). Let x and y be any tow positive real numbers.
x+y
Then their Arithmetic Mean (A.M.) is and their Geometric Mean (G.M.)
√ 2
is xy. The Arithmetic-Geometric mean Inequality for x, and y is

√ x+y
xy ≤ .
2
Equality holds if and only if x = y.

√ √ √ √
6 y. Then x > 0, y > 0 and x 6= y.
Proof. Assume that x > 0 and y > 0 with x =
Then
√ √ √
0 < ( x − y)2 = x + y − 2 xy.

Therefore,
√ 1
xy < (x + y).
2

copyright material
Next, assume that x = y with x > 0, y > 0. We may assume that x = y = m. Then

√ 1
xy = m = (x + y).
2

√ 1
Conversely, suppose that xy = (x + y). Then,
2


x + y − 2 xy = 0.
√ √ √ √
This yields ( x − y)2 = 0. That is, x = y and therefore x = y.

Geometrical Interpretation of AM-GM Inequality:

Consider a circle with center at O and diameter x + y with x > y > 0. Select a point
P on the diameter AB such that AP = x and P B = y. Let P Q be drawn perpendicular
to AB at P which meets the circle at Q. Join OQ.

48
Q

A B
O P

Figure 2.5: AM-GM Inequality

copyright material
1 1 1
Then, OQ = radius = (x + y), OP = OB − P B = (x + y) − y = (x − y). Note that
2 2 2
OP Q is a right angle triangle with right angle at P . Using the Pythagorean formula we
get
1 1
P Q2 = OQ2 − OP 2 = (x + y)2 − (x − y)2 = xy.
4 4

That is, P Q = xy. Note that OQ, being the hypotenuse of the right angle triangle
OP Q, is the longest side of the triangle. That is, P Q < OQ which is exactly the AM-GM
inequality with strict inequality. From the figure above, it is obvious that P Q = OQ if
and only if P coincides with O, that is, if and only if AP = P B. Therefore P Q ≤ OQ.
Thus follows the AM-GM inequality.

We take the next inequality as an application of AM-GM inequality.

Example 88. Prove that


xy yz zx
+ + ≥ x+y+z
z x y
for every positive real numbers x, y, and z.

Proof. Using the AM-GM inequality, we note that


xy yz xy yz
r
+ ≥2 . = 2y.
z x z x

49
yz zx zx xy
Similarly, + ≥ 2z and + ≥ 2x. The result follows by adding these three
x y y z
inequalities.

Theorem 89 (Bernoulli’s Inequality). If x > −1, then

(1 + x)n ≥ 1 + nx for every n ∈ N.

Proof. We use the principle of mathematical induction to prove this. Let

P (n) : (1 + x)n ≥ 1 + nx, n ∈ N.

copyright material
Both sides equal 1 + x for n = 1. So P (1) is true. We now assume that P (k) is true
for some k ∈ N. That is,

P (k) : (1 + x)k ≥ 1 + kx for some k ∈ N.

We want to establish that thenP (k + 1) is also true. Now, for n = k + 1,

(1 + x)k+1 = (1 + x)k (1 + x) ≥ (1 + kx)(1 + x) = 1 + (k + 1)x + kx2 ≥ 1 + (k + 1)x

because P (k) is true and kx2 ≥ 0. Thus P (k + 1) is true. Therefore, the Bernoulli’s
inequality holds for all n ∈ N.

Graphical Illustration of Bernoulli’s Inequality for n = 3:

50
13.

12.
f (x) = (1 + x)3
11.

10.

9.

8.

7.

6.

5. g(x) = 1 + 3x
4.

3.

2.
copyright
1.
material
−1. 0 1. 2. 3. 4. 5.
−1.

−2.

Figure 2.6: Bernoulli’s Inequality for n = 3.

2.6.1 Absolute Value and Inequalities

Geometrically, the absolute value of a real number is the distance of the number form the
origin on the number line. This means that absolute value of 5 is 5 and that of −5 is also
5. Absolute value of 0 is 0. Absolute value of a real number x is denoted by |x|.

51
| − 5| = 5 |5| = 5

−5 −4 −3 −2 −1 0 1 2 3 4 5 L

Figure 2.7: Absolute value

From this geometrical interpretation, it is obvious that absolute value of a real number
is non-negative. Thus the absolute value of a real number x, denoted by |x| is defined as
follows:


 x
 if x > 0,
|x| = 0 if x = 0,
−x if


x < 0.

The graph of the absolute value function f (x) = |x| is


copyright material
f

5.
Graph of f(x)=|x|
4.

3.

2.

1.

−5. −4. −3. −2. −1. 0 1. 2. 3. 4. 5.


−1.

Figure 2.8: Graph of the absolute value function f (x) = |x|.

Next we discuss the properties of the absolute value of a real number. These properties
are very important in the study of inequalities.

52
Theorem 90. (a) |x| ≥ 0 for every x ∈ R.

(b) |x| = 0 if and only if x = 0.

(c) |x| = | − x| for every x ∈ R.

(d) |x − y| = |y − x| for every x, y ∈ R.

(e) |xy| = |x||y| for every x, y ∈ R.


1 1

(f) = , x 6= 0.
x |x|

x |x|
(g) = , y 6= 0.
y |y|

(h) |x2 | = x2 for every x ∈ R.

(i) |x| ≤ c if and only if −c ≤ x ≤ c for every x ∈ R where c ≥ 0.

(j) |x| ≥ c if and only if x ≥ c or x ≤ −c where c is any positive constant.

(k) −|x| ≤ x ≤ |x| for every x ∈ R.

(l) For any m > 0, |x| = m is equivalent to x = ±m.

copyright material
Proof. (a) This follows directly from the definition.

(b) This follows from the definition.

(c) This also follows from the definition.

(d) If x − y = m then y − x = −m. So, by part (c) , it follows that

|x − y| = |m| = | − m| = |y − x|.

(e) Suppose xy = 0. Then either x = 0 or y = 0. So, part (a) follows obviously in this case.
Next suppose that xy > 0. Then |xy| = xy = |x||y| if both positive or |xy| = −x. − y =
|x||y| if both negative. If xy < 0, then either x > 0, y < 0 or x < 0, y > 0. If x > 0, y < 0
then |xy| = −xy = x.(−y) = |x||y|. If x < 0, y > 0 then |xy| = −x.y = |x||y|. Thus in
either case, |xy| = |x||y|.
1 1 1 1 1 1 1

(f) Suppose x > 0. Then > 0. So, = = . If x < 0 then < 0, and = =
x x x |x| x x −x
1
.
|x|
(g) This follows form e and f.

(h) Note that x2 ≥ 0. So x2 = |x2 | = |x||x| = |x|2 .

53
(i) Suppose |x| ≤ c. Suppose x ≥ 0. Then |x| = x and x ≤ c. Also, −c ≤ x because x > 0
and c > 0. Thus −c ≤ x ≤ c.
On the other hand, if x ≤ 0, |x| = −x and −x ≤ c, that is, x ≥ −c. Because x ≤ 0, we
have x ≤ c. In both case −c ≤ x ≤ c, and the direct implication follows.
Conversely, suppose that −c ≤ x ≤ c. Then x ≥ −c which means −x ≤ c and x ≤ c.
So, |x| ≤ c.

(j) Suppose x ≥ 0. Then |x| = x, and so x ≥ c. On the other hand, if x < 0, then |x| = −x,
and −x ≥ c, i.e., x ≤ −c.

(k) This follows from part (c) by taking c = |x|.

(l) If x ≥ 0 then |x| = x = m and if x < 0 then |x| = −x = m. Combining these we get
x = ±m.

Next, we establish one of the most useful inequalities which is frequently used in the
proofs of many results in real analysis and many other field of mathematics. This is called
a triangle inequality. This inequality is motivated by the geometrical fact that sum of the
lengths of any two sides of a triangle is always greater than the length of its remaining
third side. This fact justifies its name as triangle inequality.

Theorem 91 (Triangle Inequality).

copyright material
|x + y| ≤ |x| + |y| for every x, y ∈ R.

Proof. From k of the Theorem 90, it follows that

−|x| ≤ x ≤ |x| and − |y| ≤ y ≤ |y|.


Adding these two inequalities, we get

−(|x| + |y|) ≤ x + y ≤ (|x| + |y|).


Therefore, the Triangle Inequality follows from the part (i) of the Theorem (90).
There are other variations of this inequality. These are given as follows: For every
x, y ∈ R

||x| − |y|| ≤ |x − y|,

|x − y| ≤ |x| + |y|.
The first of these two inequalities follows form the triangle inequality (91) and part (c)
of the Theorem (90) together with

|x| = |x − y + y| ≤ |x − y| + |y|, i.e. |x| − |y| ≤ |x − y|,

and,

54
|y| = |y − x + x| ≤ |x − y| + |x|, i.e. − (|x − y|) ≤ |x| − |y|.

The second inequality follows simply from the triangle inequality (91).

2.7
copyright material
The Completeness Axioms of System of Real Num-
bers R:

We have discussed the nine axioms of the Algebraic Properties, also known as the field
axioms, and the three axioms of the Order Properties of the system of real numbers
in the last two sections. That is we have seen that set of real numbers is an ordered
field. The set of rational numbers Q also satisfies
√ these conditions. We have √ seen in
example 41 that there exists a real number 2 which is not rational. That is 2 ∈ / Q.
This observation initiates the requirement of an additional property to characterize the
system of real numbers.This additional requirement is called the Completeness Axiom.
The completeness axioms allows us to introduce the irrational numbers in the system of
real numbers. It further gives a continuity property in the system of real numbers bridging
irrationals through rationals. This is why the completeness property is also known as
continuum property of real numbers. The property of continuity through completeness
axiom is fundamental in the development of real analysis.

We will first introduce some of the terminologies and notions with some examples
before formally introducing the Completeness Axiomx.

55
Definition 92 (Upper and Lower Bounds, Supremum and Infimum). Let S
denote a nonempty set of real numbers.

(i) A real number α is said to be an upper bound of the set S if x ≤ α for every
x ∈ S. The set S is said to be bounded above if it has an upper bound. Upper
bound of a set is not necessarily unique if it exists. If the upper bound α is also
a member of the set S then α is called the largest element or the maximum
element of the set S. We write

α = max S

if it exists.

(ii) A real number α is said to be a lowper bound of the set S if x ≥ α for every
x ∈ S. The set S is said to be bounded below if it has a lower bound. Lower
bound of a set is not necessarily unique if it exists. If the lower bound α is also
a member of the set S then α is called the smallest element or the minimum
element of the set S. We write

α = min S

if it exists.

(iii) A real number α is called the least upper bound (lub) or Supremum (sup)
of S, if
copyright material
(a) α is an upper bound of S, and
(b) there does not exist an upper bound for S strictly smaller than α. That is, if
β is an upper bound of S then α ≤ β. In other words, for every  > 0 there
exists x ∈ S such that x > α − . The supremum, if it exists, is unique, and is
denoted by sup S .

(iv) A real number α is called the Greatest Lower Bound (glb) of S, if

(a) α is a lower bound of S, and


(b) there does not exist a lower bound strictly greater tha α. That is β is a lower
bound of S then α ≥ β. In other word, for every  >, there exists x ∈ S such
that x < α + . The infimum, if it exists, is unique, and is denoted by inf S.

56
Remark 93. Here we describe some of the fundamental differences between supremum
and maximum element of a set.

(i) maximum element of a set is also the supremum of the set. But the converse is
not true.

(ii) A finite set always has a maximum element which is also its supremum. But an
infinite set may have supremum but not a maximum element.

(iii) maximum element of a set is a member of the set but the supremum of a set need
not be an element of the set.

We iluustrate these differences next by examples.

Example 94. The following examples illustrate the differences or similarity between
the max S and the sup S.
1
 
(i) Let S = 2 − : n ∈ N , then sup S = 2, but 2 ∈
/ S. So max S does not exist.
n
1
 
(ii) Let S = : n ∈ N . Then sup S = 1, and inf S = 0.
n
n o
(iii) Let S = r ∈ Q : r2 < 2 . Then S does not have a maximum element. Note that the
√ √

sup S = 2. copyright material


candidate√for the max S is 2 but 2 ∈/ Q. But this set has its supremum, and the

(iv) The half-open interval S = [0, 3) is bounded above. The upper bounds of this set are
any real number greater than or equal to 3. Note that none of its upper bound belongs
to the set S = [0, 3). So, this set has no maximum element. Note that sup S = 3.

(v) Let A = {r ∈ Q : r < 0} and B = {r ∈ Q : r ≥ 0}. Then,

sup A = sup B = 0,

and 0 ∈
/ A, 0 ∈ B.

Verification of these examples are left to the reader as an easy exercise.

From the definition and the examples, we observe that maximum or minimum element,
if it exists, is not unique. But this is not the case with supremum or the infum of a set
whenever they exist.

Theorem 95. let S be a noempty bounded set of real numbers. Then,

(i) The supremum or infimum of a set is unique if it exists.

(ii) inf S ≤ sup S, if both exists.

57
Proof. Both sup S and the inf S exist because S is a bounded set.

(i) Assume that α and β are suprema of the set S. Then α ≤ β because β = sup S and
α is an upper bound of S. Similar argument implies that β ≤ α. Therefore, α = β.
This proves the uniqueness of the supremum of the set S. Similarly, we can prove
that infimum of a set is unique whenever it exists.

(ii) For any element x ∈ S, it follows from the definition of infimum and supremum
that inf S ≤ x and x ≤ sup S. That is, inf S ≤ x ≤ sup S. Therefore, it follows that
inf S ≤ sup S.

Theorem 96. Let S ⊂ R and c ∈ R. Then we define

cS = {cx : x ∈ S} .

(i) If c ≥ 0, then
sup(cS) = c sup(S), inf(cS) = c inf(S).

(ii) If c < 0, then


sup(cS) = c inf(S). inf(cS) = c sup(S).
In particular, for c = −1 this implies that

copyright
sup(−S) = − inf(S) material
and inf(−S) = − sup(S).

This is know as Reflection Trick.

Proof. (i) The result is obvious if c = 0. Suppose c > 0. Let α = sup(cS). Then,

α = sup(cS) ⇒ α ≥ cx for every x ∈ S.


α
⇒ ≥ x for every x ∈ S.
c
α
⇒ ≥ sup(S) by the definition of the supremum.
c
⇒ α ≥ c sup(S).

Therefore,
sup(cS) ≥ c sup(S) (2.2)
By definition, c sup(S) ≥ cx for every x ∈ S. This also means that c sup(s) ≥ sup(cS).
Combining this inequality with the inequality in 2.2, we can conclude that

sup(cS) = c sup(S)
for c ≥ 0. Similarly, we can prove that inf(cS) = c inf(S) for c ≥ 0.

58
(ii) Let c < 0. For x ∈ S,

sup(cS)
sup(cS) ≥ cx ⇒ ≤ x because c < 0.
c
sup(cS)
⇒ ≤ inf(S), by the definition of infimum of a set.
c

Therefore,
sup(cS) ≥ c inf(S). (2.3)

On the other hand, note that inf(S) ≤ x, x ∈ S. Multiplying this inequality by c < 0
we get,

c inf(S) ≥ cx, x ∈ S.

So, c inf(S) ≥ sup(cS). Combining this with the inequality in 2.3, we obtain

sup(cS) = c inf(S) for c < 0.

Similarly, we can prove that inf(cS) = c sup(S).

copyright material
Theorem 97. Let S and T be nonempty subsets of R with S ⊆ T . Show that

(a) inf(T ) ≤ inf(S).

(b) inf(S) ≤ sup(S).

(c) sup(S) ≤ sup(T ).

Proof. (a) Let α = inf(T ). Then α ≤ x for alll x ∈ T . So, α ≤ x for all x ∈ S because S ⊆ T.
That is α is a lower bound of the set S. Consequently, α ≤ inf(S). Therefore,inf(T ) ≤
inf(S).

(b) By the definition of infimum and supremum of a set,

inf(S) ≤ x, and x ≤ sup(S)


for every x ∈ S. Note that, we can find x ∈ S because S is nonempty. That is,

inf(S) ≤ x ≤ sup(S) for every x ∈ S.


Therefore,

inf(S) ≤ sup(S).

59
(c) Let β = sup(T ). Then β ≥ x for every x ∈ T . Then, β ≥ x for every x ∈ S because
S ⊆ T . That ism, β is an upper bound of the set S. Consequently, β ≥ sup(S).
Therefore,
sup(S) ≤ sup(T ).
This completes the proof.

Remark 98. Combining all the three inequalities in Theorem 97, we get

inf(T ) ≤ inf(S) ≤ sup(S) ≤ sup(T ) (2.4)

where S and T are nonempty subsets of R with S ⊆ T .

Gaps in rational Number System:



We know from example 41 that 2 is not a rational number. Let us consider the sets
n o
A = p ∈ Q : p > 0, p2 < 2 ,
and n o
B = p ∈ Q : p > 0, p2 > 2 .
We claim that A contains no largest element and B contains no smallest element. In
other words, we prove that for every p ∈ A we can find q ∈ A such that p < q. And for
copyright material
every p ∈ B we can find q ∈ B such that q < p.

For a rational number p > 0, let

p2 − 2 2p + 2
q = p− = . (2.5)
p+2 p+2
Then,
2(p2 − 2)
q2 − 2 = . (2.6)
(p + 2)2
Let p ∈ A. Then p2 − 2 < 0. Then from the equation 2.5 it follows that q > p > 0. From
the equation 2.6 it follows that q 2 < 2 and hence q ∈ A. Thus for every p ∈ A there exists
q ∈ A such that q > p. That is, the set A contains no maximum element.

Next, let p ∈ B. Then p2 > 2, and from the equation 2.5, we see that q > 0 and q < p.
From the equation 2.6, it follows that q 2 − 2 < 0. Therefore, q ∈ B. Thus for every p ∈ B
there exists q ∈ B such that q < p. This means that the set B contains no minimum
element.

Note that the set A is bounded above. In fact, the set B is the set of all upper bounds
of A. A has no supremum or least upper bound in Q because B has no smallest member.
On the other hand, the set B is bounded below. In fact, A {r ∈ Q : r ≤ 0} is the set of
S

all lower bounds of the set B. But, B contains no largest or maximum element. Therefore,
B has no infimum.

60
Thus A is a set of rational numbers which is bounded above without supremum in Q,
and B is a set of rational numbers which is bounded below without the greatest lower
bound or infimum in Q. This is exactly the gap in Q. This gap is the fundamental
difference between the set of real numbers R and the set of rational numbers Q.

Remark 99. The example above shows that the system of rational numbers Q has
certain gaps despite the fact that between any two rationals there always exists a rational.
For example, if m, n are two rationals with m < n then there exists a rational 12 (m + n)
1
such that m < (m + n) < n. Fulfillment of this gap is exactly the completeness axiom
2
defined below. The set of real numbers satisfies this axiom and fills the gap. This is the
fundamental difference between the set of rational numbers and the set of real numbers.

2.7.1 The Completeness Axiom:


(i) Every nonempty set S of real numbers which is bounded above has a supremum.
That is, there exists a real number α such that α = sup S.
(ii) Every nonempty set S of real numbers which is bounded below has a infimum. That
is, there exists a real number β such that β = inf S.
This axiom actually follows from the axiom i above. This is left as an exercise to
the reader.


Exixtence of 2: copyright

Next we prove the existence of
material
2 in R as one of the applications of the completeness
axiom of R.
n o √
Theorem 100. Let A = x ∈ R : x2 < 2 . Then, sup(A) = 2.

Proof. Note that A 6= φ because 1 ∈ A as 12 = 1 < 2. For any x ∈ A, x2 < 2 < 9. That is
x ≤ 3 for every x ∈ A. Therefore A is bounded above by 3. Then, by the completeness
axiom,
√ there exists a real number α such that α = sup(A). In order to prove the exixtence
of 2 as a real number we claim that α2 = 2.
2
Case I: Suppose α2 < 2. Define β = . Then αβ = 2 and
α
α2 2
< ⇒ α < β.
α α
1
Let y = (α + β). Then α < y < β. Using the AM-GM inequality 87, we get
2
1
y 2 = (α + β)2 > αβ = 2.
4
!2
2 4 4 2 2 2
Then, = 2 < = 2. This implies that ∈ A. But > = α. This contradicts
y y 2 y y β
2
that α = sup(A). This implies that α < 2 is impossible.

61
2
case II: Suppose that α2 > 2. Let β = . Then αβ = 2, and
α
α2 2
> = β ⇒ β < α.
α α
1
Let y = (α + β). Then β < y < α. AM-GM inequalti implies that y 2 > αβ = 2. Now,
2
for any x ∈ A, x2 < 2 < y 2 . This implies that x < y. That is, y is an upper bound of A which
again contradicts that α = sup(A). So, α2 > 2 is impossible. Therefore, by Trichotomy
property 2.1, the only option left is α2 = 2. This completes the proof of the theorem.

2.7.2 Properties of Supremum


Some of the fundamental properties of the supremum of a set are discussed in this section.

Theorem 101 (Approximation Property of the Supremum). Let S be a


nonempty set of real numbers which is bouded above and let α = sup(S). Then for
every  > 0 there is some x ∈ S such that

α −  < x ≤ α.

Proof. Because α = sup(S), it follows that x ≤ α for every x ∈ S. For any given  >
0, α −  < α. In other words, α −  is not an upper bound of the set S. This means that

copyright material
there exists at least one x ∈ S such that α −  < x. Otherwise, α −  would be an upper
bound of S contradicting the hypothesis that α = sup(S). Thus α −  < x ≤ α. This
completes the proof.

Theorem 102 (Additive Property of the Supremum). Let A and B be any two
nonempty subsets of R. If each of A and B has a supremum, the A + B has a supremum
and
sup(A + B) = sup(A) + sup(B)
where A + B = {x + y : x ∈ A, y ∈ B}.

Proof. Let α = sup(A) and β = sup(B). Then x ≤ α for every x ∈ A, and β ≤ y for every
y ∈ B. Let z ∈ A + B. The z = x + y for some x ∈ A and y ∈ B. Then z = x + y ≤ α + β.
This is true for any z ∈ A + B. Therefore, the set A + B is bounded above by α + β. Note
that A + B is nonempty because so are the sets A and B. Therefore, by the completeness
axiom, sup(A + B) exists. Let γ = sup(A + B). As noted, α + β is an upper bound of the
set A + B. So,

γ ≤ α + β.
We next show the reverse inequality α + β ≤ γ. For this, let  > 0 be given. Then by the
theorem 101, there exists elements x ∈ A and y ∈ B such that

α −  < x and β −  < y.

62
Note that x + y ∈ A + B and so x + y ≤ γ. By adding these two inequalities, we get

α + β − 2 < x + y ≤ γ.
This is true for every  > 0. Therefore,

α + β ≤ γ.
Therefore, γ = α + β as desired.

Theorem 103. The set of integers N = {1, 2, 3, . . . } is unbounded from above.

Proof. Suppose, if possible, N is bounded above. Clearly, N is nonempty. The completeness


axiom yields the existence of a real number α such that α = sup(N). Then by the
approximation property of the supremum ( see Theorem 101), there exists n ∈ N such that
α − 1 < n. That is α < n + 1 for this n where n + 1 ∈ N. This contradicts that α = sup(N).
This contradiction implies that the set of integers N can not be bounded above and hence
is unbounded above.

Theorem 104. For every x ∈ R, there exists n ∈ N such that n > x.

Proof. Suppose, on the contrary, that there exists x ∈ R such that n ≤ x for every n ∈ N.

copyright material
This means that this particular x ∈ R is an upper bound of the set N which contradicts
that the set of positive integers N is unbounded above ( see Theorem 103). This completes
the proof of the statement.

Corollary 4. For every x ∈ R, there exists m ∈ Z such that m < x.

Proof. Let x ∈ R. Consider its additive inverse −x ∈ R. Then, for this −x ∈ R, by


the theorem 104, there exists n ∈ N such that −x < n. Then −n < x. We may take
m = −n ∈ Z. Note that m is not necessarily in N but it is certainly in Z. Thus there
exists m ∈ Z such that m < x.

Theorem 105. Let x, y ∈ R with x + 1 < y. Then there exists m ∈ Z such that x < m < y.

Proof. Assume that x, y ∈ R and x + 1 < y. Consider a set

S = {k ∈ Z : x < k} .
By the theorem 104, for every x ∈ R, there exists n ∈ N ⊂ Z such that x < n. So A 6= ∅.
By the corollory 4 of the theorem 104, there exists p ∈ Z such that p < x. Then p < a < k
for every k ∈ A. That is, the set A has a lower bound p ∈ Z. Note that A is a subset
of Z by construction. Then by the Well-Ordering Principle of Z, A has the minimum
element m. Then m ∈ A which means that x < m by the definition of the set A. On the

63
other hand, m − 1 < m. That is m − 1 is smaller than the minimum element of the set
A. So, m − 1 ∈
/ A which means that m − 1 ≤ x. That is, m ≤ x + 1, and by assumption,
x + 1 < y. Now, combining x < m, m ≤ x + 1 and x + 1 < y we get x < m < y where m ∈ Z.
This completes the proof.

Theorem 106 (The Archimedian Propery of R). If x, y ∈ R and x > 0, then there
exists a positive integer n such that nx > y.

y
Proof. Consider the real number z = . Then by the theorem 104, there exists a positive
x
integer n such that z < n. This implies that y > nx because x > 0.

Corollary 5. For every  > 0, however small, there exists a positive integer n such that
1
0 < < .
n

Proof. The Archimedian Principle ( theorem 106) for y = 1 and x =  > 0 yields there
1
exists a positive integer n such that n > 1 which is equivalent to 0 < < .
n

Corollary 6. For every real number x > 0, there exists an integer n ∈ N such that

copyright material
n − 1 ≤ x < n.

Proof. For a real number x > 0, define the set Sx = {m ∈ N : x < m}. From the Archime-
dian Propert of R ( Theorem 106), we can find an integer m ∈ N such that x < m. So, Sx
is nonempty. That is, Sx is a nonempty subset of the set of natural numbers N. Therefore,
by the well ordering principle of the set of natural numbers N (Theorem 47), there exists
the least element of Sx . That is there exists a positive integer n in Sx such that n ≤ m for
every m ∈ Sx . This also means that x < n. Note that n − 1 ∈ / Sx . So, n − 1 ≤ x. Therefore,
n − 1 ≤ x < n.

We know that between two rationals there are infinitely many rationals. One way to
see this is to observe that the average of any two rationals is a rational number. How
about between two reals which are not necessarily rationals? We see the answer of this in
the following theorem:

Theorem 107 (The Rational Density Theorem). There always exists a rational
between any two reals.

Proof. Let x and y be any two real numbers with x < y. Then y − x > 0, and by the
Archimedian Principle (see corollory 6 of theorem 106), there exists n ∈ N such that
1
< y − x. That is, nx + 1 < ny. Then by the theorem 105, there exists m ∈ Z such that
n
nx < m < ny.

64
Dividing this inequality by n ∈ N we obtain
m
x< < y.
n
m
Then r = ∈ Q and x < r < y. This completes the proof.
n
By repeating the argument, we can conclude that there infact are infinitely many
rationals between any two reals. On the other side of the coin, it is interesting to note
that there are irrationals at least as many as rationals. The following theorem establishes
this density property of irrationals in the set of real numbers R.

Theorem 108 (Irrational Density Theorem). There always exists an irrational


between any two reals.

Proof. Let a, b ∈ R with a < b. We wish to find an irrational number q such that a < q < b.
a b a b
Consider the two real numbers √ and √ . Then √ < √ . We know from the Rational
2 2 2 2
a
Density Theorem 107 that there are infinitely many rationals between the reals √ and
2
b
√ . So, we can find a non-zero real r such that
2
a b
√ < r < √ , r 6= 0.
2 2

copyright
√ material
This is equivalent to

a< 2r < b,

where 2r is an irrational (why?). This completes the proof.

By repeated argument, we can conclude that there are infact infinitely many irrational
between any two reals.

65

View publication stats

You might also like