CHAPTER 24

Techniques of proof IV:

Induction

One thing leads to another.

Anon.

Induction is a very powerful technique used regularly by mathematicians.1 Initially, it can

be confusing because it looks like we assume what is to be proved. As we know, that never
proves theorems. On the plus side, spotting when to use it is easy and we need only check
two conditions to apply it.
Induction is applied when we have an infinite number of statements indexed by the
natural numbers such as

‘n 5 − n is even for all n ∈ N’.

It is not sufficient to prove this for a sample of natural numbers, whether that sample
involves hundreds, millions or even billions of numbers; we have to prove it for all n.
With induction we don’t prove the statements directly. What we do is perhaps best
described by analogy with domino toppling. As you are probably aware, this is where
dominoes are standing on their ends in such a way that when you push the first one over,
it knocks the second domino over, that in turn knocks the third down, and so on. Provided
the dominoes are arranged so that each knocks down the next, then all of them will fall.
The process of induction is that we prove that ‘if the kth statement is true, then the k +1th
statement is true’, i.e. the truth of one statement implies the truth of the next one. This is
analogous to one domino knocking down the next one. So, if the first statement is true (push
the first domino), then all the statements are true (all the dominoes get knocked down).

The Principle of Mathematical Induction

Let’s begin to make the idea precise. First, the Principle of Mathematical Induction
requires that we have a sequence of statements indexed by the natural numbers. There are
plenty of these, as the following examples show. Note that they are from diverse areas
of mathematics: divisibility, summation, inequalities. These statements will all be proved

1
There is a principle of induction in experimental science which is different to the one we are about to discuss.

166

https://doi.org/10.1017/CBO9780511808258.025 Published online by Cambridge University Press

 The Principle of Mathematical Induction 167

later using induction. Hopefully, by showing you this scope you will get an impression of
the power of induction – it really is an important tool in the mathematician’s toolkit.

Examples 24.1
(i) The expression 6n − 1 is divisible by 5 for all n ∈ N. (We say a is divisible by b if
there is no remainder when we divide a by b. It means that a = bq for some q ∈ Z.
We shall see more of this in Chapter 27.)
Pn
(ii) We have i=1 i = 12 n(n + 1) for all n ∈ N.
(iii) The inequality 2n−1 ≤ n! is true for all n ∈ N.

After reading these did you behave like a mathematician and try out a few cases to check
that the statements were at least plausible?
Let’s check some cases. For (i) we have

n =1: 61 − 1 = 6 − 1 = 5. Divisible by 5.
n =2: 62 − 1 = 36 − 1 = 35. Divisible by 5.
n =3: 63 − 1 = 216 − 1 = 215. Divisible by 5.
n = 10 : 610 − 1 = 60466176 − 1 = 60466175. Divisible by 5.

Notice that I did the first few cases and then threw in a higher case. This would seem to
indicate that the statement is reasonable – we haven’t found a counterexample yet!

Exercise 24.2
Check the other two statements above for small n to help convince yourself that they
are true.

This type of checking may seem pointless – ‘Aren’t we going to prove the statements are
true in a minute?’ Yes, however, what it does is engage the brain in the statement and
ensures that you see what is there – not what you think is there.
We will use the notation A(n) to denote the statement for a particular n. In (i) we have,
for example, A(3) is just ‘The expression 63 − 1 is divisible by 5.’
Now, ladies and gentlemen, I give you the grand theorem!

Theorem 24.3 (Principle of Mathematical Induction)

Let A(n) be an infinite collection of statements with n ∈ N. Suppose that

(i) A(1) is true, and

(ii) A(k) =⇒ A(k + 1), for all k ∈ N.

Then, A(n) is true for all n ∈ N.

Proof. We aim for a contradiction. Assume the conclusion is false and let j be the smallest
natural number such that A( j) is false. By assumption (i) we have that j > 1. Now we
note that A( j − 1) has to be true as j is the smallest possible. Hence by assumption
(ii) we have that A( j) is true. This is a contradiction. 

We need some definitions. The reasons for making them will become clearer later.

https://doi.org/10.1017/CBO9780511808258.025 Published online by Cambridge University Press

 168 C H A P T E R 24 Techniques of proof IV: Induction

A(4)
A(5)

A(6)
A(3)
A(2)
)
A(1
Figure 24.1 Toppling dominoes

• Checking condition (i) is called the initial step.

• Checking condition (ii) is called the inductive step.
• Assuming that A(k) is true for some k in (ii) is called the inductive hypothesis.
The idea is represented in Figure 24.1 where each statement is seen as a domino.

Examples of induction
Let’s see the principle in action. The next example is a real classic of mathematics: ‘What
is the sum of the first n numbers?’
In a famous story, when the great mathematician Johann Carl Friedrich Gauss
(1777–1855) was at school, his teacher asked the class to sum all the numbers up to
100 – presumably in the hope that it would keep them occupied for a lesson. While his
classmates worked through the exercise in the laborious, obvious way, Gauss thought like
a mathematician and avoided the drudgery by producing a simple method for calculating
the sum of the first 100 numbers, thwarting his teacher’s ambition for a quiet, easy lesson
for himself.
Gauss’ method can be applied to produce a formula for all numbers, not just 100. This
formula is that in Examples 24.1(ii) above. We shall prove the truth of this formula by
induction rather than use his method of proof.2

Example 24.4
Pn
We shall prove that i=1 i = 12 n(n + 1) for all n ∈ N.
This statement is indexed by n, so we can call A(n) the statement
n
X
i = 21 n(n + 1).
i=1
Our aim is to show that A(n) is true for each n.
Let us check condition (i), the initial case. This corresponds to pushing the first domino.
This is easy:
When n = 1, then
n
X 1
X
i= i =1
i=1 i=1

2
Gauss’ method was to note that we can pair up the numbers so that the sums of pairs is constant, i.e. 1 + 100,
2 + 99, 3 + 98, and so on. The rest of the formula is easy: we have 50 pairs (i.e. 100/2) summing to 101. The
formula in the general case uses the same argument.

https://doi.org/10.1017/CBO9780511808258.025 Published online by Cambridge University Press

 Examples of induction 169

and
1
2 n(n + 1) = 1
2 × 1 × (1 + 1) = 1.

Now for the inductive step. We now want to show that if the statement A(k) is true,
then A(k + 1) is true as well. This is the same as domino number k knocking over domino
k + 1.
Let us assume that A(k) is true for some arbitrary k. (This is a single k; the point is
that it is arbitrary. We make no assumptions for any other number n, just this particular
k.) That is,
k
X
i = 21 k(k + 1).
i=1

We investigate what implications this has for A(k + 1). First we write down the form of
A(k + 1):
k+1
X
i = 21 (k + 1)((k + 1) + 1).
i=1

How would a mathematician prove such an equality? Experience has taught us that we
should pick the complicated side of an expression and reduce it to get the other side. For
a proof by induction of a statement involving a sum the complicated side is usually the
one with the summation (as it probably has more terms).
We have
k+1 k
!
X X
i = i + (k + 1), by the definition of summation,
i=1 i=1
 
= 1
2 k(k + 1) + (k + 1), by the inductive hypothesis,
i.e. as A(k) is true,
 
= 1
2k + 1 (k + 1)

2 (k + 2) (k
1
= + 1)
2 (k + 1)((k
1
= + 1) + 1).

Thus, we have shown that if

k
X
i = 21 k(k + 1)
i=1

is true, i.e. A(k) is true, then

k+1
X
i = 12 (k + 1)((k + 1) + 1)
i=1

is true, i.e. A(k + 1) is true. In other words, we have shown that all the dominoes are lined
up so that if the kth falls, then the (k + 1)th falls.
Thus, by the Principle of Mathematical Induction the statement is true for all n ∈ N.

https://doi.org/10.1017/CBO9780511808258.025 Published online by Cambridge University Press

 170 C H APTER 24 Techniques of proof IV: Induction

Remarks 24.5
(i) Note the structure. We assume that A(k) is true for an arbitrary k. Then we look at
the A(k + 1) case, and tease it apart so that A(k) can be used. We don’t assume that
the A(k + 1) case is also true – this is a common error by beginners – we show that
it is true when A(k) is true.
(ii) Another problem for novice mathematicians is that induction seems to violate the
principle that we do not assume what we are trying to prove. I felt this myself as a
student.
However, it is vital to grasp the subtlety in induction if it is to be applied confidently.
It should be noted that the statement we wish to prove is about something holding for
all n. In our assumption in condition (ii) we assume that the statement holds for one
particular n which we call k. Ok, that one particular n is arbitrary, it is absolutely any
n you like, but it is still just one, by itself, no other assumptions made.

Example 24.6
The expression 6n − 1 is divisible by 5 for all n ∈ N.
We shall do this with less explanation than the previous example – this example will be
more like a model solution in a book or submitted for an assignment.
Initial step: The statement is true for n = 1 as 6n − 1 = 61 − 1 = 5.
Inductive step: Assume the statement is true for some3 k ∈ N; this means that 6k − 1 =
5m for some m ∈ N. Then
6k+1 − 1 = 6(6k ) − 1
= 6(5m + 1) − 1, by the inductive hypothesis,
= 30m + 6 − 1
= 5(6m + 1).
This is divisible by 5 and so the statement is true for k + 1. Hence, by the Principle of
Mathematical Induction, the statement is true for all n ∈ N.
Now for the third example.

Example 24.7
We show that 2n−1 ≤ n! for all n ∈ N.
Initial step: For n = 1 we have 2n−1 = 20 = 1 and n! = 1! = 1. Hence, 2n−1 ≤ n! for
n = 1.
Inductive step: Assume the statement is true for some k ∈ N, that is 2k−1 ≤ k!.
Then, for n = k + 1:
2(k+1)−1 = 2k
= 2(2k−1 )
≤ 2(k!), by the inductive hypothesis,
≤ (k + 1)k!, as 2 ≤ k + 1,
= (k + 1)!

3
Note: Remember it is some k – an arbitrary single k – not all k.

https://doi.org/10.1017/CBO9780511808258.025 Published online by Cambridge University Press

 Examples of induction 171

This shows that the statement is true for n = k+1. Hence, by the Principle of Mathematical
Induction, the statement is true for all n ∈ N.
Note that we took one side of the equality in the statement for n = k + 1 and played
with that. You could just as easily start with the other side, just as long as you don’t assume
A(k + 1).

Let’s try something which is more of a problem and which allows us to exercise our brains
with some real mathematical thinking.

Example 24.8
Find a formula for the sum of the first n odd numbers.
We have two problems wrapped up in one. First we must find the formula by some
method and then prove it holds for all n (which, it should come as no surprise, will be by
induction).
For the first part we can refer to our problem-solving methods (see Chapter 5). A good
method of attack is to rewrite the problem in symbols. In symbols the sum of the first n
Pn
odd numbers is i=1 (2i − 1). Let’s be good mathematicians: try the first few cases.
These results are tabulated below.
n 1 2 3 4 5
Pn
i=1 (2i − 1) 1 4 9 16 25

It doesn’t take much to see what the pattern is: The sum of the first n odd numbers is
n 2 . Well, that’s what it looks like – it could be that the next one doesn’t fit the pattern.
Nonetheless, we have a good conjecture to pursue:
n
X
(2i − 1) = n 2 for all n ∈ N.
i=1

This is a nice, simple formula which we have guessed from investigating the first few
cases. This is truly thinking like a mathematician. Of course, we need to show that it is
true!
The initial step has been done in our tabulated calculations. Let’s do the inductive step.
Pk
Assume the statement is true for some k, i.e. i=1 (2i − 1) = k 2 . Then,
k+1 k
!
X X
(2i − 1) = (2i − 1) + (2(k + 1) − 1)
i=1 i=1
= k 2 + (2k + 2 − 1), by the inductive hypothesis,
= k 2 + 2k + 1
= (k + 1)2 .

Thus, if the kth statement is true, then so is the (k + 1)th statement. Hence, our formula
is true by the Principle of Mathematical Induction.
Notice that once again in the summation for the k + 1 case all we do is split off the last
summand and follow by a simple application of the k case.

https://doi.org/10.1017/CBO9780511808258.025 Published online by Cambridge University Press

 172 C H APTER 24 Techniques of proof IV: Induction

Remark 24.9
In writing the above a mathematician would remove all mention of the calculations of
the initial few cases that led to the formula. Instead he or she would state the formula
and proceed to show it was true. This is known as covering your tracks because all
information on how we got there has been removed.

How to spot an induction proof

Obviously, to spot that induction could be used you must be suspicious of any statement
indexed by the set of natural numbers. For example, ‘For every natural number . . . ’, or
‘. . . for all n ∈ N’.
Sometimes the statements can be disguised. For example, consider the exercise

‘Show that n 2 − 1 is divisible by 8 when n is an odd natural number.’

At first glance this does not appear to be indexed by the natural numbers as none of the
even numbers are used. However, the statement is indexed by the naturals as the set of
odd numbers can be matched up with the naturals. That is, 1 is the first odd number, 3 is
the second, 5 is the third, and so on. We could rewrite the statement as

‘Show that (2n − 1)2 − 1 is divisible by 8 when n is a natural number.’

How to write an induction proof

The method for writing a proof by induction is simple:

(i) Announce that you are using induction.

(ii) Do the initial case.
(iii) State that you are assuming that the statement is true for some k. Writing out the
statement to use it later is often helpful.
(iv) Use the truth of the statement for k in the proof of the statement for k + 1. Often this
will mean breaking a mathematical expression into two pieces one of which involves
the case for k. Be sure to indicate at which point you use the inductive hypothesis.
(v) State the conclusion: ‘By the Principle of Mathematical Induction the statement is
true.’ That way, the reader knows the proof is over.

Exercises
Exercises 24.10
Pn 2
(i) Prove by induction that i=1 i = n6 (n + 1)(2n + 1).
(ii) Show that 2n ≤ 2n for all natural numbers n.
(iii) Prove that 32n − 1 is divisible by 8 for all natural numbers n.
(iv) Prove that 17 divides 34n + 43n+2 for all n ∈ N.
π
(v) Show that sin nx ≤ n sin x for all natural numbers n and 0 ≤ x ≤ 2.

https://doi.org/10.1017/CBO9780511808258.025 Published online by Cambridge University Press

 Exercises 173

(vi) Prove the Binomial Theorem, that is,

n  
X n n−r r
(x + y) =
n
x y for all n ∈ N.
r
r =0
   
n n!
Recall that = for all 0 ≤ r ≤ n.
r r !(n − r )!
(vii) Show that n 2 − 1 is divisible by 8 when n is an odd natural number.
(viii) Prove Leibniz’s Theorem for repeated differentiation of a product: If u and v are
functions of x, then prove that
dn
     
n n n
(uv) = u v
0 n + u v
1 n−1 + u v
2 n−2 + · · · + u r vn−r
dx n 1 2 r
+ · · · + u n v0 ,
di u djv
for all n ∈ N, where u i and v j denote and respectively.
dx i dx j
(You will need to use
     
k k k+1
+ =
r −1 r r
but proving this identity is true should not pose much difficulty.)
n n
!2
X X
3
(ix) Prove that r = r .
r =1
 r =1
n 
(x) Show that n is an integer for all 1 ≤ r ≤ n − 1.
r
(xi) Let X be a finite set with n elements. Show that X has 2n distinct subsets.
(xii) This exercise will, in effect, generalize the result of Exercise (iii).
(a) Show that x n − 1 is divisible by x − 1 for all n ∈ N and where x 6 = 1 is a
natural number.
xn − 1
(b) Find a formula for and use induction to show it holds for all n.
x −1
(xiii) Consider the statement ‘A(n): 2n < 2n−1 ’. Prove that A(n) is false for all n ∈ N.
Show that the inductive step holds, i.e. that A(k) =⇒ A(k + 1). Note that this shows
that we need the initial case to be true to show the statement holds for all n. This
is like ‘The Moon is made of cheese implies that the Moon is a tasty snack’ from
Chapter 7. That is, ‘A(n) =⇒ A(n + 1)’ can be true even if both A(n) and A(n + 1)
are false.
(xiv) Prove the statement on negation of quantifiers from Chapter 11:
To negate a statement of the form Q 1 x1 Q 2 x2 . . . Q n xn P(x1 , x2 , . . . , xn ), where
Q i is ∀ or ∃ for 1 ≤ i ≤ n, we do the following:
(a) Change every ∀ to ∃ and every ∃ into ∀.
(b) Replace P by its negation.
Show that in |xi | ≥ | in xi |. (This is the generalized Triangle Inequality.)
P P
(xv)
(xvi) Suppose x1 , x2 , . . . , xn are non-negative real numbers. Prove that
x1 + x2 + · · · + xn √
≥ n x1 x2 . . . , xn .
n

https://doi.org/10.1017/CBO9780511808258.025 Published online by Cambridge University Press

 174 C H APTER 24 Techniques of proof IV: Induction

Summary
I The principle of mathematical induction is useful for proving statements indexed by
the natural numbers.
I Think of applying induction when you see statements indexed by the natural numbers.
I Induction is used extensively in mathematics, e.g. summations, inequalities, divisibility,
...
I First prove A(1), then show A(k) =⇒ A(k + 1) for some (arbitrary but fixed) k ∈ N.
I Remember, it’s some k, not all k.

https://doi.org/10.1017/CBO9780511808258.025 Published online by Cambridge University Press