Proof writing Page 1

Proof writing
Extra notes for Mathematics 114 (2020)

1 What is a proof ?
Up to now your exposure to mathematics has probably consisted mostly of being told what proper-
ties numbers have (e.g. 3 + 5 = 5 + 3) and practicing various techniques (e.g. solve x2 − x + 1 = 0).
You may have seen some theorems and proofs, but you may not have learned why proofs are nec-
essary and what theorems really are. A theorem is simply a mathematical statement and writing
it as a theorem is an assertion by the author that the statement is True. Normally the statement
of a theorem (the mathematical statement that is claimed to be True) is followed by a proof, an
explanation of why the statement is True.
Some statements are easily seen to be True, e.g. 12 > 0. They would not usually be phrased as
theorems. But some are more complicated, often asserting something for a large set of numbers,
e.g. ∀x ∈ R, x2 ≥ 0. This statement is perhaps still a bit simple to be a proof, but it illustrates a
common difference. A single example can often be verified by a calculation. But a theorem takes
many such examples, often infinitely many, and asserts that they all behave in the same way. The
proof then explains why there won’t be any funny numbers that behave differently.
Theorem 1. The sum of any three consecutive positive integers is a multiple of three.
You may be tempted to try a few cases: 1 + 2 + 3 = 6; 4 + 5 + 6 = 15; 10 + 11 + 12 = 33.
They are all divisible by three, so it seems to work. However, this doesn’t constitute a proof. The
examples are evidence for the Theorem, but they do not provide irrefutable proof. For instance,
what if your friend told you it happens to be false for 2 334 543, 2 334 544, 2 334 545? Would
you believe them? In mathematics, you need to be sure enough of your proof to immediately say:
“No, it is true for those numbers. You must have made a mistake.”
So, what does a proof look like? We shall give many examples of proofs in these notes so you
can get an idea of what they look like. Let us start with a proof of the Theorem above:

Proof of Theorem 1. Let the smallest of the three numbers be n. Then the next one is n + 1 and
the one after that is n+2. Then the sum of those three consecutive numbers is n+(n+1)+(n+2) =
3n + 3 = 3(n + 1). Note that n + 1 is an integer, so the sum is three times an integer. That means
that the sum is divisible by 3. 

Note that the n we have chosen can be any positive integer. In particular we could take
n = 2 334 543 and this exact argument would work for the three numbers your friend suggested.
We can say with absolute certainty that your friend is wrong.

Thus we can define a proof as an argument which explains with absolute certainty why a given
statement is True in all possible cases. It should not rely on the reader extrapolating from a few
examples. Any possible example they come up with must be covered in the argument.

2 Techniques in proof writing

2.1 Conditional statements (implications)
When we want to prove a statement like x = 2 ⇒ x2 = 4, we need to explain why it is True for
all values of x. So, we break it into 3 cases. If x = 2, then the statement becomes True ⇒ True,
which is True. If x = −2, then it becomes False ⇒ True which is True. And otherwise (if x 6= ±2),

Copyright c 2020 Stellenbosch University. All Rights Reserved.

Proof writing Page 2

it becomes False ⇒ False which is True. Thus we have explained why the statement is True for
all values of x.
You may find this strange. Why do we also care about the cases where x 6= 2? Actually we
don’t! If x 6= 2, then the premise (the P in P ⇒ Q) is False, and as we know, False ⇒ Q is always
a True statement. This explains our first strategy for writing proofs:
Strategy 1 (Direct proof). When proving a statement of the form P ⇒ Q, assume P and try to
explain why Q is True. (I.e. you don’t have to worry about the cases where P is False.)
Theorem 2. If x = 2, then x2 = 4.

Proof. We assume the premise is True, i.e. we assume that x = 2. Then we calculate that x2 = 4,
so the conclusion is True as well. (Thus we have successfully used the direct proof strategy to
prove the statement.) 

Theorem 3. If x is an even integer, then x2 is an even integer.

Remark: Such statements should always be interpreted as universal statements: For all even
integers x, the number x2 is even.

Proof. Suppose x is an even integer. That means that x = 2y, for some integer y. Then x2 =
(2y)2 = 4y 2 = 2(2y 2 ), which is even since it is 2 times the integer 2y 2 . 

In the above proof we start by rephrasing the premise in terms of the definition of even, i.e.
an integer x is even if it equals 2 times another integer. Note that we use a different symbol for
this other integer . So that gives us the first two sentences in the above proof. From then on we
can use some algebra to rewrite x2 as 2 times another integer, so we can conclude (again, by the
definition of even) that x2 is an even integer. That’s the conclusion of the theorem, which we have
now established to be true, so the proof ends here.
Theorem 4. The sum of two odd integers is an even integer.
Proof. An integer is odd if it can be written in the form 2n + 1 for some n ∈ Z. So the sum of two
odd integers is of the form (2n + 1) + (2m + 1), for some n, m ∈ Z. Now (2n + 1) + (2m + 1) =
2n + 2m + 2 = 2(n + m + 1), which is even. 

Notice that we used two different symbols, m and n, in the above proof, since we’re not
assuming that the two odd numbers are equal, which they would be if we used the same 2n + 1
for each number.

Exercises
1. Prove that if n is an integer divisible by 3, then n2 is an integer divisible by 3.
2. Prove that if n is an odd integer, then n2 is an odd integer.

3. Prove that if 0 < a < b, then a2 < b2 . (Hint: a2 < b2 is equivalent to b2 − a2 > 0.)
4. Prove that if a < b < 0, then a2 > b2 .
5. Prove that if m and n are odd numbers, then mn is also odd.

6. Prove that if m and n are even numbers, then mn is divisible by 4.

7. Prove that if x and y are rational numbers, then so is xy. (Hint: A number z is rational if
there exists integers a and b such that z = ab .)
Proof writing Page 3

2.2 Contradiction and contrapositive

In high school you may have come across proofs by contradiction.1 The idea is that you assume
what you want to prove is False, then deduce something ridiculous, like 0 = 1, which means that
your assumption was wrong. Thus what you set out to prove must be True.
For example, in circle geometry one proves that if a quadrilateral ABCD is cyclic (its four
vertices lie on a circle), then the opposite angles add up to 180◦ . The converse is proved by
contradiction: If the opposite angles add up to 180◦ , then the four points must have lain on the
same circle. In the proof one assumes that one of the points do not lie on the same circle as the
other three, and then conclude that it must have been on the circle after all. Since it cannot be
on the circle and not on the circle at the same time, there is a contradiction. This means that the
assumption that it is not on the circle was wrong. Let us do a simpler example:
Theorem 5. If x2 < 1, then x < 1.
Proof. Assume for a contradiction that the conclusion x < 1 is False, i.e. assume that x ≥ 1. Then
x2 ≥ x ≥ 1. But this contradicts the premise x2 < 1. (The number x2 cannot simultaneously be
greater or equal to 1 and less than 1.) Thus, our assumption that x ≥ 1 is wrong, and we conclude
that x < 1. 

Note that we could not start with x < 1, since x < 1 does not imply x2 < 1 (e.g. (−2)2 > 1).
In any case, it would not be appropriate to assume x < 1, since that would attempt to prove
Q ⇒ P , not P ⇒ Q! Something else to point out is that in this example, it is natural to start
with an assumption on x and try to deduce something about x2 (even if this deduction leads to a
contradiction). It is easier to start with x, and then go to x2 than to start with x2 and then go to
x.
Before doing another example, let us describe and analyse this proof strategy:
Strategy 2 (Proof by contradiction). When proving a statement of the form P ⇒ Q, assume P
and assume ¬Q and try to deduce that from this follows a statement which is always False. In
other words, try to prove that P ∧ ¬Q ⇒ False.
Let us analyze using a truth table what happens if we can prove that P ∧ ¬Q ⇒ False is a
True statement:
P Q P ∧ ¬Q P ∧ ¬Q ⇒ False
T T F T
T F T F
F T F T
F F F T

When trying to prove P ⇒ Q, we may assume that P is True. Thus, we only need to look at
the top two rows in the truth table. If we can now prove that P ∧ ¬Q ⇒ False is a True statement,
then we must be in the first row, i.e. Q must be True.
Theorem 6. For any integer n, if n2 is an even number, then n is an even number.
Remark. Note again that if we knew the parity of n, then it would be easy to deduce the parity
of n2 , but the other way around might be more difficult. In fact, if we did not know that n is an
integer, even that would be impossible to prove. (If x were a real number and x2 an integer, x
need not be an integer.)
Proof. Assume for a contradiction that “n is an even number” is False, i.e. assume that n is odd.
Then n2 is odd. But this contradicts our assumption that n2 was even. Thus our assumption was
wrong, and n must have been even all along. 

1 Some high school textbooks also call the method Reductio ad Absurdum, to reduce to the absurd.
Proof writing Page 4

A proof strategy that is closely related to proof by contradiction is using the contrapositive.
We define the contrapositive of the implication P ⇒ Q to be the implication ¬Q ⇒ ¬P . From
the following truth table we can see that an implication P ⇒ Q is equivalent to its contrapositive
¬Q ⇒ ¬P :

P Q P ⇒Q ¬Q ⇒ ¬P
T T T T
T F F F
F T T T
F F T T

Strategy 3 (Contrapositive). When trying to prove a statement of the form P ⇒ Q, prove instead
¬Q ⇒ ¬P , i.e. assume ¬Q and try to prove ¬P .
The reason a proof of the contrapositive statement is similar to a proof by contradiction is that
in both we assume ¬Q and try to prove something. In fact if we have proved ¬Q ⇒ ¬P , then a
proof by contradiction would be complete. Indeed if we assume P and ¬Q and deduce ¬P , then
both P and ¬P are True, a contradiction!
x2 + 1
Theorem 7. If = 2, then x 6= 1.
x3 + 2
Of course we can try to solve this equation, and then say that the solutions are different from
1. But this is quite painful and the contrapositive can simplify things enormously.
x2 + 1
Proof. We prove the contrapositive, i.e. that if x = 1, then 6= 2. Indeed, if x = 1, then
x3 + 2
x2 + 1 2
3
= 6= 2, and we are done. 
x +2 3

Theorem 8. If m and n are integers such that m + n is odd, then m 6= n.

Proof. We prove the contrapositive, i.e. that if m = n, then m + n is not odd. Indeed, if m = n,
then m + n = 2m, which is twice an integer. Then m + n is even, and therefore not odd. 

We include the contrapositive, since we want you to be aware of the existence of this proof
technique. However, we advise you to always use proof by contradiction, rather than the contra-
positive, in this course.

Exercises
1. Prove that if a < b and ac ≤ bc, then c ≥ 0.
1
2. Prove that if 0 < a < b, then a > 1b .

x5 + x4 + x3 + 5
3. Prove that if = 1, then x 6= 1.
x3 + 1
4. Prove that if x + 3y ≤ 5 and x > 2, then y < 1.
5. Prove that if x 6= 0 is rational and y is irrational, then xy is irrational.
6. Prove that if x2 + y = 5 and y 6= 1, then x 6= 2.
Proof writing Page 5

2.3 Proof by cases

Sometimes it is necessary to break a proof into different cases, because the same argument doesn’t
work for all values.
Theorem 9. For any integer n, the integer n2 + n is even.
Proof. We treat the cases where n is even and where n is odd separately.
Case 1 (n is even). If n is even, we can write n = 2m for some integer m. Then n2 + n =
(2m)2 + (2m) = 4m2 + 2m = 2(2m2 + m), which is again even.
Case 2 (n is odd). If n is odd, we can write n = 2m + 1, where m is an integer. Then
n2 + n = (2m + 1)2 + (2m + 1) = (4m2 + 4m + 1) + (2m + 1) = 4m2 + 6m + 2 = 2(2m2 + 3m + 1),
which is even.
Since any integer n must be either even or odd, we have covered all cases, and in all cases,
n2 + n is even. This finishes the proof. 

Let us analyse this strategy: For example, suppose we want to prove the statement P ∨Q ⇒ R.
If P ∨ Q is True, then either P or Q (or both) is True, but we don’t know which one. Therefore
to conclude that P ∨ Q ⇒ R we must show that P ⇒ R and that Q ⇒ R.
For example, say we want to prove “If x = 2 or x = 3, then x2 − 5x + 6 = 0.” We show that
x = 2 ⇒ x2 − 5x + 6 = 0 (which is clearly True) and that x = 3 ⇒ x2 − 5x + 6 = 0 (which is also
clearly True).
Another way to explain why it is necessary to prove both these statements is to use a truth
table to show that the statements P ∨ Q ⇒ R and (P ⇒ R) ∧ (Q ⇒ R) are equivalent.

Exercise: Show that the statements P ∨ Q ⇒ R and (P ⇒ R) ∧ (Q ⇒ R) are equivalent. (Hint:

You might use a truth table.)

Strategy 4 (Cases). To prove a statement of the form (P ∨ Q) ⇒ R do it by cases. First assume

P and prove R. Then assume Q and prove R.
Let us do some more examples. In the next example we have more than two cases.
Theorem 10. For any integer n, the number n2 is either divisible by 3 or leaves a remainder of
1 when divided by 3.

Proof. We break this up into three cases based on what the remainder is when n is divided by 3.
Case 1: n = 3m for some m. In this case n = 3m, so n2 = (3m)2 = 9m2 = 3(3m2 ), which is
divisible by 3.
Case 2: n = 3m + 1 for some m. In this case n = 3m + 1, so n2 = (3m + 1)2 = 9m2 + 6m + 1 =
3(3m2 + 2m) + 1. Since 3(3m2 + 2m) is 3 times an integer, 3(3m2 + 2m) + 1 leaves a remainder
of 1 when divided by 3.
Case 3: n = 3m + 2 for some m. In this case n = 3m + 2, so n2 = (3m + 2)2 = 9m2 + 12m + 4 =
3(3m2 + 4m + 1) + 1, which again leaves a remainder of 1 when divded by 3.

Any integer n must fall in one of these cases2 , so in all cases, n2 is either divisible by 3 or
leaves a remainder of 1 when divided by 3. 

Theorem 11. For all x ∈ R one has |x − 1| + |x + 1| ≥ 2.

Proof. The absolute value function |x| is defined piecewise. Thus we can use a different formula
when x ≥ 0 than when x < 0. Thus the function |x − 1| can be broken up into the cases x ≥ 1
2 Actually this would normally also require proof. This is a consequence of the Euclidean Division Algorithm.

But for the sake of simplicity we shall assume it without proof now.
Proof writing Page 6

and x < 1, while the function |x + 1| can be broken up into the cases x ≥ −1 and x < −1. The
points −1 and 1 break the real line into three cases:
Case 1: Assume x < −1. Then |x − 1| = −(x − 1) (since x − 1 < 0) and |x + 1| = −(x + 1) (since
x + 1 < 0). Thus |x − 1| + |x + 1| = −(x − 1) − (x + 1) = −2x. However, since we are in the case
x < −1, we can conclude that −2x > −2(−1) = 2. (Remember that the sign of the inequality
changes when you multiply with a negative number.)
Case 2: Assume −1 ≤ x < 1. Then x − 1 < 0, so |x − 1| = −(x − 1), while x + 1 > 0, so
|x + 1| = x + 1. Thus |x − 1| + |x + 1| = −(x − 1) + (x + 1) = 2, which is always greater than or
equal to 2.
Case 3: Assume x ≥ 1. Now x + 1 ≥ 0 and x − 1 ≥ 0, so |x − 1| + |x + 1| = (x − 1) + (x + 1) = 2x.
Since we have assumed x ≥ 1 in this case, we get 2x ≥ 2 and we are done.

Since any x ∈ R must satisfy one of these three cases, all cases are covered, and in all cases
|x − 1| + |x + 1| ≥ 2. 

Let us contrast this proof again with the “experimental approach” of setting x equal to various
values. One might try setting x = −2, −1, 0, 1, 2 and obtain the values 4, 2, 2, 2, 4. These values
are all greater than 2, so none of them contradict the statement. However, these values do not
give proof, since there are infinitely many other values we haven’t checked.
However, if we tried more values between −1 and 1 (say −0.8, −0.6, . . . , 0.6, 0.8) we would see
that they all give the value 2. One might then conjecture that this is the case for all x ∈ [−1, 1].
Then one can try to prove it as we did in Case 2 of the previous Theorem. The point here is that
while trying special values of x can never be a complete proof, it may help in getting the idea for
a complete proof. So, it isn’t useless to try special cases, but it isn’t enough for a proof.

Exercises
1. Use proof by cases to show that if x = −2 or x = 2, then x2 = 4.
2. Prove that for all integers n, the number n2 − 3n + 5 is odd.
3. Prove that for all integers n, the number n3 − n is divisible by 3.

4. Prove that for any integer n, either n2 is divisible by 4 or n2 leaves a remainder of 1 when
divided by 4.
5. Prove that for any integers m and n, if mn is odd, then both m and n are odd. (Hint: First
assume for a contradiction and then use cases.)

2.4 Biconditional proofs

Recall that a conditional statement is a statement of the form P ⇒ Q. Thus we call a statement
of the form P ⇐⇒ Q a biconditional statement. We often read it “P if and only if Q.” The
statement P ⇐⇒ Q is True if P ⇒ Q and Q ⇒ P are True.

Strategy 5 (Biconditional statements). To prove a statement of the form P ⇐⇒ Q, assume P

and prove Q. Then assume Q and prove P .
Theorem 12. Prove that x2 − 5x + 6 = 0 if and only if x = 2 or x = 3.
Proof. First we assume that x = 2 or x = 3. We use cases to prove that x2 − 5x + 6 = 0.
Case 1: (x = 2). In this case x2 − 5x + 6 = 4 − 10 + 6 = 0, so the conclusion is True.
Case 2: (x = 3). In this case x2 − 5x + 6 = 9 − 15 + 6 = 0, so again, the conclusion is True.

Now we assume x2 −5x+6 = 0. We know that for all x ∈ R, one has x2 −5x+6 = (x−2)(x−3).
(Multiply out the right hand side.) Thus (x − 2)(x − 3) = 0. We know that if the product of two
Proof writing Page 7

real numbers is 0, then one of them must be 0. Thus x − 2 = 0 or x − 3 = 0. Thus x = 2 or

x = 3. 

Theorem 13. For all integers n, the number n is divisible by 3 if and only if n2 is divisible by 3.
Proof. First assume that n is divisible by 3. Then there is some integer m such that n = 3m.
Then n2 = (3m)2 = 9m2 = 3(3m2 ). Since 3m2 is an integer, n2 is divisible by 3.
For the converse, we proceed by contradiction. Assume that n2 is divisible by 3, but n is not
divisible by 3. Since we are assuming n is not divisible by 3, n must leave a remainder of 1 or 2
when divided by 3. Let us do this in two cases:
Case 1: n = 3m + 1. Then n2 = (3m + 1)2 = 9m2 + 6m + 1 = 3(3m2 + 2m) + 1. Thus n2 is a
multiple of 3 plus 1. This contradicts our assumption that n2 is divisible by 3.
Case 2: n = 3m + 2. Now n2 = (3m + 2)2 = 9m2 + 12m + 4 = 3(3m2 + 4m + 1) + 1, which again
contradicts the fact that n2 is divisible by 3.

Thus we have proved (by direct proof) that if n is divisible by 3, then n2 is also; and (by
contradiction and cases) that if n2 is divisible by 3, then n is divisible by 3. Thus we have proven
the biconditional statement. 

Theorem 14. The inequality

x3 − x > 0
holds if and only if x ∈ (−1, 0) ∪ (1, ∞).
Proof. We first show that if x ∈ (−1, 0) ∪ (1, ∞), then the inequality holds. There are two cases:
Case 1 (x ∈ (−1, 0)). In this case x < 0, x + 1 > 0 and x − 1 < 0, so

x3 − x = x(x − 1)(x + 1)

which is the product of two negative and one positive factors, and therefore positive.
Case 2 (x ∈ (1, ∞)). In this case x(x − 1)(x + 1) is the product of three positive factors, and
thus positive.

Now we show that if x3 − x > 0, then x ∈ (−1, 0) ∪ (1, ∞). We prove the contrapositive:
we assume that x ∈ / (−1, 0) ∪ (1, ∞) and prove the opposite inequality: x3 − x ≤ 0. If x ∈ /
(−1, 0)∪(1, ∞), then x must lie in the complement of (−1, 0)∪(1, ∞) in R, namely (−∞, −1]∪[0, 1].
Let us break this into cases again:
Case 1 (x ∈ (−∞, −1)). In this case x(x − 1)(x + 1) is the product of three negative factors and
therefore negative.
Case 2 (x = −1). Then x(x − 1)(x + 1) = 0, so x(x − 1)(x + 1) ≤ 0 is True.
Case 3 (x = 0). Then again x(x − 1)(x + 1) = 0.
Case 4 (x ∈ (0, 1). Now x(x − 1)(x + 1) is the product of two positive and one negative factors,
and therefore negative.
Case 5 (x = 1). Here again x(x − 1)(x + 1) = 0.

We have exhausted all possible values in (−∞, −1] ∪ [0, 1] and proved that in all cases the
inequality x3 − x ≤ 0 holds. Therefore we have established the contrapositive of what we wanted
to prove. 

Remark: There would be nothing wrong with grouping cases 2, 3 and 5 together in one case,
thereby saving yourself some writing.
Proof writing Page 8

Exercises
1. Prove that x2 − 2x − 3 = 0 if and only if x = 3 or x = −1.

2. Prove for any positive integer n, that n2 − 1 is divisible by 4 if and only if n is odd.
3. Prove for all real numbers x, y that x3 < y 3 if and only if x < y.
4. Prove that x(x + 1) ≥ 0 if and only if x ≤ −1 or x ≥ 0.

2.5 Counterexamples and existence proofs

A counterexample is an example that disproves a statement of the form “For all variables. . . ”.
For such a statement to be True, it needs to be True for all values of the variables. If it is False for
even one example, then it is False altogether. Such an example that proves the statement False is
called a counterexample.

Remember that to prove a theorem, it is not enough to give a handful of examples of numbers
n for which the statement is satisfied. Such examples are useful, for example to test possible
statements. If a statement seems to be true in a number of cases then one can start looking for a
proof.
However, a single counter-example is enough to show that a statement is false. For example
the statement √
∀x ∈ N, x∈Q
is false. We see that because we know of one example, namely x = 2, where it breaks down. So it
is often much easier to disprove a false statement than to prove a true statement.

Examples:
• “If (x − 2)(x + 2) = 0, then x = 2.” Recall that such a conditional statement is usually
interpreted as “For all x ∈ R, if (x − 2)(x + 2) = 0, then x = 2.”
Intuitively, this statement is False, since another solution to the equation is x = −2. We can
use this intuition to construct a counterexample. Set x = −2. Then (−4)(0) = 0 is True,
while −2 = 2 is False, so the statement (x − 2)(x + 2) = 0 ⇒ x = 2 becomes True ⇒ False,
which is False. Thus, x = −2 is a suitable counterexample.
In this example, there was only one counterexample — the one constructed. Sometimes
there will be more possibilities, but constructing one is enough to prove that the statement
is False.
• “If x > 2, then x > 5.” To construct a counterexample we must find some x which satisfies
the premise x > 2, but not the conclusion x > 5. Thus we must take x ∈ (2, 5].
We claim that x = 3 is a counterexample. In that case 3 > 2 is True, but 3 > 5 is False. Thus
the statement 3 > 2 ⇒ 5 > 2 is False. In other words, x = 3 provides a counterexample.
Note that x ∈ (2, 5] characterizes the set of counterexamples, but it does not constitute a
counterexample. A counterexample is a value of x (for which the statement is False).

Existence proofs are very similar to counterexamples, except that they prove why a certain
statement is True.
Proof writing Page 9

Examples:
• “There exists a real number x ∈ R such that x2 + x < 0.” If we take x = − 21 , then
x2 + x = 14 − 12 = − 14 < 0. Thus, there exists (at least one) such x.
Note that proving this statement was the same as giving a counterexample to its negation
“For all x ∈ R, x2 + x ≥ 0.”

• “There exists a real number y such that for all x ∈ R we have cos x > y.” We know that
the values of cos x lies in the interval [−1, 1], so all we need to do is choose y < −1. For
example, if y = −2, then cos x > y for all x. This proves the existence of some y with the
desired property.
• An advanced variant of existence proof (and therefore not required for this course) is a non-
constructive proof, i.e. an explanation of why an object (like a number) with certain
properties exists without giving an example. For example, the Mathematics 114 group
consists of about 500 students. Therefore two of you must share the same birthday. That is
because if all of you had different birthdays, there could be at most 366 of you — one for
each day of the year (including 29 February). However, there is no way of quickly saying
which two of you that is!
(A more spectacular variant of this argument can be used to prove that there are two
non-bald people in Cape Town with the same number of hairs on their head!)

Exercises
1. Find counterexamples to the following statements and explain why they are counterexamples:
(a) For all x ∈ R, one has x2 > 0.
(b) If x2 > 0, then x > 0.
(c) If xy = 0, then x = 0.
(d) If xy = 1, then x = 1 or y = 1.

2. Prove that there exists x ∈ R such that x 6= 1 and x2 = 1.

3. Prove that there exists a real number x such that x < 1 and x2 > 1.
4. Prove that for each y ∈ R, if y > 0, then there exists x ∈ R such that x2 = y. (Hint: You
may do this by giving a formula for x in terms of y.)

5. Prove that, for any positive integer n, there exists an integer k such that

k 2 − 1 = n(n + 1)(n + 2)(n + 3).

2.6 Mathematical Induction

What is the sum of the first n odd positive integers? Well, if we experiment we see

1=1
1+3=4
1+3+5=9
1 + 3 + 5 + 7 = 16

It seems that the pattern is that the sum of the first n odd numbers is n2 . But how do we know
that this pattern continues?
Take another example: For which integers n is n2 − n + 41 a prime number? If we set n equal
to 1, 2, 3, . . . , 10 in succession we get 41, 43, 47, 53, 61, 71, 83, 97, 113, 131, all of which are prime.
Proof writing Page 10

Does this mean that we have a pattern, and that n2 − n + 41 is prime for all integers n? No! If we
set n = 41 we get 412 − 41 + 41 = 412 which has a factor of 41. In fact 41 is the smallest positive
integer for which n2 − n + 41 is not prime!3
So, how do we know if a pattern continues or not? There is a proof technique called Induction
which allows us to prove that a pattern always continues. The idea is to explain that if the pattern
holds at the n-th step, it must also hold at the n + 1-st step. Therefore the pattern will always
continue.
In the first example above, what does it mean for the statement to be True at the n-th step?
It means that
1 + 3 + 5 + · · · + (2n − 1) = n2 ,
since the n-th odd number is 2n − 1. Now, what happens in the (n + 1)-st step? We add one more
number to the left hand-side — the (n + 1)-st odd number, 2(n + 1) − 1 = 2n + 1. Thus the left
hand-side becomes
1 + 3 + 5 + · · · + (2n − 1) + (2n + 1).
Since we are assuming that the statement is True at the n-th step, we can replace the 1 + 3 + · · · +
(2n − 1) by n2 . Thus, the left hand-side becomes

(n2 ) + (2n + 1).

Note that this factorizes to (n + 1)2 . But this means that the pattern continues! If it is True at
the n-th step, then it must also be True at the (n + 1)-st step. We can check that it is True at
step 1, which implies that it is True at step 2. But that means that it is True at step 3. And that
means that it is True at step 4. This process can be repeated to exhaust all positive integers.
In the second example above, what did it mean for the statement to be True at step n? It
means that
n2 − n + 41
is prime. What happens at the (n + 1)-st step? Then the expression becomes

(n + 1)2 − (n + 1) + 41 = n2 + 2n + 1 − n − 1 + 41 = n2 + n + 41.

If n2 −n+41 is prime, what can we say about whether n2 +n+41 is prime? Basically nothing, since
they have completely different prime factors. That is why the argument doesn’t work for the second
example — there is no way of connecting the statement at Step n with the statement at Step n+1.

You may ask: “How do we know that we get all the positive integers if we make this argu-
ment?” Let us make a list of all the integers for which the statement is True. If it is True for
n = 1, we put 1 in the list. But if 1 is in the list, 2 must be there as well. Then 3 must be
there, and then 4. We are essentially just counting up from 1. So, we must obtain every number
to which we can count. How do we know that all positive integers can be obtained by counting
from 1? Well, that is simply the way we think about numbers. It is the same as something like
x + y = y + x. Neither of them can be proved, but they seem so obvious that we just assume them.
We call these assumptions axioms. We need a few basic axioms, otherwise we wouldn’t be able
to prove anything. Later in the notes we shall say more about the role of axioms in mathematics.

Principle of Mathematical Induction: Let P (1), P (2), P (3), . . . be a sequence of statements,

one for each n ∈ N. Suppose that we know (can prove)

• P (1) is True;
• For each k ∈ N, P (k) ⇒ P (k + 1) is True.
3 The fact that n2 −n+41 is prime for n = 0, 1, . . . , 40 has a remarkably deep explanation using algebraic number

theory. Unfortunately this can only be treated at Honours level.

Proof writing Page 11

Then P (n) is True for all n ∈ N.

Now we can formulate a proof strategy that uses this principle.

Strategy 6 (Proof by Induction). To prove a statement of the form “For all positive integers n,
the statement P (n) is True.” prove the following:
• P (1) is True;
• Let k be an arbitrary positive integer. Assume that P (k) is True and using that, prove that
P (k + 1) is True.
If those steps are completed, the statement ∀n ∈ N, P (n) has been proven by Induction.
Note that this strategy can only be used for statements that are required to hold for all positive
integers, and not for statements about all real numbers or other objects. In the next example we
show how one writes out a proof by induction of the statement we encountered at the start of this
section.

Theorem 15. For each positive integer n we have

1 + 3 + 5 + · · · + (2n − 1) = n2 .

Proof. Let P (n) be the statement

1 + 3 + 5 + · · · + (2n − 1) = n2 .

Then P (1) is the statement 1 = 1, which is clearly True.

Now suppose, for an arbitrary k ∈ N that P (k) is True, i.e. that

1 + 3 + 5 + · · · + (2k − 1) = k 2 .

Then
1 + 3 + 5 + · · · + (2k − 1) + (2k + 1) = k 2 + (2k + 1)
using the Induction Hypothesis (the assumption that P (k) is True). Factorising the right hand-side
as (k + 1)2 we conclude that

1 + 3 + 5 + · · · + (2k − 1) + (2k + 1) = (k + 1)2 .

This is exactly P (k + 1). Thus, if P (k) is True, then P (k + 1) is also True. In other words, the
statement P (k) ⇒ P (k + 1) is True for all k ∈ N.
Thus we have proved
• P (1) is True;
• for all k ∈ N that P (k) ⇒ P (k + 1) is True.

Thus, the Principle of Mathematical Induction tells us that P (n) is True for all n ∈ N. 

Let us give two more examples of induction proofs that have been popular exam problems in
the past.
Theorem 16. Prove that for every positive integer n, the number n3 + 2n is divisible by 3.

Proof. We first prove the statement for n = 1. If n = 1, then n3 + 2n = 1 + 2 = 3, which is

divisible by 3. Thus the base case holds.
For the induction step assume that the statement is True for n = k, i.e. assume that k 3 + 2k
is divisible by 3. That is the same as saying k 3 + 2k = 3m for some integer m. Now we proceed
Proof writing Page 12

to prove that the statement is True for n = k + 1. We need to prove that (k + 1)3 + 2(k + 1) is
divisible by 3. We have

(k + 1)3 + 2(k + 1) = (k 3 + 3k 2 + 3k + 1) + (2k + 2)

= k 3 + 3k 2 + 3k + 2k + 3
= (k 3 + 2k) + (3k 2 + 3k + 3)
= 3m + 3(k 2 + k + 1)
= 3(m + k 2 + k + 1).

We used the induction hypothesis in the second to last line. Now note that the final expression
is 3 times an integer, proving that (k + 1)3 + 2(k + 1) is divisible by 3. Thus, by Mathematical
Induction, n3 + 2n is divisible by 3 for all positive integers n. 

Theorem 17. Prove that if n ≥ 4 is a positive integer, then 2n ≥ n2 .

Proof. Here the statement only makes a claim for n ≥ 4, so our base case is n = 4. (Note that
the statement is False for n = 1, 3.) If n = 4, then the left hand-side is 24 = 16 and the right
hand-side is also 24 = 16, so the statement is True.
Now assume the statement for n = k, i.e. assume 2k ≥ k 2 . Then for n = k + 1 we have

2k+1 = 2(2k ) ≥ 2(k 2 ).

We need to show that 2k+1 ≥ (k + 1)2 . So, if we can show that 2(k 2 ) ≥ (k + 1)2 (for k ≥ 2), then
we are done. But 2k 2 − (k + 1)2 = k 2 − 2k − 1 is a quadratic expression which takes its minimum
value when k = 1 and then increases. Thus for k ≥ 4 we have k 2 − 2k − 1 ≥ 42 − 2 · 4 − 1 = 7 > 0.
Thus 2(k 2 ) − (k + 1)2 = k 2 − 2k − 1 ≥ 7 and thus 2k 2 ≥ (k + 1)2 + 7. We may now complete the
proof of the induction step: If k ≥ 4, then

2k+1 ≥ 2k 2 ≥ (k + 1)2 + 7 > (k + 1)2 .

By Mathematical Induction, 2n ≥ n2 for all n ≥ 4. 

Exercises
1. Use Mathematical Induction to prove that
n(n + 1)
(a) 1 + 2 + 3 + · · · + n = ;
2
n(n + 1)(2n + 1)
(b) 12 + 22 + 32 + · · · + n2 = ;
6
2. Let Fi be the Fibonacci sequence: F0 = 1, F1 = 1, F2 = 2, F3 = 3, F4 = 5, F5 = 8, F6 = 13,
F7 = 21, where each entry is the sum of the previous two. In symbols Fi+2 = Fi+1 + Fi .
Use Mathematical Induction to show that
(a) F0 + F1 + F2 + F3 + · · · + Fn = Fn+2 − 1.
(b) F02 + F12 + · · · + Fn2 = Fn Fn+1 .
3. Use Mathematical Induction to show that the following divisibility relations hold:
(a) n2 − n is always divisible by 2;
(b) n5 − n is always divisible by 5;
(c) xn − 1 is divisible by x − 1 (as polynomials).
4. Use Mathematical Induction to show that the following inequalities hold:
Proof writing Page 13

(a) 2n ≥ 2n + 4 for all n ≥ 4.

(b) 3n ≥ 4n + 2 for all n ≥ 3.
(c) 5n ≥ 4n + 9 for all n ≥ 2
(d) (1 + x)n ≥ 1 + nx for all x ∈ [−1, ∞) and all n ∈ N.
5*. Use Mathematical Induction to show that

1 · 1! + 2 · 2! + 3 · 3! + · · · n · n! = (n + 1)! − 1,

where k! = 1 × 2 × · · · × k.

3 Advice
3.1 On mathematical writing
3.1.1 Use full sentences and correct grammar
A proof is a piece of writing that needs to convey an argument. Writing is done so that someone
can later read it, so therefore one needs to write in a way that makes it easy for the reader to
understand what you mean. The first way to do this is to use full sentences and to explain what
you are doing. Compare the two proofs below to see how much easier it is to read when there are
some words present:
Theorem 18. The tangent to the graph of the function y = x2 at (1, 1) has equation y = 2x − 1.
d 2
Proof 1. dx x = 2x
x = 1 means gradient = 2
Thus y = 2x + c.
Thus 1 = 2(1) + c, so c = −1.

So y = 2x − 1. 

Proof 2. By definition, the gradient of the tangent line is equal to the derivative evaluated at
x = 1. Since the derivative of x2 is 2x, the gradient of the required tangent line is 2(1) = 2. This
means that the tangent line has the form y = 2x + c, where c is a constant. To calculate c, we
note that the line passes through the point of tangency (1, 1). Therefore the point (x, y) = (1, 1)
must satisfy the equation y = 2x + c, in other words 1 = 2(1) + c. Solving this for c gives c = −1.
We conclude that the equation of the tangent line is y = 2x − 1. 

Can you see that it is much easier to understand the proof if more words are used? However,
it takes significantly more writing. Let us attempt a shorten the second solution a bit:
Proof 3. The derivative of x2 is 2x. Therefore the gradient of the tangent line is 2(1) = 2. The
general form of a line with gradient 2 is y = 2x + c, where c is a constant. We solve for c by
replacing x = 1, y = 1 in that equation to get 1 = 2(1) + c. We obtain c = −1 and conclude that
the tangent line has equation y = 2x − 1. 

3.1.2 Use symbols, but not too many

Symbols can help the reader understand something, but too many symbols could have the opposite
effect. Up to the 17th century mathematicians did not use symbols as much as today, and one
saw many statements like:

It is impossible to separate a cube into two cubes, or a fourth power into two fourth
powers, or in general, any power higher than the second, into two like powers.
Proof writing Page 14

Today, with the help of symbols one can rephrase this as:

If n > 2 is an integer, and x, y, z are non-zero integers, then xn + y n 6= z n .

On the other hand, an overuse of symbols can lead to something like:

Let a ∈ A, satisfy ∀x ∈ A, x ≥ a.

Everyone would understand better if one wrote:

Let a be the smallest element of A.

Thus, use your judgment to decide when symbols will simplify a thought and when words be a
clearer way to say what you want to say.

3.1.3 Do not start sentences with symbols

A proof needs to be clear, so ambiguity must be avoided. Consider the following example:

In a right-angled triangle with hypotenuse 1, the side opposite angle θ has length
sin θ. cos θ is the length of the other side.

The problem is that it is not clear that the period signals the end of a sentence. One might wonder
whether it denotes the product of two functions instead. An improvement is:

In a right-angled triangle with hypotenuse 1, the side opposite angle θ has length sin θ.
The third side has length cos θ.

3.2 How do I come up with a proof ?

Proofs are found4 by human beings, and you can learn to do this, too!
There are two very different stages in finding a proof of a theorem.
The first stage is creative (and uses mostly the right brain hemisphere). It involves coming up
with some clever ideas, some “reasons for which the theorem is true”. This stage is quite hard
to teach, and some people are naturally better at this than others, but anybody can improve this
skill through practice and experience.
The second stage is very logical and pedantic (and uses mostly the left brain hemisphere). It
involves translating the clever ideas from the first stage into a sequence of strict logical steps.
These steps must be written in such a way that the reader can verify each step, and be satisfied
at the end that the proof is correct and thus the theorem is true. This stage is a technical skill
which can be learned by almost anybody.
The exact amount of detail required at the second stage depends very much on the intended
reader.
At the one extreme, one might want to give enough detail, make the “distance” from each step
to the next so small, that the proof can be verified by a computer. Such a proof is called a formal
proof, and is usually so long (and painful to read), that formal proofs are rarely written down in
practice.
At the other extreme are proofs found in research articles. Since the intended readers are
expert researchers in that particular area of mathematics, one can expect the readers to take
fairly large steps, and basically be able to fill in much of the detail themselves. This is also done
to make the article as short as possible (so that a journal is more likely to agree to print it).
If the intended reader is a student (or the professor marking an exam!), then the level of detail
required falls between these two extremes. When you write down a proof, you should give enough
detail that your classmates can follow every step. This is the level of detail presented in the
example proofs in these notes.
4 Whether proofs are created or discovered is a philosophical question that we will not address here.
Proof writing Page 15

When reading a proof, try to understand the underlying ideas behind the proof (the clever
ideas from the first stage above). These are often obscured by the pedantic succession of logical
steps, so this may require some effort. But when you have understood these ideas, you have
understood the proof. These ideas are what’s worth remembering — once you have them, you
should be able to reconstruct the rest of the proof.

4 Axioms (optional)
Any chain of implications has to start somewhere, so before we can prove anything, we must choose
statements which we will accept as true without proof. Such statements are called axioms.
Nowadays, we are not concerned about whether axioms are actually true in any metaphysical
sense. They are merely mathematical statements from which other mathematical statements can
be deduced. If they happen to model some aspect of physical reality, then that is practically
very useful, since every theorem deduced from such axioms will again say something true about
physical reality, which is bonus.
One must be very careful to choose axioms which do not contradict each other, even very
indirectly — we say that the axioms must be consistent (technically this means that it must
not be possible to prove both a statement and its negation). Unfortunately, in most interesting
cases it is impossible to prove the consistency of a set of axioms (this is Gödel’s famous Second
Incompleteness Theorem).

To illustrate what kind of statements axioms are, let us list the axioms of the real number
system. The first four (A1–A4) are about the addition of numbers, the next four (M1–M4)
are about the multiplication of numbers and D is about the inter-relation between addition and
multiplication (called distributivity). The next four (O1–O4) are about the order relations (less
than, greater than, etc.) between real numbers. These should mostly be familiar to you as
properties learnt at school.
A1. a + (b + c) = (a + b) + c
A2. a + 0 = a = 0 + a

A3. For every a there is some number −a such that a + (−a) = 0

A4. a + b = b + a
M1. a · (b · c) = (a · b) · c
M2. a · 1 = a = 1 · a

M3. For every number a, if a 6= 0, then there is some number a−1 such that a·(a−1 ) = 1 = a−1 ·a.
M4. a · b = b · a
D. a · (b + c) = a · b + a · c

O1. If x < y, then x + z < y + z

O2. If x < y and y < z, then x < z
O3. Exactly one of the following hold: x = y, x < y or x > y
O4. If x < y and z > 0, then xz < yz

The next axiom (C) is a little technical and we will not elaborate on it in this course. The
reason for this axiom is e.g. to distinguish between the rational numbers and the real numbers.
More precisely, the sets Q of rational numbers and R of real numbers both satisfy all the previous
axioms. But R also satisfies (C), while Q does not.
Proof writing Page 16

C. Every non-empty set that is bounded above has a least upper bound.
We have two more axioms, but these are about the positive integers, not about the real numbers.
They are the axioms of Induction (I) and the Well-Ordering Principle (W). Logicians can prove
that these axioms are equivalent, so we only need to accept one of them; we then get the other one
for free. We won’t prove this equivalence and you may accept both as True during this course.
I. The Principle of Mathematical Induction.
W. The Well-ordering Principle: Every non-empty subset of N has a least element.

What (W) means is, if A ⊆ N and A 6= ∅, then A contains an element which is smaller than
all other elements of A.
All of mathematics can be built up from axioms5 For example, consider the statement

For all x, y ∈ R, if xy = 0, then either x = 0 or y = 0.

This is also something that may have been considered as an axiom. However, the next theorem
shows that it follows from the other axioms, so it is not necessary to list it as well.
Theorem 19. For all x, y ∈ R, if xy = 0, then either x = 0 or y = 0.
Proof. Suppose that x, y ∈ R and that xy = 0.
Case 1 (x = 0). In this case the statement x = 0 or y = 0 is True and there is nothing to prove.
Case 2 (x 6= 0). Now assume that x 6= 0. (Our aim is to show that y = 0.) Since x 6= 0, Axiom
(M3) tells us that there exists a real number x−1 such that x−1 x = 1. Then, multiplying the
equation xy = 0 by x−1 on both sides we get:

x−1 xy = x−1 0,

which is the same as 1y = 0, which is the same as y = 0. Thus we have proved that y = 0.

Since the cases x = 0 or x 6= 0 covers all possibilities for x, and in all cases the statement x = 0
or y = 0 is True, we have completed the proof. 

In this course we won’t expect you to prove anything from the axioms. But this begs the
question: What are you allowed to assume when writing a proof? There is really no rule in this
case, as it depends much on the type of statement that needs to be proved. It is often advised
to write as if a typical fellow student is required to read and understand it. We hope that the
example theorems in these notes will give you an idea of the level of detail we expect.

5 To be fair, the axioms we have listed are not the axioms used to build mathematics. Those axioms are called

the Zermelo-Fraenkel axioms of set theory. However, they are too abstract for the first year level, and the axioms
we listed are a good substitute.