From Mahobe Resources (NZ) Ltd

We hope you enjoy these sample pages of the book Paradoxes and Sophisms in Calculus. We
have provided the first 42 pages of the book. The actual book can be purchased from our
website www.mahobe.co.nz

This book is a supplementary resource intended to enhance the teaching and learning of a
first-year university Calculus course. It can also be used in upper secondary school. It consists
of selected paradoxes and sophisms that can be used as a pedagogical strategy by creating
surprise and interest in the subject.

In this book the following major topics from a typical single-variable Calculus course are
explored: Functions, Limits, Derivatives and Integrals. As with the author’s previous book
“Counter-Examples in Calculus” (Maths Press, Auckland, New Zealand, 2004, ISBN
0-476-01215-5, 116 p.) the intention of this book is to encourage teachers and students to use
it in the teaching and learning of Calculus, with these purposes: • for deeper conceptual
understanding • to reduce or eliminate common misconceptions • to advance one’s
mathematical thinking, that is neither algorithmic nor procedural • to enhance generic critical
thinking skills – analysing, justifying, verifying, checking, proving which can benefit students
in other areas of life • to expand the ‘example set’ - a number of examples of interesting
functions for better communication of ideas in mathematics and in practical applications • to
make learning more emotional, active and creative The book can be useful for: • upper
secondary school teachers and university lecturers as a teaching resource • upper secondary
school and first-year university students as a learning resource.

If you enjoy the book then why not purchase it! Go to the Mahobe Resources (NZ) Ltd
website www.mahobe.co.nz

Go ahead and enjoy the first pages of this very informative and thought provoking book.
Copyright © 2007 Sergiy Klymchuk

This e-book is copyright. All rights reserved. Apart from any fair dealing
for the purpose of study, teaching, review, or otherwise permitted under
the Copyright Act, no part of this publication may be reproduced by any
process without written permission from the author.

ISBN 978-0-473-12798-5

The printed version first published

December 2005 ISBN 0-473-10550-0

Maths Press
PO Box 109-760
Newmarket
Auckland 1031
New Zealand

Klymchuk, Sergiy.
Paradoxes and Sophisms in Calculus
 Contents

Preface……………………………………………………………….7

Paradoxes

1. Functions and Limits………….………………………11

2. Derivatives and Integrals………………………........15

Sophisms

1. Functions and Limits………….………………………16

2. Derivatives and Integrals………………………........28

Solutions to Paradoxes

1. Functions and Limits………….………………………34

2. Derivatives and Integrals………………………........43

Solutions to Sophisms

1. Functions and Limits………….………………………45

2. Derivatives and Integrals………………………........53

 Preface

Dear God,
If I have just one hour remaining to live,
Please put me in a Calculus class
So that it will seem to last forever.
A bored student’s prayer

There are many interesting non-routine problems, puzzles, paradoxes

and sophisms. Some of them can change the life of students forever.
They can enthuse, enlighten and inspire. They can open the mind and
drive a person into the labyrinths of knowledge. They can encourage
passion for invention and discovery.

The teaching/learning process loses its effectiveness without emotional

involvement. What we tend to remember most are the events,
situations, news and facts that bring with them strong feelings and
emotions. The purpose of this book is to create and use such feelings
and emotions as a pedagogical strategy in a Calculus course.

The book is a supplementary resource intended to enhance the teaching

and learning of a first-year university Calculus course. It can also be
used in upper secondary school. The following major topics from a
typical single-variable Calculus course are explored in the book:
Functions, Limits, Derivatives and Integrals.

The book consists of two parts: paradoxes and sophisms. The word
paradox comes from the Greek word paradoxon which means
unexpected. There are several usages of this word, including those that
deal with contradiction. In this book, the word means a surprising,
unexpected, counter-intuitive statement that looks invalid but in fact is
true. The word sophism comes from the Greek word sophos which
means wisdom. In modern usage it denotes intentionally invalid
reasoning that looks formally correct, but in fact contains a subtle
mistake or flaw. In other words, it is a false proof of an incorrect
statement. Each such ‘proof’ contains some sort of error in reasoning.

7
Most students are exposed to sophisms at school. Often finding and
analyzing the mistake in a sophism can give a student deeper
understanding than a recipe-based approach in solving standard
problems. The former works on a psychological level whereas the latter
on a procedural level. Some basic examples of the sophism include
division by zero, taking only a non-negative square root and so on.
These tricks are used to ‘prove’ statements like “1 = 2”. In this book the
tricks from Calculus are used to ‘prove’ such statements.

Some of the paradoxes presented in the book (such as ‘Cat on a Ladder’

and ‘Encircling the Earth’) are at a lower level than Calculus, but they
can still be used in Calculus classes to demonstrate that sometimes our
intuition fails, even when we are dealing with very familiar shapes (like
circles for example). Much of the book’s content can be used as
edutainment – both education and entertainment.

As with my previous book, “Counter-Examples in Calculus” (Maths

Press, Auckland, New Zealand, 2004, ISBN 0-476-01215-5) the
intention of this book is to encourage teachers and students to use it in
the teaching/learning of Calculus with these purposes:

x For deeper conceptual understanding

x To reduce or eliminate common misconceptions
x To advance one’s mathematical thinking, that is neither
algorithmic nor procedural
x To enhance generic critical thinking skills – analysing, justifying,
verifying, checking, proving which can benefit students in other
areas of life
x To expand the ‘example set’ - a number of examples of interesting
functions for better communication of ideas in mathematics and
in practical applications
x To make learning more emotional, active and creative

The book can be useful for:

x Upper secondary school teachers and university lecturers as a
teaching resource
x Upper secondary school and first-year university students as a
learning resource

8
Acknowledgement

My thanks to Dr Farida Kachapova and Brody Radford from the

Auckland University of Technology for proofreading and formatting the
text and putting the material on AUT Online, the University’s web-based
learning system.

Contact Details for Feedback

Please send your questions and comments about the book to the postal
address:

Dr Sergiy Klymchuk
Associate Professor
School of Mathematical Sciences
Faculty of Design and Creative Technologies
Auckland University of Technology
Private Bag 92006
Auckland 1020
New Zealand

or e-mail to:

[email protected]

Sergiy Klymchuk
June 2006

9
 Paradoxes
I see it but I don't believe it!
Georg Cantor (1845 – 1918), in a letter to R. Dedekind (1877)

1. Paradoxes: Functions and Limits

1. Laying bricks
Imagine you have an unlimited amount of the same ideal
homogeneous bricks. You are constructing an arc by putting the
bricks one on top of another without using any cementing solution
between them. Each successive brick is further to the right than
the previous (see the diagram below).

How far past the bottom brick can the top brick extend?

2. Spiral curves
Construct two similar-looking spiral curves that both rotate
infinitely many times around a point, with one curve being of a
finite length and the other of an infinite length.

3. A tricky curve
Construct a curve that is closed, not self-crossing, has an infinite
length and is located between two other closed, not self-crossing
curves of a finite length.

11
4. A tricky area
A square with sides of 1 unit (and therefore area of 1 square unit)
is divided into 9 equal squares, each with sides 1/3 unit and areas
of 1/9 square unit, then the central square is removed. Each of the
remaining 8 squares is divided into 9 equal squares and the central
squares are then removed. The process is continued infinitely many
times. The diagram below shows the first 4 steps. At every step 1/9
of the current area is removed and 8/9 is left, that is at every step
the remaining area is 8 times bigger than the area removed. After
infinitely many steps what would the remaining area be?

5. A tricky ‘next term’

What is the next term in the sequence 2, 4, 8, 16?

6. A tricky shape
It looked like the cross-section of an object of circular shape. To
determine whether it was a circle, a student suggested measuring
its several diameters (the length of line segments passing through
the centre of symmetry of the figure and connecting its two
opposite boundary points). The student reasoned that if they all
appeared the same, the object had a circular shape. Was the
student’s reasoning correct?

7. Rolling a barrel
A person holds one end of a wooden board 3 m long and the other
end lies on a cylindrical barrel. The person walks towards the
barrel, which is rolled by the board sitting on it. The barrel rolls
12
 without sliding. This is shown in the following diagram:

What distance will the person cover before reaching the barrel?

8. A cat on a ladder
Imagine a cat sitting half way up a ladder that is placed almost
flush with a wall. If the base of the ladder is pushed fully up
against the wall, the ladder and cat are most likely going to fall
away from the wall (i.e. the top of the ladder falls away from the
wall).

Part 1: If the cat stays on the ladder (not likely perhaps?) what will
the trajectory of the cat be? A, B or C?

A B C

Part 2: Which of the above options represents the cat’s trajectory if

instead of the top of the ladder falling outwards, the base is pulled
away? A, B or C?

9. Sailing
A yacht returns from a trip around the world. Different parts of the
yacht have covered different distances. Which part of the yacht has
covered the longest distance?

13
10. Encircling the Earth
Imagine a rope lying around the Earth’s equator without any bends
(ignore mountains and deep-sea trenches). The rope is lengthened
by 20 metres and the circle is formed again. Estimate how high
approximately the rope will be above the Earth:
A) 3 mm B) 3 cm C) 3 m?

11. A tricky equation

To check the number of solutions to the equation
x
§1·
log 1 x ¨ ¸ one can sketch the graphs of two inverse functions
16 © 16 ¹
x
§1·
y log 1 x and y ¨ ¸ .
16 © 16 ¹

0
-2 0 2 4

-1

-2

From the graphs we can see that there is one intersection point and
x
§1·
therefore one solution to the equation log 1 x ¨ ¸ . But it is easy
16 © 16 ¹
1 1
to check by substitution that both x and x satisfy the
2 4
equation. So how many solutions does the equation have?

14
 2. Paradoxes: Derivatives and Integrals

1. An alternative product rule

The derivative of the product of two differentiable functions is the
product of their derivatives: (uv)c u cv c . In which cases is this ‘rule’
true?

2. A ‘strange’ integral
dx
Evaluate the following integral ³ dx .

3. Missing information?
At first glance it appears there is not enough information to solve
the following problem: A circular hole 16 cm long is drilled through
the centre of a metal sphere. Find the volume of the remaining part
of the sphere.

4. A paint shortage
1
To paint the area bounded by the curve y, the
x
x-axis and the line x = 1 is impossible. There is not enough paint in
f
1
the world, because the area is infinite: ³ dx lim(ln b  ln 1) f.
1
x b of

However, one can rotate the area around the x-axis and the
resulting solid of revolution would have a finite volume:
f
1 1 1
S ³ 2 dx S lim(  ) S . This solid of revolution contains the area
b of b 1
1 x

which is a cross-section of the solid. One can fill the solid with
S cubic units of paint and thus cover the area with paint. Can you
explain this paradox?

15
 Sophisms
1 + 1 = 3 for large values of 1.
A student joke

1. Sophisms: Functions and Limits

1. 1 = 0
n
1
Let us find the limit lim ¦ using two different methods.
n of 2
k 1 n k
a)
n
1 1 1 1
lim ¦ lim  lim  ...  lim 0  0  ...  0 0
n of
k 1 n2  k n of
n2 1 n of
n2  2 n of
n2  n
n 1
b) The lower and upper boundaries for the sum ¦ are:
k 1 n2  k
n 1 n 1 n 1
¦ <¦ <¦
2 2
k 1 n n k 1 n k k 1 n2
n 1 n n 1 n
Since lim ¦ 2
lim 2
1 and lim ¦ 2
lim 1
n of k 1
n n n of
n n n of k 1
n n of
n2
n 1
then by the Squeeze Theorem lim ¦ 1.
n of k 1
n2  k
Comparing the results in a) and b) we conclude that 1 = 0.

2. 1 = 0
1
Let us find the limit lim( x sin ) using two different methods.
xo0 x
1 1
a) Since  1 d sin d 1 then  x d x sin d x . Applying the Squeeze
x x
1
Theorem we conclude that lim( x sin ) 0 .
x o0 x
sin x
b) It is known that lim 1 . Rewriting the limit we receive
x o0 x
1 sin(1 / x)
lim( x sin ) lim 1.
x o0 x x o0 (1 / x)
Comparing the results in a) and b) we conclude that 1 = 0.

16
3. 1 = 0
Let us find the limit lim( x x ) using two different methods.
x o0

a) First find the limit of the base and then the other limit:
lim( x x ) (lim x) x lim 0 x 0 .
x o0 x o0 x o0

b) First find the limit of the power and then the other limit:
( lim x )
lim( x x ) x xo0
lim x 0 1.
x o0 x o0

Comparing the results in a) and b) we conclude that 1 = 0.

4. 1 = ’
Let us find the limit lim n n using two different methods.
n of

a) First find the limit of the expression under the radical and then
the other limit: lim n n n lim n lim n f f .
n of n of nof

b) First find the limit of the nth root and then the other limit:
1 1
lim
lim n n lim n n n nof n
lim n 0 lim1 1 .
n of n of nof nof

Comparing the results in a) and b) we conclude that 1 = ’.

5. sin kx k sin x
sin x
It is known that lim 1.
x o0 x
Using this fact we can find the following two limits:

sin kx sin kx sin u

a) lim k lim k lim k.
x o0 x x o0 kx u o0 u

k sin x sin x
b) lim k lim k.
x o0 x x o0 x

sin kx k sin x sin kx k sin x

Since lim lim then and sin kx k sin x .
x o0 x x o0 x x x

6. 1 = 0
We know that the limit of the sum of two sequences equals the sum
of their limits, provided both limits exist. We also know that this is
true for any number k of sequences in the sum.
1
Let us take n equal sequences and find the limit of their sum
n

17
 when nń’:
1 1 1 1 1 1
lim(   ...  ) lim  lim  ...  lim 0  0  ...  0 0 .
n of n n n n of n nof n n of n

1 1 1 1
On the other hand, the sum   ...  is equal to n u 1.
n
n

 n n
n

So we receive 1 = 0.

7. 1 = -1
Let us find two limits of the same function:
x
1
x y y
a) lim lim lim lim lim(1) 1 .
x of y of x  y x o f y of x x of
1
y
y
1
x y x lim 1 1 .
b) lim lim lim lim
y of x of x  y y of x of y y of
1
x
x y x y
Since lim lim lim lim , the results from a) and b) must
x of y of x  y y of x of x  y

be equal. Therefore we conclude that 1 = -1.

8. a = 1/a
Let a be any non-zero number.
Let us find two limits of the same function:
ax
1
ax  y y 1 1
a) lim lim lim lim lim .
x of y of x  ay x of y o f x x of a a
a
y
y
a
ax  y x lim a a .
b) lim lim lim lim
y of x of x  ay y o f x of ay y of
1
x
ax  y ax  y
Since lim lim lim lim , the results from a) and b) must
x of y of x  ay y of x of x  ay

be equal. Therefore we conclude that a = 1/a.

18
9. 1 = 2
a) Take two line segments of length 1 unit and 2 units and
establish a one-to-one correspondence between their points as
shown on the diagram below:

The number of points on the line segment of length 1 unit is the

same as the number of points on the line segment of length 2 units,
meaning 1 = 2.

b) Take two circles of radius 1 unit and 2 units and establish a

one-to-one correspondence between their points as shown on the
diagram below:

The number of points on the circumference of the inner circle is the

same as the number of points on the circumference of the outer
circle, so we can conclude that 1 = 2.

10. R = r
Two wheels of different radius are attached to each other and put
on the same axis. Both wheels are on a rail (see the diagrams on
the next page). After one rotation the large wheel with radius R
covers the distance AB which is equal to the length of its
circumference 2›R. The small wheel with radius r covers the
distance CD which is equal to the length of its circumference 2›r. It
is clear that AB = CD, therefore 2›R = 2›r and R = r.

19
 Wheel

Rail

Cross-section showing the wheel and the rail.

C D

A B

11. 2 > 3
We start from the true inequality:
2 3
1 1 §1· §1·
! or ¨ ¸ ! ¨ ¸
4 8 © 2¹ © 2¹
Taking natural logs of both sides:
2 3
§1· §1·
ln¨ ¸ ! ln¨ ¸
© 2¹ © 2¹
Applying the power rule of logs:
§1· §1·
2 ln¨ ¸ ! 3 ln¨ ¸
© 2¹ © 2¹
§1·
Dividing both sides by ln¨ ¸ :
© 2¹
2 > 3.

20
12. 2 > 3
We start from the true inequality:
2 3
1 1 §1· §1·
! or ¨ ¸ ! ¨ ¸
4 8 © 2¹ © 2¹
1
Taking logs with the base of both sides:
2
2 3
§1· §1·
log 1 ¨ ¸ ! log 1 ¨ ¸
2© 2¹ 2© 2¹

Applying the power rule of logs:

§1· §1·
2 log 1 ¨ ¸ ! 3 log 1 ¨ ¸
2© 2¹ 2© 2¹

§1·
Since log 1 ¨ ¸ 1 we obtain: 2 > 3.
2© 2¹

1 1
13. !
4 2
We start from the true equality:
1 1
2 2
Taking natural logs of both sides:
1 1
ln ln
2 2
Doubling the left hand side we obtain the inequality:
1 1
2 ln ! ln
2 2
Applying the power rule of logs:
2
§1· §1·
ln¨ ¸ ! ln¨ ¸
© 2¹ © 2¹
2
§1· 1 1 1
Since y = ln x is an increasing function ¨ ¸ ! or ! .
© 2¹ 2 4 2

21
14. 2 = 1
Let us take an equilateral triangle with sides of 1 unit. Divide the
upper sides by 2 and transform them into a zig-zagging, segmented
line as shown on the diagram below:

1 1

1 1 1 1

a) The length of this segment line is 2 units because it is

constructed from two sides of 1 unit each. We continue halving the
triangle sides infinitely many times. At any step the length of the
segment line equals 2 units.

b) On the other hand from the diagram we can see that with more
steps, the segment line gets closer and closer to the base of the
triangle which has length 1 unit. That is lim S n 1 , where S n is the
n of

length of the segment line at step n.

Comparing a) and b) we conclude that 2 = 1.

15. › = 2
Let us take a semicircle with diameter d. We divide its diameter into
d
n equal parts and on each part construct semicircles of diameter
n
as shown on the following diagram:

22
 Sd
a) The arc length of each small semicircle is . The total length Ln
2n
Sd Sd
of n semicircles is Ln un . Therefore the limit of Ln when
2n 2
Sd Sd
nń’ is: lim Ln lim .
n of n of 2 2
b) From the diagram we can see that when n increases, the curve
consisting of n small semicircles gets closer to the diameter, which
has length d. That is lim Ln d.
n of

Sd
Comparing a) and b) we see that d and conclude that › = 2.
2

16. › = 0
Let us find the lateral surface area of a cylinder with height 1 unit
and radius 1 unit using the following approach. Divide the cylinder
into n horizontal strips. Divide the circumference of each
cross-section by points into m equal parts. Rotate all odd-numbered
circumferences in such a way that the points on them are exactly
midway between the points on the even-numbered circumferences.
Form 2mn equal isosceles triangles by joining any two adjacent
points on each circumference with a point midway between them
on the circumferences above and below.

23
Using simple geometry of a right-angled triangle it can be shown
that the area of the resulting polyhedral surface is:
S S
S mn 2m sin 1  4n 2 sin 4 .
m 2m
When both m and n tend to infinity this area tends to the lateral
surface area of the cylinder. The limit of Smn is found using the well
sin x
known formula lim 1 . Let us consider 3 cases.
x o0 x

a) n = m

4
S § S ·
sin 4 2 ¨ sin ¸
S S m S m ¨ 2 m
lim S m lim 2m sin 1  4m 6 sin 4 lim 2S 1 ¸ f.
m of mof m 2m m of S 4 ¨ S ¸
¨ ¸
m © 2m ¹

b) n = m2
4
S § S ·
sin 4 ¨ sin
S S m S ¨ 2m ¸¸
lim S m lim 2m sin 1  4m 4 sin 4 lim 2S 1
m of m of m 2m m of S 4 ¨ S ¸
¨ ¸
m © 2m ¹
S4
2S 1  .
4

24
 c) n = m3
4
S § S ·
sin 4 2 ¨ sin ¸
lim S m lim 2m sin
S
1  4m 6 sin 4
S
lim 2S m 1 S m ¨ 2m ¸ f.
m of m of m 2m m of S 4 ¨ S ¸
¨ ¸
m © 2m ¹
In all three cases a), b) and c) when mń’ the polyhedral surface
tends to the lateral surface of the cylinder. So the limit lim S m in all
m of

three cases a), b) and c) must be the same.

This is only possible if › = 0.

17. Achilles and the Tortoise

This is one of the sophisms created by the Greek philosopher Zeno
in the 5th century B.C. Sometimes they are called Zeno’s paradoxes,
but in the sense of “paradox” and “sophism” accepted in this book
they are considered sophisms.
In a race between Achilles, the fastest of Greek warriors, and a
tortoise that had a head start, Achilles will never pass the tortoise.
Suppose the initial distance between them is 1 unit and Achilles is
moving 100 times faster than the tortoise. When Achilles covers the
1 th
distance of 1 unit the tortoise will have moved of a unit
100
further from its starting point. When Achilles has covered the
1 th 1
distance of of a unit the tortoise will move of a unit
100 100 2
1
further, and so on. The tortoise is always ahead of Achilles by
100 n
of a unit no matter how long the race is. This means that Achilles
will never reach the tortoise.

18. A snail
Imagine a snail moving at a speed of 1 cm/min along a rubber rope
1 m long. The snail starts its journey from one end of the rope. After
each minute the rope is uniformly expanded by 1 m. Below is a
‘proof’ that at some stage the snail will eventually reach the other
end of the rope.
1 th
In the first minute the snail will cover the first of the rope.
100
1 th
In the second minute the snail will cover of the rope.
200

25
 1 th
In the third minute the snail will cover of the rope, and so on.
300
The distance covered by the snail after n minutes will be
1 § 1 1 1·
¨1    ...  ¸ . The sum in the brackets represents the first n
100 © 2 3 n¹
terms of the harmonic series, which is divergent. The sum in the
brackets (and therefore the distance) can be made bigger than any
number. So no matter how big the length of the rope is, the
distance covered by the snail at some stage will be bigger than the
length of the rope. This means that the snail will reach the other
end of the rope.

19. 1,000,000 § 2,000,000

If we add 1 to a big number the result would be approximately
equal to the original number. Let us take 1,000,000 and add 1 to it.
That is 1,000,000 § 1,000,001.

Similarly 1,000,001 § 1,000,002.

And 1,000,002 § 1,000,003.

And so on…

1,999,999 § 2,000,000.

Multiplying the left-hand sides and the right-hand sides of the

above equalities we receive:

1,000,000 u 1,000,001 u ... u 1,999,999 | 1,000,001 u 1,000,002 u ... u 2,000,000.

Dividing both sides by 1,000,001 u ... u 1,999,999 we conclude that
1,000,000 | 2,000,000.

20. 1 = -1
Since a u b a u b , it follows that
1 1 (1) u ( 1) 1 u 1 i u i i2 1.

21. 2 = -2
Two students were discussing square-roots with their teacher.
The first student said: “A square root of 4 is -2”.

26
 The second student was sceptical, and wrote down 4 2.
Their teacher commented: “You are both right”.
The teacher was correct, so 2 = -2.

22. 2 = 1
Let us find the equation of a slant (or oblique) asymptote of the
x2  x  4
function y using two different methods.
x 1
x2  x  4 6
a) By performing long division: x2 . The last term,
x 1 x 1
6
tends to zero as x o f . Therefore as x o f the function
x 1
approaches the straight line y x  2 , which is its slant asymptote.

b) Dividing both numerator and denominator by x we receive:

4
x 1
x . Both 4 and 1 tend to zero as x o f . Therefore
2
x x4
x 1 1 x x
1
x
as x o f the function approaches the straight line y x  1 which is
its slant asymptote.

x2  x  4
The function y has only one slant asymptote. Therefore
x 1
from a) and b) it follows that x  2 x  1 .
Cancelling x we receive 2 = 1.

27
 2. Sophisms: Derivatives and Integrals

1. 1 = C, where C is any real number

Let us apply the substitution method to find the indefinite integral
³ sin x cos xdx using two different methods:
ªu sin x º u2 sin 2 x
a) ³ sin x cos xdx « » ³ udu 2  C1  C1
¬du cos xdx ¼ 2
ªu cos x º u2 cos 2 x
b) ³ sin x cos xdx « »  ³ udu  2  C 2  2  C 2 ,
¬du  sin xdx ¼
where C1 and C2 are arbitrary constants. Equating the right hand
sides in a) and b) we obtain
sin 2 x cos 2 x
 C1 =   C2 .
2 2
Multiplying the above equation by 2 and simplifying we receive
sin 2 x  cos 2 x 2C 2  2C1 or sin 2 x  cos 2 x C since the difference of
two arbitrary constants is an arbitrary constant. On the other hand
we know the trigonometric identity sin 2 x  cos 2 x 1 . Therefore 1 = C.

2. 1 = 0
1
Let us find the indefinite integral ³ x dx using the formula for

integration by parts ³ udv uv  ³ vdu :

ª 1 1º
1 «u du  §1· § 1· 1
³ x dx x x2 » ¨ ¸ x  ³ x¨  2 ¸dx 1  ³ dx .
« » © x¹ © x ¹ x
¬dv dx v x ¼
1 1
That is, ³ dx 1  ³ dx .
x x
1
Subtracting the same expression ³ x dx from both sides we receive

0 = 1.

3. Division by zero
dx
Let us find the indefinite integral ³ 2x  1 by the formula

f c( x) dx
³ f ( x) ln f ( x)  C using two different methods:

dx 1 dx 1 1
a) ³ 2x  1 2 ³ 1 ln x   C1 .
2 2
x
2

28
 dx 1 2dx 1
b) ³ 2 x  1 2 ³ 2 x  1 2 ln 2 x  1  C 2 .
Equating the right hand sides in a) and b) we obtain
1 1 1
ln x   C1 ln 2 x  1  C 2
2 2 2
Since C1 and C2 are arbitrary constants that can take any values,
1 1 1
let C1 = C2 = 0. Then ln x  ln 2 x  1 .
2 2 2
1 1
Solving for x we receive x  2 x  1, x  .
2 2
Substituting this value of x into the original integral gives zero in
the denominator, so division by zero is possible!

4. sin 2 x 1 for any value of x

Let us differentiate the function y = tan x twice:
1 2 sin x
yc , y cc .
cos 2 x cos 3 x
The second derivative can be rewritten as
2 sin x 2 sin x 1
y cc 2 yy c ( y 2 )c .
2 tan x u
cos 3 x cos x u cos 2 x 2
cos x
Integrating both sides of the equation y cc ( y 2 )c we receive
2 1 2 1 sin 2 x
y c y or tan x , or .
cos 2 x cos 2 x cos 2 x
From here sin 2 x 1 .

ªS º
5. 0  « , S »
2 ¬ ¼
S
dx
Let us estimate the integral ³ 1  cos 2
.
0 x
1 1 S
1 S
dx S
a) Since d d 1 on [0, S ] then ³2 dx d ³ 1  cos 2 x ³ dx or
d
2 1  cos 2 x 0 0 0

S S dx
d³ dS .
2 0 1  cos 2 x

29
 b) On the other hand
dx
ªt tan x x 0 t 0º
S
dx S
cos 2 x
0
dt
³ 1  cos 2 x ³ 1  tan 2 x  1 ««dt dx
x S t 0
»
» ³ 2  t2 0.
0 0 0
¬ cos 2 x ¼
ªS º
Comparing a) and b) we conclude that 0  « , S » .
¬2 ¼

6. ln 2 is not defined
1
Let us find the area enclosed by the graph of the function y ,
x
the x-axis and the straight lines x = -2 and x = -1 using two
different methods (see the diagram on the next page).
1
a) On one hand, the derivative of the function y = ln x is y c and
x
1
therefore an antiderivative of f ( x) is F ( x) ln x . We can apply
x
1
1
the Newton-Leibnitz formula to the integral ³ dx (the limits are
2 x

finite and the function is continuous on [-2,-1]) to find the required

1
1
area: A  ³ dx (ln(1)  ln(2)) . The area is undefined since the
2 x

logarithm of a negative number does not exist.

b) On the other hand, this area is the same as the area enclosed by
1
the graph of the function y , the x-axis, and the straight lines
x
x = 1 and x = 2 due to the symmetry of the graph about the origin.
2
1
Therefore the area equals: A ³ xdx ln 2  ln1 ln 2.
1

Comparing a) and b) we conclude that ln 2 is not defined.

30
 1.5

0.5

0
-4 -2 0 2

-0.5

-1

-1.5

-2

7. › is not defined
Sx  sin x
Let us find the limit lim using two different methods.
x ofx  sin x
sin x
S
Sx  sin x x
a) lim lim S.
x of x  sin x x of sin x
1
x
b) Since both numerator and denominator are differentiable we can
f ( x) ª f º f c( x)
use the well known rule lim
x of g ( x ) «¬ f »¼ xof g c( x) which gives us:
lim

Sx  sin x ªf º S  cos x
lim «¬ f »¼ lim , which is undefined.
x of x  sin x x of 1  cos x

Comparing the results in a) and b) we conclude that › is not

defined.

8. 0 = C, where C is any real number

We know the property of an indefinite integral:
³ kf ( x)dx k ³ f ( x)dx where k is a constant.
Let us apply this property for k = 0.
a) The left hand side of the above equality is ³ 0 f ( x)dx ³ 0dx C ,
where C is an arbitrary constant.
a) The right hand side is 0 ³ f ( x) dx 0.
Comparing a) and b) we conclude that 0 = C, where C is any real
number.

31
9. 1 = 2
Let us find the volume of the solid of revolution produced by
rotating the hyperbola y 2 x 2  1 about the x-axis on the interval
[-2, 2] using two different methods.
2 2
x3 2 4
a) V S ³ y 2 dx S ³ ( x 2  1)dx S (  x) S (cubic units).
2 2
3 2 3
b) Since the hyperbola is symmetrical about the y-axis we can find
the volume of a half of the solid of revolution, say on the right from
the y-axis and then multiply it by 2. Obviously the point (1,0) is a
vertex to the right of the origin and the right branch of the
hyperbola is to the right of the vertex (1,0). Therefore the volume of
2 2
2 2 x3 2 4
the right half is V1 S ³ y dx S ³ ( x  1)dx S (  x) S (cubic
1 1
3 2 3
8
units) and the total volume V 2V1 S (cubic units).
3
4 8
Comparing a) and b) we obtain S S or 1 = 2.
3 3

10. An infinitely fast fall

Imagine a cat sitting on the top of a ladder leaning against a wall.
The bottom of the ladder is pulled away from the wall horizontally
at a uniform rate. The cat speeds up, until it’s falling infinitely fast.
The ‘proof’ is below.

y l

32
 By the Pythagoras Theorem y l 2  x 2 , where x x(t ) , y y (t )
are the horizontal and vertical distances from the ends of the
ladder to the corner at time t. Differentiation of both sides with
xxc
respect to t gives us y c  . Since the ladder is pulled
l 2  x2
uniformly x c is a constant. Let us find the limit of y c when x
§ xxc ·
approaches l: lim y c lim¨¨  2 ¸¸ f . When the bottom of the
x ol x ol
© l  x2 ¹
ladder is pulled away by the distance l from the wall, the cat falls
infinitely fast.

11. A positive number equals a negative number

sin x
a) The function f ( x) is continuous and non-negative on
1  cos 2 x
3S
the interval ª«0, º» . Therefore by the definition of the definite
¬ 4¼
integral the area enclosed by the function f(x) and the x-axis on the
3S
interval ª«0, º» is a positive number.
¬ 4¼
b) On the other hand, since the function F ( x) tan 1 (sec x) is an
antiderivative of the function f(x) (this is easy to check by
3S
differentiation) calculating the area as the integral of f(x) on ª«0, º»
¬ 4¼
3
S
4
sin x S
we receive a negative number: ³ 1  cos dx  tan  1 2  .
0
2
x 4
Hence a positive number equals a negative number.

33
 Solutions to Paradoxes
1. Solutions to Paradoxes: Functions and
Limits

1. Laying bricks
The top brick can be infinitely far from the bottom brick! The
x-coordinate of the position of the centre of mass of a system of n
objects with masses m1, m2,…, mn is defined by the formula:

m1 x1  m2 x 2  ...  mn xn
x0 .
m1  m2  ...  mn

Let us consider two bricks. For the upper brick not to fall from the
lower brick the perpendicular distance from the centre of mass of
the upper brick should not be beyond the right edge of the lower
brick. That is, the maximum value of the x coordinate of the centre
of mass of the upper brick is l: x0 l . So the maximum shift is:
l
'x1 .
2

y
y

x
l

34
Let us consider three bricks.

x
l

l
For the top brick we have the maximum possible shift: 'x1 . Let
2
us find the maximum possible shift for the middle brick. Again, the
perpendicular distance from the centre of mass of the system of the
middle and top bricks should not extend past the right edge of the
lower brick. In other words, the maximum value of the x coordinate
of the centre of mass of the system of the middle and the top bricks
is l: x0 l . Expressing x0 for the system of the middle and the top
bricks from (1) we obtain:
l l l
m( 'x 2  )  m('x2   )
2 2 2 l.
2m

l
From here 'x 2 .
4
l l l
In a similar way we can obtain: 'x3 , 'x4 , ... , 'x n .
6 8 2n
l 1 1
Adding all shifts we receive: 'x1  'x 2  ...  'x n (1   ...  ) .
2 2 n
When nń’ the sum in the brackets tends to infinity. This means
that the maximum possible shift of the top brick with respect to
the bottom brick can be made as large as we want.
Comment: In practice it is of course impossible. Starting from a
certain value of n we will not be able to make shifts of the length
l
as they will be too small to perform.
2n

35
2. Spiral curves
a) Let us construct a spiral curve of a finite length.

Draw a line segment of length d. Draw a semicircle with diameter d

on one side of the line segment. Then on the other side of the line
segment draw a semicircle of diameter d/2. Then on the other side
draw a semicircle of diameter d/4, and so on.

The length of the curve is:

d d d 1 1 1
S  S  S  ... Sd (    ...) Sd .
2 4 8 2 4 8

b) Let us construct a spiral curve of infinite length. Draw a line

segment AB of length d with midpoint P. Draw a circle with the
centre C on the line segment at distance a from P. On one side of
the line segment draw a semicircle of the diameter d. On the other
side of the line segment draw a semicircle of the diameter AE,
where point E is the midpoint of PB. Then on the other side of the
line segment draw a semicircle of the diameter EF, where point F is
the midpoint of AD and so on (see the following diagram).

36
 The curve has infinitely many rotations around point C and each
rotation has a length bigger than the circumference 2›a, so the
length of the curve is infinite.

3. A tricky curve
One example of such a curve is the famous Koch snowflake. We
start with an equilateral triangle and build the line segments on
each side according to a simple rule. At every step each line
segment is divided into 3 equal parts, then the process is repeated
infinitely many times. The resulting curve is called the Koch curve
or Koch snowflake and is an example of a fractal. The first four
iterations are shown below:

The initial triangle and all consecutive stars and snowflakes are
located between the circumferences inscribed into the triangle and
circumscribed around it. Both circumferences have finite lengths.
If the perimeter of the initial triangle is 1 unit, then the perimeter
37
 4 1
of the star in the first iteration is 12 u units. The perimeter of
3 9
2
1 16 § 4 ·
the snowflake in the second iteration is 48 u ¨ ¸ units.
27 9 © 3 ¹
n
§ 4 ·
The perimeter of the snowflake in the nth iteration is ¨ ¸ units.
© 3¹
As nń’ the perimeter of the snowflake tends to infinity. The Koch
curve has an infinite length but bounds a finite area, which is
between the area of the circle inscribed into the initial triangle and
the area of the circle circumscribed around it.

4. A tricky area
Although at every step the remaining area is 8 times bigger than
the area removed, after infinitely many steps the remaining area
will be zero and the area removed will be 1 square unit. Let us
1 8
show this. After the first step the remaining area equals 1  .
9 9
2
8 1 64 §8·
After the second step the remaining area is  8u ¨ ¸ .
9 81 81 ©9¹
n
§8·
Similarly, after the nth step the remaining area is ¨ ¸ and if n
©9¹
tends to infinity this area tends to zero. This figure is called the
Sierpinski carpet and is another example of a fractal.

5. A tricky ‘next term’

a) The expected answer is 32, but there are infinitely many other
correct answers. Actually the next term in the sequence 2, 4, 8, 16
can be any number.
Let the nth term be a n 2 n  (n  1)(n  2)(n  3)(n  4) x .
The first 4 terms are 2, 4, 8, 16. One can make a5 equal to any
number by determining x from the formula for a5 . This can be done
to obtain the formula for the nth term. For example, let the 5th term
be –4. From the equation  4 2 5  4 u 3 u 2 u 1u x we obtain x = -1.5
and the formula for the nth term is
a n 2 n  (n  1)(n  2)(n  3)(n  4)(1.5) .

b) Another interesting example has a geometrical flavour. Draw a

circle, and put two dots on the circumference and connect them
with a line segment. The circle is divided into 2 regions. Put a third
38
 dot and connect all dots. The circle is now divided into 4 regions.
Put a fourth dot and connect all dots. The circle is now divided into
8 regions. Put a fifth dot and connect all dots. The circle is now
divided into 16 regions. It looks like we have a clear pattern. But
when you put the sixth dot and connect all dots the circle is divided
into 30 regions!

6. A tricky shape
No, the student was not right. A figure can be of constant diameter
yet not be a circle. As an example, consider the following curve. In
an equilateral triangle draw circular arcs with the radius equal to
the side of the triangle from each vertex. The resulting figure is a
curved triangle, which is called the Reuleaux triangle (see the
diagram below). One of its properties is that it has a constant
diameter. When it rolls on a horizontal surface its center moves
along a sine curve with ups and downs (unlike a circle whose
center does not move up and down – only along a straight
horizontal line). For this reason it is not practical to use it as a
wheel, but it does have practical applications. It is used in the
Wankel rotary engine, and in some countries manhole covers are
shaped like the Reuleaux triangle.

7. Rolling a barrel
The barrel rolls as long as the person continues walking. The
velocity of the point on the top of the barrel equals the velocity of
the walking person and is twice the velocity of the axis of the barrel,
so the person will cover 6 m by the time he reaches the barrel.

8. A cat on a ladder
Part 1. Most people are confident that C is the answer to Part 1.
Without much difficulty, one can imagine the ladder rotating about
a central point, i.e. where the base of the ladder touches the wall.

39
An arc is the result and in this case represents a quarter of the
circumference of a circle.

Part 2. However, Part 2 isn’t so easy! Many people conclude that A

is the correct answer. It sounds reasonable that as the ladder
slides outwards away from the wall that it would appear to drop
quickly, then level out as it approaches the horizontal.
Surprisingly, the answer to this problem is also C. Try it out by
making a model (see the sequential sketches below). With a paper
ladder, with a point marking half way, slowly slide the ladder down
and away from the wall. After each small amount of movement, put
a dot on the page at the place where the centre of the ladder lies.
Note that as the ladder approaches the horizontal, further lateral
movement is minimal.

Here is a simple proof:

40
 B D
C

O A

Let AB be the ladder. Point C (where the cat sits) is always at the
same distance (half the length of the ladder) from the point O
regardless of the position of the ladder. This comes from the fact
that diagonals in a rectangle are the same and are divided in half
by the point of their intersection.

You may well be very surprised to see that the trajectory is the
same in both cases. Do not be alarmed however – in this case the
intuition of many people fails. My colleague tried this test out with
a class of 100 4th year engineering students in Australia, Germany,
New Zealand and Norway. These young men and women, aged
about 21, are expected to be able to quickly conceptualise shapes,
dimensions, movements and forces. The students were given 40
seconds to find the answer. They were told that it was a mental
exercise, with no calculations or drawings permitted. The results
were startling, for although 74% of the students gave C, the correct
answer to Part 1, 86% were wrong in Part 2. 14% and 34% gave B
as the answer in Parts 1 and 2 respectively.

9. Sailing
The top of the yacht has covered the longest distance. The shape of
the Earth is approximately spherical, so the top of the yacht has
the longest radius compared to lower parts and therefore has the
longest circumference.

41
10. Encircling the Earth
Approximately 3 m high. This is a surprising answer for many
people. Let r be the radius of the Earth and R be the radius of the
circle after adding 20 metres to the rope. The difference between
the two circumferences is 20 m: 2SR  2Sr 20 or 2S ( R  r ) 20.
From here the difference between the two radii is R  r | 3m. The
answer does not depend on the original length of the rope.

11. A tricky equation

The rough sketch of the graphs is “too rough”. Both functions are
decreasing for all x in their domains but they are very close to both
axes, and in fact have 3 intersection points. It can be shown that
the equation has 3 solutions.

42