A Basic Course in Real Analysis

Prof. P. D. Srivastava
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture - 4
Continuum and Exercises (Contd.)

So, last time we were discussing about some example showing that a square root of a rational
number positive rational number which is not a perfect square is a irrational number; similarly
square root of a integer positive integer, it is also not a perfect square, it will be a irrational
number. And in fact we have seen the first part is ok, second part we were just checking, and we
come to that.

(Refer Slide Time: 00:54)

Suppose m is a positive integer which is not a perfect square, is it not? That is square root of m
then we will show m is an irrational number, is it not, that we were discussing. So, we have

proved this by contradiction. Suppose it is an irrational number suppose m is an irrational

p
number is a rational number, sorry, is a rational number which is of the form, say , where p
q
p
and q where is in its lowest form and q is not equal to 0. Then no common factor between p
q
 p
and q, and it is in the form, okay. So, what we get it and this number is a rational number. So,
q
we can identify the two integer in between this number will lie. So, let  ,   1 be the two
p
integers such that      1 . That is what we get is  q  p  (  1)q .
q

Now from here we can say that p   q  q let it be 1, because this will be required; that is what
we want, this will be required, okay. Again it is given that root m we are assuming it is a rational
p p2
number which is the form . So, 2  m ; that is p square minus q square equal to mq square,
q q

p 2  mq 2 or p 2  mq 2 will be zero. So, this is also second. Now let us consider the expression

(mq   p)2 , consider this, minus m( p   q ) , oaky; let us consider this. So, basically what we

are doing that this expression if I open it then it can be put it in the form of ( 2  m)( p 2  mq 2 ) ;
one can easily just verify it.

Now p 2  mq 2 because of the two so basically this will come out to the 0 because of two, okay.
Therefore from here m can be written as, m will be written as; oh this is square of this sorry this
is square of; otherwise this problem will be, mq minus lambda this is square will come. So, m
(mq   p) 2
will come out to be what? or this square this square. So, m will be this; so, m
( p   q)2
(mq   p)
will be of the form , just taking positive well roots okay. Now p   q  q from one.
( p   q)
So, from one what it shows that this shows but from one shows that m square root of m that is m
can be written as in the form of the, yes in another form is square root of m is another rest form
where the numerator is this and denominator is this with denominator is lower than the
denominator which we have assumed q.

p
But what we assume is in its lowest form, so a contradiction is reached, because m under root
q
p
m is written in the form of which in lowest form, but what we are getting under root m can
q
also be expressed into this form where the denominator is lower than the q less than it. It means
 p
again the denominator is lower. So, it cannot be lowest form; that cannot be lowest form. So,
q
it is contradiction. So, it shows that p   q  q that is denominator is strictly less than q. So, a
p
contradiction that is in lowest form, and this contradiction is because our wrong assumption it
q

is a rational one; therefore, m is an irrational number, okay. So, that is what we get it, clear. So,
this much we get.

(Refer Slide Time: 07:43)

Now let us come to what is the continuum. The aggregate of the all real numbers that is rational
and irrational is called a continuum. So, basically the set of all real numbers aggregate of the all
real numbers rational irrational we call it a continuum. Now this continuum is complete; this is a
complete set. This continuum is closed that is completeness is there; what do you mean by this?
We have got this real discontinuum by applying the Dedekind cuts. So, when we apply the
Dedekind cuts over the set of rational numbers rationales apply the Dedekind cuts or sections
then this collection of the rational numbers in fact it is equivalent to the rational numbers when
you apply the Dedekind cuts we are basically bringing this thing into form of the sections. So,
each section will correspond to some number  .
Now this  may a rational number or may be an irrational number. When we apply the
Dedekind cut or Dedekind theory over the set of rational number then it basically gives you the
sections and collection of all these sections gives the real number. That is the continuum,
because the sections which you are getting correspond to a number which may be real which
may be rational or may be irrational. So, over the rational number we are applying the Dedekinds
cut but we are getting a class which is bigger than the rational numbers, because the sections
which you are getting is larger than this elements itself that is the number of rational points,
because the rational point will also be add it there. So, this gives you the total aggregate as a
continuum.

Now the question is if suppose I apply again Dedekind cut or Dedekinds sake method theory
over the set of real numbers; that is we again find the sections or cuts by using the Dedekinds
method over the set of real number, whether will you get a set of a numbers bigger than this set
of real numbers. The question is, the answer is no. By this method we cannot enhance further the
set of real number to a bigger class. We will get basically the set of real number itself. So that is
why we call it this continuum is closed, but then somebody says why we are having the set of
complex numbers also which is an extension of the real number. But when you talk about the
complex number this basically is generated or obtained by some other trick, but the trick is that
we are choosing the square root of negative quantity, because if you look the complex number,
how this was introduced number.

The complex number is introduced basically when we look the solution of the equation
x 2  1  0 ; is it not. There is no real number which satisfies this equation, say, one which
satisfies this equation, but this is also an equation. This may be some physical phenomenon may
be governed by this equation. So, the question is whether can we further extend the system of
real numbers not because of the cut because cut is not helping us; Dedekind cut is not helping us
to extend it, but from other method and that method comes out to be choosing the square root of
minus one as an imaginary quantity i. We choose square root of minus one as an imaginary
number, and then once it is imaginary number then we develop the set of points where it is of the
form x +iy where x and y both are real, and basically this will be represented in the form of the
ordered pair (x, y) here x and y.
If x takes the position of real x axis that we call it as real and this is we called as an imaginary
axis. So, the position of this point (x, y) basically is the same as x + iy in a complex plane C in a
complex plane C. Now this leads to a extensive complex number system of the set of all complex
numbers which is an extension of real, because once you y 0 then basically it gives you entire
real line. So, it is an extension no doubt, but the way you have extended the real number is
entirely different the Dedekind’s method, clear. So, this is separate, and by this method when
you are extending the real to complex this collection of the complex number does not behave as
a smooth as a work set of real numbers, because between two real numbers one can easily
ordered one can say this two given number either they are equal or one is greater than the other.

But in case of the complex no ordering can be defined; in fact if suppose the complex numbers
are not ordered set. It means that is there no order can be derived. If z 1 z 2 are two complex
numbers you cannot say z 1 is greater than z 2 or z 1 is less than z 2 except when equal to of
course when you say z 1 is equal to z 2 we assume the real part is equal to real imaginary part is
equal to imaginary. But otherwise greater than less than not be in fact we did a contradiction,
because suppose there is an ordering defined. Suppose i is greater than 1 so i is complex number,
one is also complex number, because one can written as 1+ i0, then what happens? 1 is positive
so i is greater than one means you are assuming to be a positive quantity. So, let us multiply by i
again. So, what you get? i 2  1 again but i square is minus one is greater than one, so a
contradiction.

Similarly if we assume less than we also lead to a contradiction, okay. So, this shows that
ordering relation cannot defined over a set of complex numbers. So it is a different system
differently steam where we discuss the complex number, but this is a very important area where
the all analytic functions, all entire functions and all these things Cauchy various integrals real
integrals can be completed with the help of complex integration. So, it is very important field
where the people can use it in the application part. So, that is what?
(Refer Slide Time: 17:01)

Now let us see a few exercises which I wanted to give you. Show that by using Dedekind cuts
that if  and  are two irrational numbers then    is positive or negative according as
   or    . So, using the Dedekind cuts we wanted to prove this part, okay. So, let us see
suppose we have solution. Now  and  are given as two irrational numbers; so obviously it
can be represented by means of cuts. So, let  is represented by a cut (L 1, R 1) while the  is
represented by a cut (L 2, R 2), okay; now    ., so here we are having, so this is    . So, 
is here;  is somewhere here.

This is L 1, this is R 1, here this is L 2, here is R 2. Now when    it means every elements
alpha is greater than every number or every element every number of R 1 or belongs to R 1 is in R
2. All the elements of the R 1 will be in R 2 but every member of L 1 does not belong to L 2, is it
not; that is what is seen. So, every member of R 1 is in R 2, but every member of L 2 L 1 is not
there. So, there may be some member which is in R 2. So, let a 1 be such a member of L 1 which
does not belong to L 2. So, if it does not belongs to L 2 that is it belongs to R 2. So, here is some,
say, a 1 this is number a 1 which is in L 1 but not in L 2, so it must be in R 2. Let us choose a
number b 2.
Now let b2 be a member of R 2 such that a1  b2 . Suppose b2 is I am choosing this number, okay.

So, a1  b2 is positive b 2 be a number of R 2 such that a1  b2 no, a 1 oh sorry then yes. We

should write like this b 2 here, is it not. So, let it a 1; this is b 2, okay. So, a1  b2  0 , is this clear.

Now a 1 this    we wanted to prove this, what is to prove    is positive. So, when it is a
positive number, what do you mean by that? A cut if I apply the definition of Dedekind cuts then
a number is set to be a positive when its lower class contain some positive numbers, is it not.
Then only it is considered to be positive, okay. So, lower class will be some L 3,  
represented by a class (L 3, R 3).

So, let    is represented by a section, say, (L 3, R 3).; if I prove L 3 contains some positive
quantity positive number also then the    becomes greater than 0, okay. Now this L 3 what is
this alpha minus L 3 is the sum of so it is alpha minus is given by the L 3. Now L 3 is the sum of
what    ;  is in where? Alpha this is  , is it not, and  is here. Now, this a1  b2 is
positive. So, L 3 we can say like this that L 3 is the sum of any member of L 1 and any member of
minus L 2, and any member of let it not be such, let us write this thing be such that L 3 is the sum
of any member of L 1.

(Refer Slide Time: 24:30)

And any member of minus R 2 and any member of R 3 is the sum of any member of R 1 and any
member of minus L 2. Let us see what is the meaning of this? We wanted to show the    to
be positive. So, let us suppose    represents a cut by (L 3, R 3), okay, and we are assuming
that this cut is such that L 3 any member of L 3 is the sum of any member of L 1 and any member
of minus R 2. So, if you take any element of L 3 it can be written as the sum of the elements of L 1
and minus R 2 member of minus R 2. Similarly if you take a R 3 then any member of R 3 we are
assuming as a sum of the member of R 1 and any member of minus L 2, is it okay or not.

So therefore a1  b2 , what is this; is it not the same as a1  (b2 ) . Now a 1 belongs to what? a 1
belongs to L 1; b 1 b 2 is in R 2. So, minus b 2, b 2 is in R 2. So, minus b 2 will be in minus R 2, is it
not. Therefore this is the sum of L 1, this belongs to L1  ( R2 ) , and that is nothing but L 3, and

this a1  b2 is greater than 0 positive. So, a section (L 3, R 3) which is assumed to represent the

number    is such that it contains the positive numbers. So,    therefore this    is a
positive cut is positive, clear. So, that is what is written, is it okay or not.

Now let us take another exercise, okay. Specify the sections corresponding to the irrational
number e and prove that the sections so specified satisfies all the requirements satisfies
Dedekind’s theorem or satisfy all the postulates of Dedekind’s theorem, because there are three
postulates in Dedekind’s theorem that we can divide the whole system into real number into two
classes lower and upper class, and each class is nonempty. Second one is that at least lower class
every element in the lower class is the less than the elements of the upper class, and third is any
rational number will belong to any number either this class rational number or that class like this,
is it not. So, three postulates are there, so we can show. So, what it says is that corresponding to e
you first defines the sections, and then show that these sections satisfy the condition, okay.

1
Let us see this, what is e? If you remember e is the limiting value of (1  ) n . This is the e, is it
n
not, but here we cannot take the limit because we wanted in terms of the sections. So, let us say
1
let an  (1  ) n . here n is a positive integer. Now let us consider and define the class is as
n
follows, define the lower class L as the set of those rational number x such that an  x from and
after sometime from and after some fixed value of n; that is those rational number belongs to L
for which an  x for after some number n n0 for some value of n fixed value, say, after n equal

to n0 , n  x then we say x will be in L, and row upper class R, I am defining the this with upper
class R, okay, the rational number y is an upper class R if y is or such that write if a n is greater
than, okay, if y  an is strictly greater than a n for all values of n.

So, lower class I am defining the element x rational number belongs to the lower class if a n this
number a n is greater than x after a certain stage, say, a n is n0 and R is the those rational number

if y  an for each. We claim that this will be this section the way in which we have defined will
satisfy all the conditions of Dedekind theorem. So, what this first condition is both the class L
and R must be nonempty. So, there must be at least one element should be available in R from a n
as well as some element of a an must be in R, okay. Now if we look what is n? n is a positive
integer. Now if I take n any positive integer this sum will always be greater than 1; when n is
greater than 1 this number will always be greater than 1, is it not.

(Refer Slide Time: 33:17)

So, this one belongs to the lower class clearly number one belongs to lower class because a n

1
which is (1  )n will be greater than 1 for n  1 or it may be equal to 1; n is equal to 1 also. You
n
just check greater than sorry greater than 1; this is an  x no, so this is greater than 1, sorry. So,
x is basically 1. This number will always be greater than 1 when n is greater than equal to 1. So,
this therefore L is nonempty, okay. Similarly, x=3 if I take then all the elements of an s are less
1
than 3 for all n (1  ) n whatever the n you choose it will remain less than 3. So, 3 is an element
n
belongs to; therefore, 3 belongs to R. So, L and R both are nonempty. This is the first postulate
results that be shown in the Dedekind. Second one condition which is in the Dedekind shows that
every rational number p will be belongs to either L or R.

So, second one every rational number p belongs to L and R. Second is to show every rational
number p belongs to L or R; this we want to show, okay, which is obviously true which is
obvious, why? Suppose an  p for all n, p is a number, p is a number I am choosing. So, there

are two possibility either n will be greater than for all p or may be an  p for some after certain

stage, okay. So, an  p for all n then this p belongs to the upper class, is it not, upper class, and
if a n is less than if a n is greater than, no, no, this is if a n is less than, sorry, what was the upper
case. Yes, an  y , y is greater than a n for all a n. So, if an  p for all n then it is in upper class,

and if an  p for some n then this p belongs to the lower class.

1
So take any p. Suppose I take p any number say 2 then you can say (1  ) n then you can choose
n
n such a way that this class after a certain stage will satisfy this condition a n is greater than
because limiting value of this is 3, e is greater lying between this, is it not. Limiting value of this
1
will be this; what is this? If I expand it (1  ) , this is the number actually, 1 plus n plus n minus
2
1 and so on. So, when you take the 2 plus something. So, if you take a number if all the numbers
are less than 3, say, then this number, say, 4 then all the numbers are less than 4. If I take a
number 2 then there are some number where it is greater than. So, we can identify all the
numbers which is either in lower class or in the upper class. So, this is also true.

And third part is what? Every number a in the lower class is less than than the lower class. So, if
x belongs to the lower class it means that x  an for some n greater than equal to, say, m onward,
is it not, for all n greater than it. After certain stage it is less than this, but if x will be in y but any
element in the upper class R satisfy this condition; condition is that an  y for each n. So, from
here x  an for some n but this is less than y if this is an element in R. This is true for every n;
this is true for some n after some m after some m onward, okay. Then this will always be less
than y. So, any element in lower class will always be less than the elements in the upper class,
okay. So, this shows the property. So, all the conditions are satisfied Dedekind’s, okay; all
requirements of the Dedekind’s theorem is satisfied, so this… Let us see the next problem.

(Refer Slide Time: 39:38)

Suppose we have show that if a, b, x, y are rational numbers such that

(ay  bx)2  4(a  x)(b  y )  0 , then such that either x= a, y = b or then either this part or (1-ab)
and (1-xy) are squares of rational numbers, okay. It is basically mathematics things, not much
that Dedekind’s cuts is required but somewhere is required the concepts, okay. So, let
p  a  x, q  b  y , then if I solve this ay-bx then we get. If I just substitute these values we will
get the value to bp-aq which is also the same as yp- xq; this is what, okay. Just I take this p and q
then. So, our given relation one so from given relation we can write ay minus a means
(bp  aq)2  4 pq  0 , or we can also write this thing as, because this is also equivalent to this.

So, we can write ( yp  xq)2  4 pq  0 , okay.

Now clearly if p  0 , q is also 0; that is if x  a , y  b , then p 0 q 0 the equations these two are
satisfied, is it not. Obviously true, that is nothing. Now if they are not 0 if p is not 0 q is also not
0 then one can divide by 4 pq and one cam divide the equation two. So, from equation two we
get the relation okay; we get from here now we can write like this. This is now here, so we
wanted to write (bp  aq)2 . If I just made the plus sign then two times of this will come in
picture. So, here this is when you take the minus b and whole square so what you are getting is;
you are getting this like equal to plus 4 (1-ab) pq equal to 0, okay. Then only it is balanced.

Similarly if I take here ( yp  xq) 2 whole square then you are getting (1-xy) four times of this

(bp  aq)2
into pq is 0. This is from two. Now divide by pq. So, from here we get 1  ab  , but -
4 pq
4pq from here is nothing but which is equivalent to (bp+aq) divide by minus 4 p we can write as
(bp  aq)2 . Similarly 1-xy we can also write in a similar way, and this will be equal to if I put it

in this form we can (yp+xq) divided by whole square divided by ( yp  xq)2 . So, what we
conclude that if first is satisfied then also this equation is satisfied; otherwise, second 1-ab can be
literally square of the two rational numbers you see 1-ab is square of the rational number, 1-xy is
also square of the rational number and that proves the result, okay. Then last exercise let us see.

(Refer Slide Time: 46:00)

This is also interesting one. If d is a positive integer but not in the square of an, when d is a
positive integer not the square of an integer, it means square root of d becomes that it is not a
square of this integer. So, what we can say? It is not a perfect square, okay, square of integer, and
x( x 2  3d )
y is written in the form of suppose y  where x is given x is a positive rational
3x 2  d
2 x(d  x 2 ) ( x 2  d )3
number, then show that y  x  , and y 2
 d  .
3x 2  d (3x 2  d )2

Hence show that the section of the positive rational number determined by assigning to the upper
class all rational numbers whose square is greater than d and to the lower class all the other
rational numbers, so that the section of the positive rational number determined by assigning to
the upper class all rational numbers whose square is greater than d and to the lower class all other
rational number, is not generated by a rational number, okay.

Let us see, first let us see the first part of this I am not interested in this, because this is a routine
thing. You are having the value y is given, x is already given, so y-x you will find this
expression, y 2  d you will find the expression just simple calculation. So, we are not interested
in solving this, but this required by showing that this section is not a rational number, okay.

(Refer Slide Time: 50:55)

So what is the number of this? meaning is that this is our lower class, this is the upper class, we
are taking those positive rational numbers which are whose rational number says x such that x
square is greater than d and lower class rest of the rational numbers. So, this will include all
negative and the remaining positive also some positive, but here exclusively positive square is
greater than d. Now, this number say  we wanted to show  is not rational; this is all, okay.

(Refer Slide Time: 51:47)

So, let see the proof is very simple, just go ahead, okay. Clearly the upper class R has no least
number because if suppose x belongs to this class, suppose x belongs to the upper class R. So, by
definition so this implies that x 2  d ; when x 2  d and the function is defined like this so x 2  d
then y 2  d ; and here x square so y-x is negative, okay. So, we get from here is this one defined

so that then we find y as defined which is such that y  x and y 2  d . So, if I take any number x
in the upper class then what we are getting another number y which is lower than the x but still
stays in the upper class. This implies that y is in the upper class R. So, upper class cannot have a
least number.

Similarly lower class cannot have an upper number largest number we cannot have the lower
class as greatest number, why? Because the reason is for if suppose x belongs to the class L then
it x belongs to the class L then x 2  d ; otherwise x will be in upper class, okay. So, x 2  d then
the y so define will give what? If x 2  d then this is positive. So, y-x is positive, so y  x . So,
again the y belongs to lower class. So, if x is in the lower class then we are also getting another
number y which also in lower class but it is greater than L. So, L cannot have a largest number.
If x 2  d then this is positive. So, this is positive means ( y  x)  0 . So, y  x . It means by
another class another point in lower class which in this shows a contradiction. Therefore section
(L, R) represents irrational number; that is all.

Thank you. Clear?

Thanks all.