Proof Techniques

How to prove a statement

Formulating a proof of a statement is often difficult. In this paper we discuss how to formalize
a statement, what kind of proofs there are and how to start formulating a proof. One could
wonder why we care so much about writing arguments so precisely. Often, a picture or short
argument will give the feeling that the statement is true – why bother about making it formal
theorems and proofs?

The answer to this question lies in generalization. If you have a specific problem, it might not
need a formal proof to solve it. If someone else has another problem that is slightly related,
this problem might be solved in the same manner - or it might turn out to be not the case.
Now if that first person analyses his problem to see under which conditions the solution is
true, and publishes that, this knowledge is available for others. Then if someone else has a
problem, he can formulate his problem and search for publications about it – and will soon
find the solution of the other. Thus, when formulating a problem mathematically we all have
our problems in the same language and can thus compare. Something about not having to
re-invent wheels...

And why we need to prove the statement if it the answer is clear easy to see? First of all,
others should be able to verify your solution. You can present a very nice theorem which
solves all problems in the world – if it is not true, it is useless. So by giving a formal proof
someone else can easily verify that your statement is indeed true. Furthermore, he can study
your proof and modify it to prove another statement with different conditions. By reading and
analyzing yours, he does not have to start from scratch but can use the insights that have
already been developed. Thus at first it looks like unnecessary extra work, but it will save
time in the long run.

To see why this indeed works, consider the following example. Remember the chemistry
lessons at high school where a molecule equation was given and you had to determine the
right coefficients. Now if it included only a few particles they were easy to solve, but as soon
as they got larger the puzzling got heavier. Mathematicians came up with a more structured
approach to solve this to replace just trying: linear algebra. But it turns out that many other
problems can be solved with this theory as well: economic models of supply and demand
use it, social studies use it to represent who knows who in a group of people, physicists use it
to calculate how heat flows through objects, engineers use it to calculate the forces an object
can handle, and there are many more applications. Mathematics is the same, regardless of
the application at hand.

Finally, learning how to formulate and prove a statement helps you to structure thoughts
and make an argument solid. When statements get complicated and proofs long, the skill
you obtain is how to keep the overview and branch a problem into several smaller, easier
problems. This is a skill that is very useful in everyday life – it is no coincidence that many
mathematicians end up in high, influential positions in society where you would not have
expected them if you recall they studied theoretical concepts for at least five years. So
although the theory in this course might be difficult and useless at first sight, think of it as a
long-term investment.

Most mathematics you have learned so far usually had a recipe: there is a certain procedure
you have to follow, and as soon as you have learned that procedure you are able to do all of
the exercises. It might take some time to learn to procedure, but once you know how it works
it gets easy.

Proving a statement can be more difficult. It is like solving a puzzle where all the pieces are
all somewhere lying around. You first have to find out which pieces belong to yours, and then
you can put them together to get the desired result. There is no standard procedure that will
enable you to prove all statements that are given to you. There are however some guidelines
that will help:

1. Decide what is given and what you can assume about the problem. Decide what you
must show or find.

2. Write down the mathematical definitions of what you have to show, and the definitions
of the things you may use. These first two steps require some reading skills: what
exactly is asked? What exactly have we given already?

3. Without starting the formal proof yet, look at the problem and see if the statement
is reasonable. Things that might help here are making a picture, try an example, or
discuss the problem with others. Do not give up quickly on this part, as the other steps
will be easier as soon as you have the feeling the statement must be true. Conversely,
if you think the statement is not true, you can try to find a counterexample. Then if you
find it you are done, and if you cannot find it then think about why it is not possible to
find one.

4. If it is not clear quickly why the statement is true, flip through related theorems and
examples to see if they have something to do with what you are given or what you
need to show. Check if they can bring you closer to the result.

5. As soon as you have the feeling the statement is true, analyze why that is the case.
A proof often exists of several steps which lead from the givens to the desired result.
Write down these consecutive steps. It should be like a construction work: orderly,
systematic and with a reasonable pace. Thus, this is the step in which you actually
write the proof.
 6. Now that you have written down everything there is, take some distance and see if
it is a good argument. Are all steps justified? Is there something missing? Is the
conclusion what was asked for?

When studying a proof, be careful to accept all that is said without verifying the steps. At first
sight the following proof might be reasonable, but the result is not true at all:
Example 1. Theorem: 4=3.
Proof: Suppose that there exist numbers a, b, c such that:
a+b=c
This can also be written as:
4a − 3a + 4b − 3b = 4c − 3c
After reorganizing:
4a + 4b − 4c = 3a + 3b − 3c
Take the constants out of the brackets:
4(a + b − c) = 3(a + b − c)
Remove the same term right and left:
4=3

What went wrong?

Thus, be precise in your argument. In the Inclusion-Exclusion principle, we used the sum
rule. This was allowed, since the sets we applied them on were disjoint. If we do not check
conditions of theorems before we apply them, awkward things like Example 1 can happen.
In case you had not noticed the mistake there yet: The last step is only valid if a + b − c 6= 0.
However, we assumed a + b = c, so a + b − c = 0 for all a, b, c.

In the following pages, some often used proof techniques are discussed.

0.1 Counterexample

The counterexample is used to show that a theorem is not true. It is an example that satisfies
the conditions of the stated theorem, but the theorem does not hold for that particular exam-
ple. In that case, we have shown that the theorem is not true in general. This saves all the
work of formulating a proof, therefore it is often wise to start searching for a counterexample
if you are not convinced yet of the validity of the theorem.
Example 2. Consider the statement that all prime numbers are odd numbers. This is true
for most prime numbers, except for 2: that is an even number. Thus, we have found a
counterexample.

However, the counterexample can be used more widely. A theorem always relies on a set of
assumptions. If there are 1000 different conditions needed for one theorem, the result might
not be so useful. Conversely, if there is just little conditions needed, the theorem is much
more widely applicable: it is much stronger. Therefore, the counterexample can be used to
verify that all conditions are indeed necessary for the theorem.

Example 3. Statement:
√ p
If x ≥ 0, y ≥ 0 and x2 = y 2 , then x = y.

√ p
This statement has three conditions: x ≥ 0, y ≥ 0 and x2 = y 2 . Each of theses
conditions is necessary to make a true statement. If, for example, the first condition ”x ≥ 0”
is skipped, then a counterexample for the statement can be constructed by taking x = −1
and y = 1. The remaining conditions are satisfied , but not the conclusion: x 6= y .

0.2 Proof by contradiction

The idea behind this method is that we do not try to prove that the given theorem is true. In-
stead, we assume the theorem is false. Then we show that it has impossible consequences.
Thus, it cannot be that the theorem is false. Therefore, it must be true. An example of a
theorem that uses this is the Pigeonhole Principle:

Theorem 1. If m pigeons occupy n pigeonholes and m > n, then at least one pigeonhole
has two or more pigeons roosting in it.

Proof of Theorem 1. If the theorem is not true, then each pigeonhole has at most one pigeon
roosting in it. Since there are n pigeonholes, that means that at most n pigeons occupy a
pigeonhole. Thus, there are at least m − n > 0 pigeons that are not in a pigeonhole, which
is a contradiction with the given that there are m pigeons occupying the pigeonholes. Thus,
at least one pigeonhole has two or more pigeons roosting in it.
0.3 Direct proof

The previous methods proved that a theorem was not true, or showed that the theorem could
not be false. A more elegant approach is to show a theorem is to make use of a direct proof.
This is given by a straightforward combination of established facts. Often, other theories and
results are used. An example of a theorem that can be proved using a direct proof is the
following:

Example 4. Consider two even integers a and b. Then the sum a + b is an even integer
as well. Proof: if a and b are even integers, they can be written as a = 2x and b = 2y ,
where x and y are integers. So if a number is dividable by two, then it is even. Then
a + b = 2x + 2y = 2(x + y). Since all even numbers can be divided by two and a + b can be
divided by two, a + b is an even integer.

Other examples can be found in set theory. To prove that two sets are equal, we show that
each element of the one set is an element of the other set, and vice versa. The proofs of
Lemma ?? on page ?? and Theorem ?? on page ?? are also direct: an arbitrary but fixed
element is used to show that a property is true.

0.4 Proof by Cases

Sometimes it is difficult to state a proof directly, but are there several cases to distinguish.
These cases are then mostly easy to prove directly. Consider the following example:

Example 5. The absolute value of a number x is denoted by |x|. Thus if x = −5 then |x| = 5,
if x = 4 then |x| = 4 and if x = −3 then |x| = 3. Thus, the absolute value returns the positive
value of the number. Now prove that for every real number, x ≤ |x|. We will distinguish two
cases. If x is a real number, then we have x ≥ 0 or x < 0. If x ≥ 0, we have that x = |x|. If
x < 0, we have that |x| = −x. Furthermore, we know in that case x < 0 < |x|. Combining
the cases yields that x ≤ |x|.

0.5 Mathematical induction

Imagine a long table where you stand on the head end. On the table is a long, long, possibly
infinite row of blocks. Some of them are marked with a dot. Suppose that you know two
things:
 (a) The first block is marked with a dot.

(b) If the block with number n is marked with a dot, then the block with number n + 1 is
also marked with a dot.

What do these two statements imply? We know because of (a) that the first block is marked.
Since the first block is marked, because of (b) the second block must be marked as well.
Now that the second block is marked, we apply (b) again and know that the third block is
marked. When this argument is repeated, we know that also block 402 and block 385492
are marked with a dot. Thus, all blocks are marked with a dot.

A similar example uses domino blocks. If one block falls, the next block will fall as well. So
if you push the first block, all of the blocks will (if you placed them correctly) fall down. This
illustrates the principle of mathematical induction. It does not only work for blocks, but for
many more proofs:

Definition 1. Suppose that we have a function S(n) whose domain is the set of positive
integers. Suppose that

(a) S(1) is true;

(b) for all n ≥ 1, if S(n) is true, then S(n + 1) is true.

Then S(n) is true for every positive integer n. This is the principle of mathematical induction.
Step (a) is called the basis step, step (b) is the induction step.

This principle of mathematical induction is very useful when functions with integers are in-
volved. All the blocks mentioned had an integer number, and the corresponding function
would be whether the block would be marked or not. With the domino example, the number
of the stone would be the integer number that is used as input of the function, then the output
of the function would be whether the stone had fallen or not.

Example 6. Suppose we want to add the integers 1 to n. Thus, denote this sum as:
n
X
S(n) = i = 1 + 2 + ... + n (1)
i=1

We claim that for all integers n ≥ 1:

n(n + 1)
S(n) = (2)
2
This can easily be proven using mathematical induction. The basis step is to verify that (2)
is true for n = 1. This is the easy part:

1(1 + 1)
S(1) = =1 (3)
2
Now we do the induction step and assume that (2) is true for some integer k ≥ 1. Then we
will show that it is also true for k + 1. We will first use the definition (1) of S(k + 1), then apply
our knowledge (2) about S(k), and rearrange terms to see whether we can end up with what
we want to show:
k+1
X
S(k + 1) = i
i=1
= S(k) + k + 1
k(k + 1)
= +k+1
2
(4)
k 2 + k 2k + 2
= +
2 2
k 2 + 3k + 2
=
2
(k + 1)(k + 2)
=
2
Thus, we have shown that (2) is true for n = 1 and for all n + 1 if n is true. Thus, with
mathematical induction we have shown that (2) is true.

You might have known the previous result already, or have been able to reason why state-
ment (2) would be true. An example that is somewhat more difficult to reason but shown with
mathematical induction is the following:

Example 7. Show that 5n − 1 is divisible by 4 for all n ≥ 1. Start with the basis step for
n = 1: 51 − 1 = 4, and 4 is divisible by 4. Now we will assume that 5k − 1 is divisible by
4 and will try to show that 5k+1 − 1 is divisible by 4 as well. How to do this? First of all, we
should try to write 5k+1 − 1 in such a form that we can use the induction hypothesis. Thus,
we want to write:
5k+1 − 1 = 5k − 1 + something
This can be done in the following way:

5k+1 − 1 = 5 · 5k − 1 = 4 · 5k + 5k − 1

Since we know that 5k − 1 is divisible by 4, we only have left to determine whether 4 · 5k is

divisible by 4. This however is easy since there is a factor 4 before the 5k which makes 4 · 5k
divisible by 4. Thus, we have shown the induction step and the basis step. Therefore, the
statement is proven by mathematical induction.
0.6 Exercises

Proofs

Exercise 1. Show that the following statement cannot be true: If a and b are real numbers
and a2 > b2 , then a > b.

Exercise 2. Prove the following statement:

For all positive real numbers x and y , if the product xy exceeds 25, then x > 5 or y > 5.

Exercise 3. We will prove that the difference of any rational number and any irrational num-
ber is irrational.

a) Explain why a proof by contradiction can be useful in this case.

b) Suppose that there exist a rational number x and an irrational number y such that
(x − y) is rational. Explain that there must exist some integers a, b, c and d with b 6= 0
and d 6= 0 such that x = a/b and x − y = c/d.

c) Show that then y = (ad − bc)/bd

d) Now explain that y must be rational and that we can thus prove by contradiction that
the theorem is true.

Exercise 4. Use a direct proof to show that for all even integers n, (−1)n = 1.

Exercise 5. Use mathematical induction to show that the following is true for all positive
integers n:
n(n + 1)(2n + 1)
12 + 22 + 32 + . . . + n2 =
6