Proofs and Mathematical Reasoning

University of Birmingham
Author:

Supervisors:

Agata Stefanowicz

Joe Kyle
Michael Grove

September 2014

c
University
of Birmingham 2014

Contents
1 Introduction

2 Mathematical language and

2.1 Mathematics is a language
2.2 Greek alphabet . . . . . .
2.3 Symbols . . . . . . . . . .
2.4 Words in mathematics . .

symbols
. . . . . .
. . . . . .
. . . . . .
. . . . . .

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

6
6
6
6
7

3 What is a proof ?
9
3.1 Writer versus reader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.2 Methods of proofs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.3 Implications and if and only if statements . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4 Direct proof
4.1 Description of method
4.2 Hard parts? . . . . . .
4.3 Examples . . . . . . .
4.4 Fallacious proofs . .
4.5 Counterexamples . . .

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

11
11
11
11
15
16

5 Proof by cases
17
5.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.2 Hard parts? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.3 Examples of proof by cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
6 Mathematical Induction
6.1 Method . . . . . . . . . .
6.2 Versions of induction. . .
6.3 Hard parts? . . . . . . . .
6.4 Examples of mathematical

.
.
.
.

19
19
19
20
20

7 Contradiction
7.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.2 Hard parts? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.3 Examples of proof by contradiction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

26
26
26
26

. . . . . .
. . . . . .
. . . . . .
induction

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

8 Contrapositive
29
8.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
8.2 Hard parts? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
8.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
9 Tips
9.1 What common mistakes do students make when trying
9.2 What are the reasons for mistakes? . . . . . . . . . . .
9.3 Advice to students for writing good proofs . . . . . . .
9.4 Friendly reminder . . . . . . . . . . . . . . . . . . . . .

c
University
of Birmingham 2014

to present
. . . . . .
. . . . . .
. . . . . .

the proofs? . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .

.
.
.
.

.
.
.
.

.
.
.
.

31
31
32
32
32

10 Sets
10.1 Basics . . . . . . . . . . .
10.2 Subsets and power sets . .
10.3 Cardinality and equality .
10.4 Common sets of numbers
10.5 How to describe a set? . .
10.6 More on cardinality . . .
10.7 Operations on sets . . . .
10.8 Theorems . . . . . . . . .

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

34
34
34
35
36
37
37
38
39

11 Functions
11.1 Image and preimage . . . . . . .
11.2 Composition of the functions . .
11.3 Special functions . . . . . . . . .
11.4 Injectivity, surjectivity, bijectivity
11.5 Inverse function . . . . . . . . . .
11.6 Even and odd functions . . . . .
11.7 Exercises . . . . . . . . . . . . .

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

41
41
42
42
43
44
44
45

12 Appendix

c
University
of Birmingham 2014

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

47

Foreword
Talk to any group of lecturers about how their students handle proof and reasoning when
presenting mathematics and you will soon hear a long list of improvements they would wish for.
And yet, if no one has ever explained clearly, in simple but rigorous terms, what is expected it is
hardly a surprise that this is a regular comment. The project that Agata Stefanowicz worked on
at the University of Birmingham over the summer of 2014 had as its aim, clarifying and codifying
views of staff on these matters and then using these as the basis of an introduction to the basic
methods of proof and reasoning in a single document that might help new (and indeed continuing)
students to gain a deeper understanding of how we write good proofs and present clear and logical
mathematics. Through a judicious selection of examples and techniques, students are presented
with instructive examples and straightforward advice on how to improve the way they produce
and present good mathematics. An added feature that further enhances the written text is the
use of linked videos files that offer the reader the experience of live mathematics developed by
an expert. And Chapter 9, that looks at common mistakes that are made when students present
proofs, should be compulsory reading for every student of mathematics. We are confident that,
regardless of ability, all students will find something to improve their study of mathematics within
the pages that follow. But this will be doubly true if they engage with the problems by trying
them as they go through this guide.
Michael Grove & Joe Kyle
September 2014

c
University
of Birmingham 2014

Acknowledgements
I would like to say a big thank you to the Mathematics Support Centre team for the opportunity
to work on an interesting project and for the help and advice from the very first day. Special
gratitude goes to Dr Joe Kyle for his detailed comments on my work and tips on creating the
document. Thank you also to Michael Grove for his cheerful supervision, fruitful brainstorming
conversations and many ideas on improving the document. I cannot forget to mention Dr Simon
Goodwin and Dr Corneliu Hoffman; thank you for your time and friendly advice. The document
would not be the same without help from the lecturers at the University of Birmingham who took
part in my survey - thank you all.
Finally, thank you to my fellow interns, Heather Collis, Allan Cunningham, Mano Sivantharajah and Rory Whelan for making the internship an excellent experience.

c
University
of Birmingham 2014

Introduction

From the first day at university you will hear mention of writing Mathematics in a good style and
using proper English. You will probably start wondering what is the whole deal with words, when
you just wanted to work with numbers. If, on top of this scary welcome talk, you get a number of
definitions and theorems thrown at you in your first week, where most of them include strange notions
that you cannot completely make sense of - do not worry! It is important to notice how big difference
there is between mathematics at school and at the university. Before the start of the course, many of
us visualise really hard differential equations, long calculations and x-long digit numbers. Most of us
will be struck seeing theorems like a 0 = 0. Now, while it is obvious to everybody, mathematicians
are the ones who will not take things for granted and would like to see the proof.
This booklet is intended to give the gist of mathematics at university, present the language used and
the methods of proofs. A number of examples will be given, which should be a good resource for further
study and an extra exercise in constructing your own arguments. We will start with introducing the
mathematical language and symbols before moving onto the serious matter of writing the mathematical
proofs. Each theorem is followed by the notes, which are the thoughts on the topic, intended to give
a deeper idea of the statement. You will find that some proofs are missing the steps and the purple
notes will hopefully guide you to complete the proof yourself. If stuck, you can watch the videos which
should explain the argument step by step. Most of the theorems presented, some easier and others
more complicated, are discussed in first year of the mathematics course. The last two chapters give
the basics of sets and functions as well as present plenty of examples for the readers practice.

2
2.1

Mathematical language and symbols

Mathematics is a language

Mathematics at school gives us good basics; in a country where mathematical language is spoken,
after GCSEs and A-Levels we would be able to introduce ourselves, buy a train ticket or order a pizza.
To have a fluent conversation, however, a lot of work still needs to be done.
Mathematics at university is going to surprise you. First, you will need to learn the language to
be able to communicate clearly with others. This section will provide the grammar notes, i.e. the
commonly used symbols and notation, so that you can start writing your mathematical statements in
a good style. And like with any other foreign language, practice makes perfect, so take advantage
of any extra exercises, which over time will make you fluent in a mathematical world.

2.2

Greek alphabet

Greek alphabet - upper and lower cases and the names of the letters.

2.3

Symbols

Writing proofs is much more efficient if you get used to the simple symbols that save us writing long
sentences (very useful during fast paced lectures!). Below you will find the basic list, with the symbols
on the left and their meaning on the right hand side, which should be a good start to exploring further
mathematics. Note that these are useful shorthands when you need to note the ideas down quickly. In
general though, when writing your own proofs, your lecturers will advise you to use words instead of
the fancy notation - especially at the beginning until you are totally comfortable with the statements
if. . . , then. . . . When reading mathematical books you will notice that the word implies appears
more often than the symbol = .
7
c
University
of Birmingham 2014

A
B

E
Z
H

I
K

alpha
beta
gamma
delta
epsilon
zeta
eta
theta
iota
kappa
lambda
mu

T
Y

nu
xi
omicron
pi
rho
sigma
tau
upsilon
phi
chi
psi
omega

Table 1: Greek letters

Quantifiers
(universal quantifier)
(existential quantifier)

for all
there exists

Symbols in set theory

, ( or &

union
intersection
subset
proper subset
composition of functions

Common symbols used when writing proofs and definitions

implies
if and only if
is defined as
is equivalent to
such that
therefore
contradiction
end of proof

:=

: or |

E or
 or 

2.4

Words in mathematics

Many symbols presented above are useful tools in writing mathematical statements but nothing
more than a convenient shorthand. You must always remember that a good proof should also include
words. As mentioned at the beginning of the paper, correct English (or any other language in which
you are literate) is as important as the symbols and numbers when writing mathematics. Since it is
important to present proofs clearly, it is good to add the explanation of what is happening at each
step using full sentences. The whole page with just numbers and symbols, without a single word, will
nearly always be an example of a bad proof!
Tea or coffee?
Mathematical language, though using mentioned earlier correct English, differs
slightly from our everyday communication. The classic example is a joke about a mathematician,
8
c
University
of Birmingham 2014

who asked whether they would like a tea or coffee, answers simply yes. This is because or in
mathematics is inclusive, so A or B is a set of things where each of them must be either in A or in B.
In another words, elements of A or B are both those in A and those in B. On the other hand, when
considering a set A and B, then each of its elements must be both in A and B.
Exercise 2.1. Question: There are 3 spoons, 4 forks and 4 knives on the table. What fraction of the
utensils are forks OR knives?
Answer: Forks or knives means that we consider both of these sets. We have 4 of each, so there
8
are 8 together. Therefore we have that forks or knives constitute to 11
of all the utensils.
If we were asked what fraction of the utensils are forks and knives, then the answer would be 0,
since no utensil is both fork and knive.
Please refer to section 10, where the operations on sets are explained in detail. The notions or
and and are illustrated on the Venn diagrams, which should help to understand them better.

9
c
University
of Birmingham 2014

What is a proof ?
The search for a mathematical proof is the search for a
knowledge which is more absolute than the knowledge accumulated by any other discipline.
Simon Singh

A proof is a sequence of logical statements, one implying another, which gives an explanation of why
a given statement is true. Previously established theorems may be used to deduce the new ones; one
may also refer to axioms, which are the starting points, rules accepted by everyone. Mathematical
proof is absolute, which means that once a theorem is proved, it is proved for ever. Until proven
though, the statement is never accepted as a true one.
Writing proofs is the essence of mathematics studies. You will notice very quickly that from day
one at university, lecturers will be very thorough with their explanations. Every word will be defined,
notations clearly presented and each theorem proved. We learn how to construct logical arguments and
what a good proof looks like. It is not easy though and requires practice, therefore it is always tempting
for students to learn theorems and apply them, leaving proofs behind. This is a really bad habit (and
does not pay off during final examinations!); instead, go through the proofs given in lectures and
textbooks, understand them and ask for help whenever you are stuck. There are a number of methods
which can be used to prove statements, some of which will be presented in the next sections. Hard
and tiring at the beginning, constructing proofs gives a lot of satisfaction when the end is reached
successfully.

3.1

Writer versus reader

Kevin Houston in his book[2] gives an idea to think of a proof like a small battle between the
reader and the writer. At the beginning of mathematics studies you will often be the reader, learning
the proofs given by your lecturers or found in textbooks. You should then take the active attitude,
which means working through the given proof with pen and paper. Reading proofs is not easy and may
get boring if you just try to read it like a novel, comfortable on your sofa with the half-concentration
level. Probably the most important part is to question everything, what the writer is telling you.
Treat it as the argument between yourself and the author of the proof and ask them why? at each
step of their reasoning.
When it comes to writing your own proof, the final version should be clear and have no gaps in
understanding. Here, a good idea is to think about someone else as the person who would question
each of the steps you present. The argument should flow and have enough explanations, so that the
reader will find the answer to every why? they might ask.

3.2

Methods of proofs

There are many techniques that can be used to prove the statements. It is often not obvious at
the beginning which one to use, although with a bit of practice, we may be able to give an educated
guess and hopefully reach the required conclusion. It is important to notice that there is no one ideal
proof - a theorem can be established using different techniques and none of them will be better or
worse (as long as they are all valid). For example, in Proofs from the book, we may find six different
proofs of the infinity of primes (one of which is presented in section 7). Go ahead and master the
techniques - you might discover the passion for pure mathematics!

10
c
University
of Birmingham 2014

We can divide the techniques presented in this document into two groups; direct proofs and indirect
proofs. Direct proof assumes a given hypothesis, or any other known statement, and then logically
deduces a conclusion.
Indirect proof, also called proof by contradiction, assumes the hypothesis (if given) together with
a negation of a conclusion to reach the contradictory statement. It is often equivalent to proof by
contrapositive, though it is subtly different (see the examples). Both direct and indirect proofs may
also include additional tools to reach the required conclusions, namely proof by cases or mathematical
induction.

3.3

Implications and if and only if statements

If our hypothesis is about anything and everything and not about one
or more particular things, then our deductions constitute mathematics.
Thus mathematics may be defined as the subject in which we never know
what we are talking about, nor whether what we are saying is true.
Bertrand Russell

The formula A = B means A implies B or if A then B, where A and B are two statements.
Saying A = B indicates that whenever A is accepted, then we also must accept B. The important
point is that the direction of the implication should not be mixed! When A = B, then the
argument goes from A to B, so if A holds, then B does too (we cannot have A without B). On the
other hand, when we have that B is accepted, then it does not have to happen that A is also accepted
(so we can have B without A). This can be illustrated by the following example:
it is raining = it is cloudy.
Now, if the first statement is true (so it is raining), then we automatically accept that it is also cloudy.
However, it does not work the other way round; the fact that it is cloudy does not imply the rain.
Notice further, that accepted does not mean true! We have that if it is raining, then it is cloudy and
we accept both statements, but we do not know whether they are actually true (we might have a nice
sunny day!). Also, genuineness of the second statement does not give any information whether the first
statement is true or not. It may happen that the false statement will lead to the truth via a number
of implications!
If and only if, often abbreviated iff, is expressed mathematically A B and means that if A
holds, then B also holds and vice versa. To prove the theorems of such form, we must show the
implications in both directions, so the proof splits into two parts - showing that A B and that
B A. The proof of the statement
it is raining it is cloudy,
requires from us showing that whenever it is raining, then it is cloudy and showing that whenever it
is cloudy, it is always raining.
Necessary and sufficient.
A B means that A is sufficient for B;
A B means that A is necessary for B;
A B means that A is both necessary and sufficient for B.

11
c
University
of Birmingham 2014

Direct proof

4.1

Description of method

Direct proof is probably the easiest approach to establish the theorems, as it does not require
knowledge of any special techniques. The argument is constructed using a series of simple statements,
where each one should follow directly from the previous one. It is important not to miss out any steps
as this may lead to a gap in reasoning. To prove the hypothesis, one may use axioms, as well as the
previously established statements of different theorems. Propositions of the form
A = B
are shown to be valid by starting at A by writing down what the hypothesis means and consequently
approaching B using correct implications.

4.2

Hard parts?

it is tempting to skip simple steps, but in mathematics nothing is obvious - all steps of reasoning
must be included;
not enough explanations; I know what I mean is no good - the reader must know what you
mean to be able to follow your argument;
it is hard to find a starting point to the proof of theorems, which seem obvious - we often
forget about the axioms.

4.3

Examples

Below you will find the theorems from various areas of mathematics. Some of them will be new
and techniques used not previously seen by the reader. To help with an understanding, the proofs are
preceded by the rough notes which should give a little introduction to the reasoning and show the
thought process.

Theorem 4.1. Let n and m be integers. Then

i. if n and m are both even, then n + m is even,
ii. if n and m are both odd, then n + m is even,
iii. if one of n and m is even and the other is odd, then n + m is odd.
Rough notes. This is a warm-up theorem to make us comfortable with writing mathematical arguments. Start with the hypothesis, which tells you that both n and m are
even integers (for part i.). Use your knowledge about the even and odd numbers, writing
them in forms 2k or 2k + 1 for some integer k.

Proof.

i. If n and m are even, then there exist integers k and j such that n = 2k and m = 2j. Then
n + m = 2k + 2j = 2(k + j).

And since k, j Z, (k + j) Z. n + m is even.

ii. and iii. are left for a reader as an exercise.
12
c
University
of Birmingham 2014

Theorem 4.2. Let n N, n > 1. Suppose that n is not prime = 2n 1 is not a prime.
Rough notes. Notice that this statement gives us a starting point; we know what it
means to be a prime, so it is reasonable to begin by writing n as a product of two natural
numbers n = a b.
To find the next step, we have to play with the numbers so we receive the expression
of the required form.
We are looking at 2ab 1 and we want to factorise this. We know the identity
tm 1 = (t 1)(1 + t + t2 + + tm1 ).
Apply this identity with t = 2b and m = a to obtain
2ab 1 = (2b 1)(1 + 2b + 22b + + 2(a1)b ).
Always keep in mind where you are trying to get to - it is a useful advice here!
Proof. Since n is not a prime, a, b N such that n = a b, 1 < a, b < n. Let x = 2b 1 and
y = 1 + 2b + 22b + + 2(a1)b . Then

xy = (2b 1)(1 + 2b + 22b + + 2(a1)b )

b

=2 +2

2b

+2

2b

12 2
ab

3b

+ + 2

(substituting for x and y)

ab

23b 2(a1)b

(multiplying out the brackets)

(taking away the similar items)

= 2 1.

(as n = ab)

=2

Now notice that since 1 < b < n, we have that 1 < 2b 1 < 2n 1, so 1 < x < 2n 1. Therefore,
x is a positive factor, hence 2n 1 is not prime number.
Note: It is not true that: n N, if n is prime = 2n 1 is prime; see the counterexample of this
statement in section 4.5.

Proposition 4.3. Let x, y, z Z. If x + y = x + z, then y = z.

Rough notes. The proof of this proposition is an example of an axiomatic proof, i.e.
the proof that refers explicitly to the axioms. To prove the statements of the simplest
form like the one above, we need to find a starting point. Referring to axioms is often a
good idea.
Proof.
x+y =x+z
= (x) + (x + y) = (x) + (x + z)

(by the existence of additive inverse)

= ((x) + x) + y = ((x) + x) + z

(by the associativity of addition)

= (x + (x)) + y = (x + (x)) + z

(by the commutativity of addition)

= 0 + y = 0 + z

(by existence of additive inverse)

= y = z.

13
c
University
of Birmingham 2014

Proposition 4.4. x Z, 0 x = x 0 = 0
Rough notes. Striking theorem seen in the second year lecture! We all know it since our
early years, though now is the time to proove it! Again, we will refer to the axioms.
Proof.
x 0 = x (0 + 0)

(because 0 = 0 + 0)

= x 0 = x 0 + x 0

(by the distributivity)

= 0 + (x 0) = x 0 + x 0

(by the existence of zero)

= 0 = x 0

(by the cancellation)

Similarly, 0 = 0 x (try to prove it yourself!)

Theorem 4.5.

n2 +5
n+1

as n .

Rough notes. Before we proceed, we need to recap the definitions.

Definition 4.1. A sequence (of real numbers) is a function from N to R.
Definition 4.2. A sequence (an ) of real numbers tends to infinity, if given any A >
0, N N such that an > A whenever n > N .
The above definitions are the key to proving the statement. We follow their structure,
so we assume A being given and try to find N such that an > A whenever n > N .
Proving statements of this form is not very hard, but requires practice to be able to get
the expression of required form. We will play with the fraction to make it smaller,
which will prove that it tends to infinity.

Proof. Let an :=
an :=

n2 +5
n+1

and let A > 0 be given. Observe that

n2
n2 + 5

n+1
n+1
n2

n+n
n2
=
2n
n
=
2

(find an expression smaller than an by taking away 5 in the numerator)

(decrease an expression by increasing the denominator; holds as n N)
(adding ns in the denominator)
(cancelling ns in the numerator and denominator)

and n2 > A provided that n > 2A.

So, let N be any natural number larger that 2A. Then if n > N , we have an >
Therefore, an tends to infinity.

14
c
University
of Birmingham 2014

n
2

>

N
2

> A.

Lemma 4.6 (Gibbs Lemma). Assume (p1 , . . . , pm ) is probability distribution and let (r1 , . . . , rm )
r
X
X
X
1
1
be such that ri > 0 i and
ri 1. Then
pi log

pi log .
pi
ri
i=1
i
i
Rough notes. You will probably not see this lemma until year 3, although a year 1 student
should be able to prove it, as it requires only manipulations with logs and using simple
claims. Notice, that we want to show the equivalent statement:
X 
1
1
0.
pi log log
pi
ri
i
Then there are just laws of logs and a simple claim that help us to arrive at the statement.
CLAIM: for x > 0, ln x x 1.
This is illustrated in the sketch below, however the formal proof is given in the appendix.
y =x1

2
1

y = ln(x)
0

1
2
3

We now have all the tools needed to write proof of the lemma. See how it works step by
step and then check if you can do it yourself!

Proof. We want to show that

X
i


1
1
pi log log
0.
pi
ri

Write
X
i



1 X 
1
1
pi log log
=
pi log + log ri
pi
ri
pi
i
X 
ri 
=
pi log
pi
i

ri 
1 X
pi ln
=
ln m i
pi
(using the fact that logm a =


1 X  ri

pi
1
ln m i
pi


1 X
=
ri pi
ln m i
X 
1 X
=
ri
pi
ln m
i
i
| {z } | {z }
1

=1

0.
15
c
University
of Birmingham 2014

(log(a1 ) = log a)
(log a + log b = log(ab))

ln a
;
ln m

m is a base of logarithm)
(by the claim)

(multiplying out the brackets)

(by the assumptions)

4.4

Fallacious proofs

Example 4.7. Study the sequence of sentences below and try to find what went wrong. You can find
the answer in the footnote1 . We prove that 1 = 2.
a=b
= a2 = ab
= a2 + a2 = a2 + ab
= 2a2 = a2 + ab
= 2a2 2ab = a2 + ab 2ab
= 2a2 2ab = a2 ab
= 2(a2 ab) = a2 ab
= 2 = 1.

Example 4.8. This example will be similar to the previous one (again, we will prove that 2 = 1),
although it does not contain the same mistake. Try to find what goes wrong here and again, the
solution is given at the bottom of the page2 .

2 = 2
= 4 6 = 1 3
9
9
= 4 6 + = 1 3 +
4
4

3 2
3 2 
= 1
= 2
2
2
3
3
= 2 = 1
2
2
= 2 = 1.

Example 4.9. Below we will present the classic mistake

of assuming

true the statement which is yet

to be proved. The task is to prove that the statement 2 + 6 < 15 is true.

6 < 15

= ( 2 + 6)2 < 15

= 8 + 2 12 < 15

= 2 12 < 7
2+

= 48 < 49.

1 Notice that we assumed that a = b, so (a2 ab) = 0 and hence we cannot cancel these expressions out! The last
step is not correct, hence the proof is not valid.
2 The problem occurs when we take the square root of both sides. Remember that the square root function returns a

positive
output, however its input might come from a negative number raised to the power: |x| = x2 and therefore x =

x2 . You can check that taking the left hand side negative, we will actually arrive at the true statement.

16
c
University
of Birmingham 2014

It may seem that the above argument is correct as we have reached true statement (48 < 49), but
this is not the case. It is important to remember that
statement X = true statement

does NOT mean that statement X is necessarily true! We assumed that 2 + 6 < 15 is true,
where this is what we need to prove. Therefore, our implications are going in the wrong direction (go
back to section about the implications if you are still confused). Valid proof would be of the form
true statement = statement X
showing that X is true.
The proof above is not correct, however it is not totally useless! Check if you can reverse the
implications to obtain the proof we are looking for. Note that it is alright to write arguments in the
wrong direction when finding the proof but not when writing it in the final form.
Reversed implications would give a valid argument, however, presented in its final form might make
a reader wonder where did the idea of starting from 48 < 49 came from (looks pretty random).
Generally, the easier approach would be the proof by contradiction (see section 7).

4.5

Counterexamples

Having in mind a little writer - reader battle, we should be sceptical about any presented statement
and try to find a counterexample, which will disprove the conjecture. It may happen that the theorem
is true, so it is not obvious in which direction to go - trying to prove or disprove? One counterexample
is enough to say that the statement is not true, even though there will be many examples in its favour.
Example 4.10. Conjecture: let n N and suppose that n is prime. Then 2n 1 is prime.
Counterexample: when n = 11,
211 1 = 23 89.

Example 4.11. Conjecture: every positive integer is equal to the sum of two integer squares.
Counterexample: n = 3. All integer squares, apart from (1)2 , (0)2 , (1)2 , are greater than 3 and
we need only consider the situation when one of the squares is either 0 or 1. Neither (3 1), nor (3 0)
is an integer square. Hence result.

Example 4.12. Conjecture: every man is Chinese.

Counterexample: it suffices to find at least one man who is not Chinese.

17
c
University
of Birmingham 2014

Proof by cases

5.1

Method

Proof by cases is sometimes also called proof by exhaustion, because the aim is to exhaust all
possibilities. The problem is split into parts and then each one is considered separately. During
lectures you will usually see proofs containing two or three cases but there is no upper limit for the
number of them. For example, the first proof of the Four Colour Theorem used 1936 cases. Over the
time, mathematicians managed to reduce this number to over 600 - still lots!
It is often very useful to split the problem into many small problems. Be aware though, the more
cases, the more room for errors. You must be careful to cover all the possibilities, otherwise the proof
is useless!

5.2

Hard parts?

split the problem wisely - it is sometimes not obvious how to divide the problem into cases;
a big number of cases may result in skipping one of them in the proof - make sure each possibility
is included in your resoning!

5.3

Examples of proof by cases

Theorem 5.1. The square of any integer is of the form 3k or 3k + 1.

Rough notes. This is a simple example of the proof, where at some point it is easier to
split the problem into 2 cases and consider them separately - otherwise it would be hard
to find a conclusion. Start by expressing an integer a as 3q + r, (q, r Z) and then square
it. Then split the problem and show that the statement holds for both cases.
Proof. We know that every integer3 can be written in the form: 3q + 1 or 3q + 2 or 3q.
So let a = 3q + r, where q Z, r 0, 1, 2. Then
a2 = (3q + r)2 = 9q 2 + 6qr + r2 = 3 (3q 2 + 2qr) +r2
|
{z
}
Z as q,rZ

So let 3q 2 + 2qr := k, k Z. We have a2 = 3k + r2 . Now,

case I: if r = 0 or r = 1, we are done;
case II: if r = 2 = r2 = 4 and then a2 = 3k + 4 = 3k + 3 + 1 = 3(k + 1) + 1 which is in the required
form.

Theorem 5.2. Let n Z. Then n2 + n is even.

Rough notes. To show that the expression is even, it may be helpful to consider the cases
when n is even and odd - what does it mean?
Click here to see a video example.
3 The

proof is given in section Examples of Mathematical Induction

18
c
University
of Birmingham 2014

 CASE I: n is even (express it mathematically);

CASE II: n is odd;
now, the simple algebra should bring us to the required conclusion.
Proof. Exercise for a reader

Theorem 5.3 (Triangle Inequality). Suppose x, y R. Then |x + y| |x| + |y|.

Notes. To split the proof into small problems, we need to recall the modulus function,
which is defined using cases:

x
for x 0,
|x| =
x for x < 0.
Then, using the definition, carefully substitute x or (x) for |x|, depending on the case.
The triangle inequality is a very useful tool in proving many statements, hence it is worth
to study the proof and memorise the inequality - you will see it lots in the future.
Proof.
case I: x 0, y 0, so by the definition, |x| = x and |y| = y. Hence, x + y 0.
So
|x + y| = x + y = |x| + |y|
case II: x < 0, y < 0. So, |x| = x, |y| = y. Then x + y < 0.
So
|x + y| = (x + y) = x + (y) = |x| + |y|
case III: One of x and y is positive and the other is negative. Without loss of generality, assume that x is
positive (x 0so|x| = x) and y is negative (y < 0, |y| = y). Now we need to split the problem
into 2 subcases:
i. x + y 0,
So
|x + y| = x + y x + (y) = |x| + |y|
ii. x + y < 0,
So
|x + y| = x + (y) x + (y) = |x| + |y|

19
c
University
of Birmingham 2014

Mathematical Induction

6.1

Method

How to use it, when to use it? Mathematical induction is a very useful mathematical tool to
prove theorems on natural numbers. Although many first year students are familiar with it, it is very
often challenging not only at the beginning of our studies. It may come from the fact that it is not as
straightforward as it seems.
Formally, this method of proof is referred to as Principle of Mathematical Induction.

Principle of Mathematical Induction

Let P (n) be an infinite collection of statements with n N. Suppose that
(i) P (1) is true, and
(ii) P (k) = P (k + 1), k N.
Then, P (n) is true n N.

When constructing the proof by induction, you need to present the statement P (n) and then follow
three simple steps (simple in a sense that they can be described easily; they might be very complicated
for some examples though, especially the induction step):
INDUCTION BASE check if P (1) is true, i.e. the statement holds for n = 1,
INDUCTION HYPOTHESIS assume P (k) is true, i.e. the statement holds for n = k,
INDUCTION STEP show that if P (k) holds, then P (k + 1) also does.
We finish the proof with the conclusion since P (1) is true and P (k) = P (k + 1), the statement
P (n) holds by the Principle of Mathematical Induction.
Dominoes effect. Induction is often compared to dominoes toppling. When we push the first
domino, all consecutive ones will also fall (provided each domino is close enough to its neighbour);
similarly with P (1) being true, it can be shown by induction that also P (2), P (3), P (4), ... and so on,
will be true. Hence we prove P (n) for infinite n.

6.2

Versions of induction.
Principle of Strong Mathematical Induction
Let P (n) be an infinite collection of statements with n, r, k N and r k. Suppose that
(i) P (r) is true, and
(ii) P (j) = P (k + 1), r j k.
Then, P (n) is true n N, n r.

Changing base step. There are different variants of Mathematical Induction, all useful in slightly
different situations. We may, for example, prove a statement which fails for the first couple of values
of n, but can be proved for all natural numbers n greater than some r N. We then change the base
step of Principle of Mathematical Induction to
check if P (r) is true, for some r N00
20
c
University
of Birmingham 2014

and continue with the induction hypothesis and induction step for the values greater or equal than r.
More assumptions. In the hypothesis step, we are allowed to assume P (n) for more values of n
than just one. Sometimes to be able to show that the statement P (k + 1) is true , you may have to
use both P (k) and P (k 1), so assume that both of them are true. In this case the induction base will
consist of checking P (1) and P (2). It may also happen that we will deduce P (k + 1) once we assumed
that all P (1), P (2), . . . , P (k) hold.
Mixture The most complicated case would combine the last two, such that we start the induction
base for some r N and then prove that P (r), P (r +1), P (r +2), ..., P (k 1), P (k) imply that P (k +1).
Then by induction P (n) is true for all natural numbers n r.

6.3

Hard parts?

the induction hypothesis looks like we are assuming something that needs to be proved;
it is easy to get confused and get the inductive step wrong. Ethan Bloch [1] gives an example of
proof by induction which fails to be true - see exercise 6.5.

6.4

Examples of mathematical induction

Example 6.1. Show that 23n+1 + 5 is always a multiple of 7.

Notes. This is a typical statement which can be proved by induction. We start by checking
if it holds for n = 1. Then if we are able to show that P (k) = P (k + 1), then we know
that statement is true by induction.
Proof. The statement P (n) : 23n+1 + 5 is always a multiple of 7.
BASE (n=1)
231+1 + 5 = 24 + 5 = 16 + 5 = 21 = 7 3
P (1) holds.
INDUCTION HYPOTHESIS: Assume that P (k) is true, so
23k+1 + 5 is always a multiple of 7, k N.
INDUCTION STEP: Now, we want to show that P (k) = P (k + 1), where
P (k + 1) : 23(k+1)+1 + 5 = 23k+4 + 5 is a multiple of 7.
We know from induction hypothesis that 23k+1 + 5 is always a multiple of 7, so we can write
23k+1 + 5 = 7 x for some x Z
= (23k+1 + 5) 23 = 7 x 23

(multiplying by 23 )

= 23k+4 + 40 = 7 x 8
= 23k+4 + 5 = 56x 35
3k+4

= 2

+ 5 = 7 (8x 5)
| {z }
Z

21
c
University
of Birmingham 2014

(35 from both sides)

So 23k+4 + 5 is a multiple by 7 (P (k + 1) holds), provided that P (k) is true.

We have shown that P (1) holds and if P (k), then P (k + 1) is also true. Hence by the Principle
of Mathematical Induction, it follows that P (n) holds for all natural n.

Theorem 6.2 (De Moivres Theorem). If n N and R, then [cos() + i sin()]n = cos(n) +
i sin(n).
Notes. The well known De Moivres Theorem can be easily proved using Mathematical
Induction. We show it is true for n = 2, remembering the rule for the product of complex
numbers:
z1 z2 = r1 (cos 1 + i sin 1 ) r2 (cos 2 + i sin 2 )
= r1 r2 ((cos 1 cos 2 sin 1 sin 2 ) + i(sin 1 cos 2 + cos 1 sin 2 ))
= r1 r2 (cos(1 + 2 ) + i sin(1 + 2 )).
Then we assume the statement is true for n = k and use this assumption to show it holds
for n = k + 1. At the induction step we will need the trigonometric identities:
cos(A + B) = cos A cos B sin A sin B,
sin(A + B) sin A cos B + cos A sin B.

Proof. The theorem is true for n = 1, trivially.

BASE (n = 2):
z 2 = (cos + i sin )2 = (cos 2 + i sin 2)
INDUCTION HYPOTHESIS: assume that it is true for n = k, so
[cos() + i sin()]k = cos(k) + i sin(k).
INDUCTION STEP: Now,
(cos + i sin )k+1 = (cos + i sin )k (cos + i sin )
= (cos(k) + i sin(k))(cos + i sin )

(by the induction hypothesis)

= cos(k) cos + i cos(k) sin + i cos sin(k) sin(k) sin

(multiplying out the brackets)

= cos(k + ) + i sin(k + )

(follows from the trigonometric identities)

= cos(k + 1) + i sin(k + 1)

(taking outside the bracket)

Hence we have shown that P (1) and P (2) hold and k 2, P (k) = P (k + 1). Therefore P (n)
is true n 2 by the Mathematical Induction.

22
c
University
of Birmingham 2014

Proposition 6.3. Let an+1 =

1
5


a2n + 6 and a1 = 52 . Then (an ) is decreasing.

Notes. We have defined a sequence earlier and here is the definition of the decreasing
sequence.
Definition 6.1. A sequence (an ) is decreasing if an+1 an for all n N.
We will use the definition to prove the statement. Notice that we need to show an+1 an
for all n - this should suddenly bring to your mind induction.
As always, we start by checking the base, (here for n = 1) and then we assume that
P (n) is true for n = k. The hard part is usually the induction step, although it is not
very complicated here. That we want to show that ak+2 ak+1 using our previous
assumption.
Proof. We will show that the statement P (n) holds for all n.
P (n) : an+1 an for all n.
BASE:
an =
Note: a2 =

49
20

<

5
2

 49
 1  25
1  5 2
+6 =
.
+6 =
5 2
5 4
20

= a1 . Hence, P (1) holds.

HYPOTHESIS: Suppose that for some k 1, ak+1 ak .

INDUCTION STEP:
(ak+1 )2
6
+
5
5
6
(ak )2
+

5
5
= ak+1 .

ak+2 =

Hence ak+2 ak+1 .

Since P (1) is true and P (k) = P (k + 1), it follows that the sequence is decreasing by the Mathematical Induction.

Exercise 6.4. For which positive integers n is 2n < n! ?

Notes. Notice that this is a different example to the ones we have presented above. Here,
you must find n first and then show that it actually holds. You may want to check the
first couple of values of n and then formulate the statement P (n) for which you can use
Induction. Structure your proof as above, the notes on side should also help.
Click here to see a video example.

23
c
University
of Birmingham 2014

Proof. First, let us find the value of n for which we will prove the statement.
(does the statement hold for n = 1, 2, 3, . . . ? Have you
found n for which this is true?)

Let P (n) : 2n < n! be the statement. We will show that it holds for all n . . .
BASE STEP
(show P (n) holds for n smallest possible)

INDUCTION HYPOTHESIS

P (k) : . . .
(state the assumption for P (k))

INDUCTION STEP
(keep in mind what you are trying to prove - it helps
to note it on the side)
(hint: notice that 2 < k + 1 k > 1)

CONCLUSION
(finish the proof by writing the conclusion)

Exercise 6.5. Use Principle of Mathematical Induction to show that

 S
n
k=1

Ak

0

n
T
k=1

A0k n 2.

The above theorem is one of the De Morgans laws for an arbitrary collection of subsets (see section
on sets for De Morgans Laws in case of two subsets). Below there are two examples of the first
year students approach to prove this theorem by mathematical induction. Have a look at their work
(figures 1 and 2). Can you see which students work gained more marks and why? Are all the steps of
induction correct?

24
c
University
of Birmingham 2014

Figure 1: Proof by induction attempted by student A

Figure 2: Proof by induction attempted by student B

Two proofs, both written by first year students, are a good example to see why the induction is
hard. The fact that the argument looks as if it contains all the required steps, like base, hypothesis
25
c
University
of Birmingham 2014

and induction step, does not make it correct proof. We will now analyse both arguments and point
out where the problems are. You have probably spotted already that the proof written by student B
is much better. It received 4 out of 5 points, because the argument flows well and contains all required
parts of induction, though the induction step needs more explanation. You must also watch the details
closely, because the statement is proved for n 2, so this should be mentioned in our conclusion.
Student A received 0 marks for their work and the details are discussed below.
STUDENT B:

STUDENT A:
the argument starts with the wrong statement
- we want to state P (n) at the beginning, so
prove the equality in terms of n;
De Morgans Laws are necessary to show that
base step holds and it should be checked for
n=2, because we are proving the statement
for all n 2;

it is not necessary to check P (1) since we are

proving P (n) for n 2;
induction step is missing one line of argument, we should not jump to the required form straight away but show all
reasoning (the highlighted step is missing);

there is hypothesis stated clearly;

induction step is not stated properly - steps
are missing when deducing P (k + 1);
conclusion is incorrect as it was not in fact
shown that P (k) = P (k + 1).

in the conclusion P (k) P (k+1) true k

N, we should also add k 2.

Exercise 6.6. In Blochs book[1] we read an argument, which clearly fails at some point. It is hard
to detect the mistake though and it seems that induction is correct. See if you can spot a problem the answer is given in a footnote4 .

Proof. P (n) : in any collection of n horses, all of them have the same colour.
Since there are finite number of horses in the world, the statement means that all horses in the
world have the same colour!
BASE (n=1): P (n) clearly holds as in any group of only 1 horse, it is trivially true that all
horses have the same colour
HYPOTHESIS: Now we assume that P (k) holds, so in any group of k horses, it is true that all
of them have the same colour.
INDUCTION STEP: Now imagine a collection of k + 1 horses, lets call it {H1 , H2 , . . . , Hk+1 }.
Now, if we take the first k of them, then by induction hypothesis we know that they are all of the
same colour. We may also consider another set of k horses {H2 , H3 , . . . Hk+1 } which, again, are
all of the same colour. Since {H1 , H2 , . . . , Hk } and {H2 , H3 , . . . Hk+1 } all have the same colour,
then we may deduce that all k + 1 horses have the same colour. Hence all horses have the same
colour.
Since P (1) holds and P (k) = P (k + 1), we have that P (n) is true for all natural n by the Principle
of Mathematical Induction.
4 The problem lies in an inductive step which will fail for some particular value of n. So if we take n big enough,
then we may not be able to find set of n horses, where all would have the same colour

26
c
University
of Birmingham 2014

Contradiction

7.1

Method

Proof by contradiction is a very powerful technique, but the method itself is simple to understand.
When trying to prove
statement A = statement B,
assume that A is true and that not B is true and try to reach a contradiction.
The method is often used in proofs of the existence theorems, so the statements of the form there
is no x such that. . . . Here, instead of proving that something does not exist, we can assume that
it does and try to reach nonsense. We finish the proof with the word contradiction!, where some
people prefer the lightning symbol or a double cross (see section 2 on symbols) to indicate that they
reached the contradiction.

7.2

Hard parts?

when proving more complex theorems, it is easy to get confused and make a mistake. Then we
arrive at contradiction, which does not come from the original assumption but form the error in
the middle of the proof.

7.3

Examples of proof by contradiction

Theorem 7.1. Let a be rational number and b irrational. Then

i. a + b is irrational
ii. if a 6= 0, then ab is also irrational.
Notes. First of all we need to recall what it means to be rational (can be expressed as
a fraction) or irrational (cannot be expressed as a fraction). So if we want to show that
a + b is irrational, we do not really know how to describe it generally. Instead, we may
assume the opposite, express it as a rational number (which is easy to do in general) and
show that it leads to a contradiction. Notice how a big role the definitions play when
constructing the proof!
Proof. i. Suppose that a + b is rational, so a + b := m
n . Now, as a is rational, we can write it as
a := pq . So
m p
mq pn
b = (a + b) a =
=
,
n
q
nq
hence b is rational, which contradicts the assumption.
ii. left as an exercise

Exercise 7.2. Prove that

2+

6<

15

27
c
University
of Birmingham 2014

Notes. We have seen this statement before as an example of fallacious proof and now
we will show how to prove such expressions using contradiction. It has been shown that
it is easy to fall in the trap of assuming that the statement is true and then arguing from
there; it is popular mistake probably because there is no other starting point that seems
sensible. To avoid the mistake of assuming of what has to be proved, it is better suppose
the opposite of a given statement. If we reach a contradiction, then our assumption was
wrong and the statement is proven true.

Proof. Assume for a contradiction that 2 + 6 15

= ( 2 + 6)2 15

= 8 + 2 12 15

= 2 12 7
= 48 49
The
is clearly not true, hence we reached the contradiction. Therefore, we proved that
last
statement

2 + 6 < 15.

Theorem 7.3. Let f : A B and g : B C be functions. If g f is bijective, then f is injective

and g is surjective.
Notes. Refer to the section on functions to recall that bijective means injective and
surjective - have a look at the definitions in section 11, because again these will help to
construct the proof. Now, you may want to start with the statement g f is bijective
and follow from there, however it may be hard to conclude that f is injective and g
is surjective. The quickest way to bring us to the required statement is to assume the
opposite and try to reach the contradiction.
Proof. Suppose the statement does not hold, so f is not injective or g is not surjective. Let us consider
both cases:
f is not injective, which means that a, a0 A such that f (a) = f (a0 ).
Now,
(g f )(a) = g(f (a)) = g(f (a0 ))
so we have (g f )(a) = (g f )(a0 ) but a 6= a0 . So g f is not injective, hence it is not bijective.
g is not surjective, which means that c C such that for all b B, g(b) 6= c. Moreover, g f
is surjective, so a A such that (g f )(a) = c. Now, if b = f (a), then g(b) = c, which is a
contradiction!
Both cases lead us to the contradiction, hence we may conclude that if g f is bijective, then f is
injective and g is surjective.

28
c
University
of Birmingham 2014

Theorem 7.4. There are infinitely many primes.

Notes. Having seen many theorems and proofs, let us try to prove the famous theorem
stating the infinity of primes. This comment should guide you to writing the statements
in the mathematical language - do not worry if you dont get it first time; a final proof
often needs changing and polishing a few times! Click here to see a video example.
Proof. Suppose for a contradiction that. . .
(write the statement)

Since. . .

(insert the statament)

,we can list them: . . .

(list the primes - you need to pick a suitable notation)

Now, consider the number n, which is not a prime.

(you can define this number in many different ways, but you need a number which is not
a prime (consider multiplying all the listed primes by each other - then n is greater that
any of them) and that leads us to the contradiction (this is a tricky part) - we may want to
come back to this point later, because the next lines of the argument should help us to pick
appropriate n here)

Since n is not a prime,. . .

(what does it tell us? Write down what does it mean mathematically that n
is not a prime. Think about the factor of n - what if we take it the smallest
possible?)

Take the factor the smallest possible, so it is prime.

Now, it follows that z Z, such that

n = z . . .,
hence
z=

n
...
=
...
...

(1)
(2)

At this stage, we can get the contradiction to our assumption, but it depends on our choice of n. Let us
come back to the definition of n - what would guarantee it? What did we assume about the factor of n? What
assumptions did we state at the equation(1)? Try to change n slightly if your argument does not reach the
contradiction yet.

29
c
University
of Birmingham 2014

Contrapositive

8.1

Method

Proof by contrapositive is in essence a submethod of a proof by contradiction. The argument

begins in the same way in both cases, by assuming the opposite of the statement. So when showing
that
statement A = statement B,
we assume not B, but this time we argue to arrive at not A. The trick here is the fact that the
statements A = B and not B = not A are equivalent.
To understand this relationship better, you may want to read more about the Wason selection task,
which is the logic puzzle formulated by Peter Wason in 1966. The example below is based on his work
and it illustrates very well the reasoning used in proof by contrapositive.
Imagine four cards placed on the table, each with a letter on one face and a number on the other
one.
A D 9 5
You are given the rule which states: if A is on a card, then 5 is on its other side. Now, the task is
to indicate which card(s) need to be turned over to check whether the rule holds?
While most of the people give the automatic response A and 5, the correct answer is A and
9. Notice that according to our rule, if the card shows A on one face, it must have 5 on the
other. The rule however does not say anything about the card showing 5!
Hence, only checking A and 9 can test the rule:
if A does not have 5 on the other side, the rule is broken;
if D has (or not) 5 on the other side - it does not tell us anything;
if 5 has (or not) A on the other side - again, the rule is not broken;
if 9 does have A on the other side - the rule is broken.

8.2

Hard parts?

the method itself requires the knowledge of the fact A B is equivalent to not B not A;
similarly as in the proof by contradiction, the theorems proved may be complex and it is easy to
make mistakes and arrive at the incorrect conclusion.

8.3

Examples

Theorem 8.1. Let n Z. If n2 is odd, then n is odd.

Notes. The direct proof would not really work here, because writing n2 in the form 2k + 1
(for some k Z) does not really help to continue. Neither is the contradiction useful, as
we cannot find the required conclusion. In this case we need a special technique, which is
the contrapositive. This means that we start by assuming the opposite of the statement
(here: n is NOT odd) and then use the fact that
A = B is equivalent to: not B = not A.
In this example statement A is: n2 is odd and B: n is odd.
30
c
University
of Birmingham 2014

Proof. Let n be even (which is not B).

= n = 2k, k Z
= n2 = (2k)2 = 4k 2 = 2 2k 2
= n2 is even
(up to this point we proceed as we would in the proof by contradiction, although
the conclusion would be: n even = n2 even, which is not what we are trying
to prove)

So we proved that n is even = n2 is even. Now using the contrapositive we conclude that
n2 not even (odd) = n not even (odd),
which proves the statement.

Theorem 8.2. If mn is odd, then m and n are odd.

Notes. Representing mn as an odd integer does not really give us any tips on how to carry
on with our proof. Much easier thing is to assume the opposite of the second part of the
statement and see where we arrive at - is it contradiction? Or opposite of the first part
of the statement? We have the tools to arrive at the conclusion in any case, so let us see
where will our assumption get us to.
Click here to see a video example.

Proof. Assume that . . .

, so m and n are. . . .
(assume not B)

So m = 2 . . . , . . . Z and n = 2 . . ., . . . Z.

Now, mn = . . . . . . = . . .
(substitute for m and n)

Hence, we conclude that. . .

(you should arrive at not A)

So, by contrapositive, . . .
(remember not B = not A A = B)

31
c
University
of Birmingham 2014

Tips
Proofs are hard to create - but there is a hope.
K. Houston

This section has been created with help of lecturers from the University of Birmingham who took
part in a short survey, so the problems and common mistakes in areas of pure mathematics could be
detected. All the quotations throughout this section come from the short questionnaire. Many good
tips have also been found in [1] and [2] and the reader is strongly advised to look into this literature.
It is good to find out what the common mistakes are, and to watch out closely and try to avoid them.
Writing proofs is hard. It often requires a good amount of time and paper, before the neat proof
is finally produced. You will find yourself crossing things out and changing it many times before you
reach what is required. Making mistakes is natural and it is better to think about them as a step in
learning to write proofs. Writing proofs in pure mathematics university courses is different from what
has been done during school, so it will take some time to get used to doing this properly.

9.1

What common mistakes do students make when trying to present the

proofs?

Misunderstanding of definitions The absolute number one problem mentioned by most lecturers
taking part in a survey is the fact that students do not learn definitions. It is important to be on track
of the terminology - there will be a lot of it, so it will be hard to learn it all if you leave it for later!
(Students) do not know where to start because they do not know the definitions of the
objects they are working with
Not enough words Lecturers are waiting for your words! Prioritising symbols to words especially
at the beginning of mathematics studies is a common mistake, because it seems like a page with lots
of symbols looks clever.
I have found it common in particular for first year students not to explain the steps in
their arguments, as if they think they are not allowed to use words, only symbols
Lack of understanding
When a student gets to a point in a proof that they cannot proceed from, often the
conclusion of the result follows immediately, and it is clear that the student does not
understand the necessary missing arguments
They are trying to memorise the proofs rather than understand them
Incorrect steps Although your argument should start at the beginning and then lead to the final
statement, while constructing the proof you may want to look at the conclusion and imagine how it
may be arrived from the hypothesis. You may then be able to reverse the steps to produce a good proof
- not always though, be careful! Another idea is to work from both ends (both from the beginning and
from the end) and meet somewhere in the middle. This is all allowed during constructing the proof
but remember to then produce a neat final version with all steps well explained. Good knowledge of
different methods of proofs is also essential.
Contradiction arguments often have incorrect conclusions; induction arguments often
start at the incorrect point, and the induction hypothesis is often abused

32
c
University
of Birmingham 2014

9.2

What are the reasons for mistakes?

Again we come back to definitions, as incorrect arguments may come from the fact that students
do not appreciate the importance of mathematical definitions. Sometimes we lack the motivation
and sometimes also ability, hence practise is key to success. Mistakes may also come from the fact that
school maths does not prepare students adequately in how to present a mathematical argument. A
good method to develop mathematical reasoning is to create your own examples on which you can
practice writing proofs. How many teachers challenge their students in this way though? How much
is there of understanding maths and how much of getting the required grade?
Additionally, students are often asked questions which have nice looking answers and the final
steps of a calculation can be somewhat guessed to obtain an answer that looks correct. However, it is
very difficult to guess parts of a proof correctly. It seems that there is no magic fix for this, other
than practice and feedback. Finally though, this should not necessarily be considered a mistake
rather a step in learning. Writing proofs in pure mathematics university courses is different from what
has been done during school, so it will take some time to get used to doing this properly.

9.3

Advice to students for writing good proofs

Here are the tips collected together, which come from the lecturers at University of Birmingham.
Have a look at what they think and what they are looking for when marking your work!
Often youll need to do some rough work to figure out how your proof should go. Some
of the time, you will be able to get a good idea of how a proof should go by noticing that
it can be similar to the proof of something else.
Understand every line that you write, and do not make bogus claims.
Understanding the proofs in the lectures/ lecture notes as a way of understanding of
the type of reasoning that is involved in a proof.
How you lay out a proof is at least as important as the content (its not ok if its all in
there somewhere).
Try to make your proofs easier to follow by including brief phrases where appropriate.
For example rather than writing (statement A) implies (statement B) it may be more
enlightening to write (statement B) follows from (statement A) because. . .
Try to write out many of them, understanding every step. Ask others about unclear
points. Try to follow proofs in class or in books, and ask about unclear points.

9.4

Friendly reminder
The importance of proofs goes well beyond a university degree. It is eventually about using reason in everyday life. This could contribute to solving
major and global problems.

You have seen many methods of proofs presented in previous sections and they are all used in
different areas of mathematics. It has been underlined many times that writing proofs is not easy, but
with a lot of practice and open mind, pure mathematics is not as scary. Here are some final tips to
keep in your head when starting the next proof. Good luck!
Experiment! If one method does not work, try a different one. Lots of practice allows for an
educated guess in the future;
do not start with what you are trying to prove;
33
c
University
of Birmingham 2014

 use correct English with full punctuation;

begin by outlining what is assumed and what needs to be proved; do not skip this step!
remove initial working when writing up the final version of the proof, but include all steps of
reasoning.

34
c
University
of Birmingham 2014

10
10.1

Sets
Basics

To be able to write well in mathematical language, you will need at least the basics of the set theory.
Most lecturers will assume that you know it, or will include it in the zero chapter in their lecture
notes, so you must know all these definitions and symbols before you start working on your first proofs.
Definition 10.1. A set is a collection of objects, which are called the elements or members of a set.
Sets are usually denoted by capital letters and the members by lower case letters. We usually write
all elements in curly brackets. The notation
A = {a, b, c}
means that the set A consists of 3 elements: a, b and c. We can say that the element a belongs to the
set A, write a A, or that d is not a member of A, write d
/ A.
Example 10.1. Examples of the sets:
B = {1, 2, 3}, C = {0}, D = {x, y, }, E = {{red, white}, blue}, F = {1, 0, cos , 1}.
Note: the order of the elements in a set does NOT matter, i.e. {2, 3, 1} is exactly the same as set B
in the example given above. Observe also the difference between the sets E = {{red, white}, blue} and
{red, white, blue}. Here blue is an element of both sets, but red
/ E. We have that {red, white} E.
So a set can be an element of another set.
The symbol represents the empty set, which contains no elements. Note that {} is NOT an
empty set!

10.2

Subsets and power sets

Definition 10.2. If A is a subset of B (write A B), then all elements of A are also elements of B;
A B x A, x B. So A is contained in B.
If you want to say that A is NOT subset of B, write mathematically A * B.
Example 10.2.
{1, 10} {1, 10, 100},
{1000, 10} * {1, 10, 100},
{}.
Notice that the empty set is a subset of any set.
To show that A B, you need to show that every element of A is also an element of B.
We will present a very simple theorem and its proof. It should give you an idea of how to structure
your argument. Note the amount of words used!
Theorem 10.3. Let A, B and C be sets. Then
(a) A A
(b) If A B and B C, then A C.

35
c
University
of Birmingham 2014

Proof. (a) Start by choosing an arbitrary element a A (we mean A on the left hand side of A A).
Then it follows that a A (A on the left right side).
We chose a arbitrarily, hence the argument holds for all a A. So by the definition of a subset,
A A.
(b) Let a A. Then, since A B, it follows that a B. Since B C, we have that a C.
So a A implies a C. Therefore A C.

If A is a subset of B but it they are not equal, then we say that A is a proper subset of B and
write it A B (or A ( B or A & B).
To show that A is a proper subset of B, you need to show that A B and find at least one element
of B which is not an element of A.
Example 10.4.
{a, b, c} {a, b, c, d},
{1, 2, 3} 6 {1, 2, 3} but {1, 2, 3} {1, 2, 3}.

Definition 10.3. The power set of a set A consists of all subsets of A and is usually denoted by
P (A) (some writers use 2A ).
The power set of A = {x, y, } is P(A) = {{x}, {y}, {}, {x, y}, {x, }, {y, }, {x, y, }, }.

10.3

Cardinality and equality

Definition 10.4. In mathematics, the cardinality of a set A (card(A) or |A|) is a measure of the
number of the elements of the set5 .
Example 10.5. If A = {a, b, c}, then |A| = 3.
Example 10.6. Important to notice:
|{}| = 1, while || = 0.
|{0}| =
6 0, but |{0}| = 1.
Notice that the repetitions are ignored when we are counting the members of the set. The convention
is to list each element only once, however as you see in the Example 4.1. the same number can be
written in different forms.
F = {1, 0, 1, cos } = {1, 0, 1}, as cos = 1. Hence, |F | = 3.
Definition 10.5. Two sets A and B are equal when they have exactly the same elements, i.e. every
element of A is an element of B and every element of B is an element of A. So
A = B A B and B A.
Example 10.7.
{1, 0, 1} = {1, 0, 1, cos },
{2, 3, 3, 3, 3, 2, 3, 2, 2} = {2, 3},
{2, 3, 4} = {4, 2, 3}.
5 This

works for finite sets - see section 4.6. for the cardinality of the infinite sets

36
c
University
of Birmingham 2014

To show that two sets A and B are equal, pick an arbitrary x A and show that x B and vice
versa.
Exercise 10.8. Let A = {1, 2, 3, 4} and B = {x : x N, x2 < 17}, where N is the set of natural
numbers. Show that A = B.
Answer. To prove the equality of the sets, we must show that for every x,
x B x A (B A)
and
x A x B (A B).

So if x B, then x < 17, which implies x < 17. Therefore x 4. Since x is a positive
integer, therefore for every x B we have that 0 < x 4. Hence, x B x A.
2

Now assume x A, so x {1, 2, 3, 4}. To prove that x B, it suffices to show that the largest
element x A satisfies x2 < 17. Then it is also true for the smaller values since they all belong
to N.
Since x A, x2 42 16 < 17, we have that x A x B.

Exercise 10.9. Show that {(cos t, sin t) : t R} = {(x, y) : x2 + y 2 = 1}.

Answer. Let A = {(cos t, sin t) : t R} and B = {(x, y) : x2 + y 2 = 1}. Now, to show that
A = B we need to show that A B and B A.
Let x = cos t and y = sin t. Then
x2 + y 2 = cos2 t + sin2 t = 1
because cos2 t + sin2 t = 1 is a known identity. Hence we have that A B.
Now, to show that B A we appeal to geometry. Let (x, y) B, hence x2 + y 2 = 1. So (x, y)
lies on the unit circle.
(x, y)
t

Therefore, we have that cos t = x and sin t=y. As x2 + y 2 = 1, substituting in for x and y gives
cos2 t + sin2 t = 1
and hence we have shown that B A and so A = B.

10.4

Common sets of numbers

The commonly used sets of numbers are:

The set of natural numbers, N = {1, 2, 3, 4...}. Careful! Some mathematicians include 0 in N,
37
c
University
of Birmingham 2014

the others do not. Your lecturers will usually tell you at the beginning which version do they
mean by writing N;
The set of integers, Z = {3, 2, 1, 0, 1, 2, 3, 4, ...};
The set of rational numbers Q = { m
n : m, n Z, n > 0};
The set of real numbers R, which is the union6 of both
rational Q and irrational numbers (which
cannot be expressed as a fraction, for example log 2, 2, , e).
Notice that one set is a subset of another, in the following order: N Z Q R.

10.5

How to describe a set?

As mentioned before, we write the elements of a set inside curly brackets. The elements should be
defined clearly, for example
{x R : 2 < x < 3}.
Notice that it is often tempting to write just {2 < x < 3}, but it is meaningless! Naturally, x being
a natural number rather than a real number makes a big difference! Hence, make sure that you define
your sets properly. It may give rise to a very long and scary looking sentence with lots of notation,
but there will be no room for ambiguity.
We can represent the intervals on the real number line using set notation. Let a, b R be any two
numbers with a < b. Then we can define the following:
[a, b] = {x R : a x b}
(a, b) = {x R : a < x < b}
[a, b) = {x R : a x < b}
(a, b] = {x R : a < x b}
[a, ) = {x R : a x}
(a, ) = {x R : a < x}
(, b] = {x R : x b}
(, b) = {x R : x < b}
R = (, )

closed interval
open interval
half open intervals
infinite intervals

It is important to understand that is not a real number! It is a symbol to represent the fact that
the interval goes on forever, hence infinity is a concept rather than a big numerical value. Notice that
no interval is closed at or , as there is no number there to be included in the real line.

10.6

More on cardinality

We have seen earlier the definition, stating that the cardinality of a set is a measure of the number
of the elements of the set. This is indeed the case when the sets, say A and B, are finite, i.e. they
consist of a finite number of elements. Then the notion |A| = |B| means the equality of integers. Do
not confuse it with A = B, which means that not just a number of the elements is the same, but
elements themselves are equal.
The problem arises when we consider the infinite sets. The number of their elements is not finite
and so we cannot count them and arrive at finite number. In this case we say that two sets A and
B have the same cardinality whenever we can match, or pair off, the elements of the set A with the
elements of the set B. Put more exactly, two sets A and B have the same cardinality whenever there
is a bijection7 between A and B.
6 See

section 10.7
correspondence or bijection - see section 11

7 One-to-one

38
c
University
of Birmingham 2014

For example, given set A = {1, 2, 3}, its cardinality equals 3 (card(A) = 3). It is more complicated
to talk about the cardinalities of the sets of, say, real or natural numbers. It automatically brings up
the interesting discussion about infinity and its different sizes (!). One of the most striking theorems
of set theory was proved in 1878 by George Cantor, who stated that card(R) = card(R2 ). The proof
to this powerful statement is quite simple (see appendix), but who would have thought that a straight
line has as many points as the plane?

10.7

Operations on sets

Below we present the basic operations on sets, which can be nicely presented on the Venn diagrams.
Unions

A B = {x : x A or x B}

Intersections

A B = {x : x A and x B}

note: when A B = , then A and B are said to be disjoint.

Complements

A\B = {x A : x
/ B}

The above notation is sometimes called a relative complement of B in A. We also distinguish

the universal complement or simply a complement, U\A (denoted by A0 or Ac ).
U

Ac = {x : x
/ A}

Here U is the universal set, which contains all objects, including itself. Notice that U must be
specified (which is very often omitted) for Ac to be well-defined8 .
8 well-defined

means that the expression is unambiguous, so it has a unique value or interpretation assigned to it

39
c
University
of Birmingham 2014

 Cartesian Product
A B = {(a, b) : a A and b B}
Note, that here (a, b) is not an open interval , but an ordered pair, which is a pair of elements
written in a particular order. This means that (x, y) and (y, x) represent different ordered pairs,
unless x = y. Sometimes x and y are called coordinates, with x being the first one and y the
second.

10.8

Theorems

In this section we will present statements and proofs of simple theorems, which are a good warm-up
to the more complicated material on the set theory.
Theorem 10.10. For any sets A, B and C, the following statements are true:
i. A A = A,
ii. A B = B A, (commutative law)
iii. A (B C) = (A B) C, (associative law)
iv. A A B and B A B,
v. A = A.

Proof. i. Take any x in A. But then the statement x A or x A is also trivially true ( x, x
A (x A or x A) ). Therefore the statement holds.
ii. (x A or x B) (x B or x A). Trivially, the statements are equivalent, hence (ii) is true.
iii. We will start from the left hand side of the statement to arrive at the right hand side, using
definitions of the union of sets.
x, (x A (B C)) (x A or x (B C)) (x A or (x B or x C)) ((x
A or x B) or x C) (x (A B) or C) (x (A B) C).
iv. This follows from the definition of the union of the sets: x A B means x A or x
B. Hence, A A B and B A B.
v. x A x A or x . But does not contain any elements, hence x A x A.
The next theorem is similar to the previous one, but it deals with the intersections of the sets. The
proofs are not provided and the reader is strongly encouraged to mimic the above arguments to prove
the following statements.
Theorem 10.11. For any sets A, B and C, the following statements hold:
i. A A = A,
ii. A B = B A, (commutative law)
iii. A (B C) = (A B) C, (associative law)
iv. A B A, A B B,
v. A = .

40
c
University
of Birmingham 2014

The next result provides the rules when taking the unions and intersections with three different sets.
Theorem 10.12 (Distributive laws for sets). For any three sets A, B and C,
i. A (B C) = (A B) (A C),
ii. A (B C) = (A B) (A C).
When working with sets, it is a good idea to picture the statement first, using Venn diagrams, to see
that the theorem is in fact true. Remember that the diagrams are not proofs and the formal argument
needs to follow.
Draw separate pictures, one for the left hand side and the other for the right hand side of the
statement. You should get exactly the same graphs, which can reassure you that the theorem is
correct. Then you can start your formal proof.
A

Figure 3: A (B C)

Figure 4: (A B) (A C)

Proof. i.

x (A (B C)) x A and x (B C)
x A and (x B or x C)
(x A and x B) or (x A and x C)
x (A B) or x (A C)
x (A B) (A C).
ii. This part is left to the reader as an exercise.
Three of the earlier described operations on sets (intersection, union and complement) have been
related to each other in the 19th century by the British mathematician Augustus De Morgan. We
present his theorem below.
Theorem 10.13 (De Morgans Laws). Let A and B be sets. Then the following statements are true:
i. (A B)0 = A0 B 0 ,
ii. (A B)0 = A0 B 0
The proof is not provided, but is an excellent exercise so the reader is strongly encouraged to try
writing it.

41
c
University
of Birmingham 2014

11

Functions

A function can be defined by specifying three elements: set of inputs (called domain), set of outputs
plus possibly extra elements9 (called codomain) and a rule assigning each element from the domain
to a unique member of the codomain. Formally, functions can be defined in many ways, for example
in terms of sets, like below.
Definition 11.1 (Function). Let A and B be sets. A function (also called a map) f from A to B,
f : A B is a subset F A B such that for each a A there is one and only one pair of the form
(a, b) in F .
For example, we may have the set of real numbers R as a domain and the codomain and the rule
f (x) = x2 . In this case, we define the function in the following way:
let f : R R be given by f (x) = x2 .
Writing just f (x) = x2 on its own does not describe the function fully; it is tempting to state
f (x) = x2 is a function, but you need to specify the domain and codomain as well. Notice that they
are a really important part of our definition of a function. Imagine that we changed the codomain
from R to Z; the resulting function is completely different! The functions are equal only if all three
elements are the same.
Many authors of the mathematical books, like for example D. Bloch [1], will point out the difference
between f and f (x). This is one of those things that are often assumed as unimportant by students
and needs to be pointed out at the beginning, since it may lead to misunderstandings in the future.
Informally, you may still find people referring to the function x2 , but this is really an abuse of
notation. Notice how we defined the function above; f : R R suggests that the name of the
function is f and NOT f (x).

11.1

Image and preimage

There is an important difference between the codomain and the image, sometimes also called range.
It is common to mix these two terms and interchange them when describing the functions. You can
think of a codomain as the target set of a function, which all the outputs are constrained to fall into.
It may however include the elements which are not the outputs of the function for any of the elements
of the domain. The image of the function, however, is a set of values that the function can produce.
The formal definition is given below.
Definition 11.2 (Image). Let f : A B be a function from set A to B and let X be a subset of A.
The image of a subset X A under f is the subset f (X) B defined by
f (X) = {b B|b = f (a) for some a X}.
The image f (A) of the entire domain is called the image of f.
We also use the term image when talking about a single element of a set. Letting f : A B to
be a function from the set A to B, we can find the image of any element a A under f . The function
f (a) = b gives us the output, i.e. the element of the codomain B, which is our required image of a
point a.
So far, we have only looked at the function as the process of inputting the values and receiving the
output values. We can also reverse this operation.
Definition 11.3 (Inverse image, pre-image). Let f be a function from A to B. The pre-image or
inverse image of a set Y B under f is the subset of A defined by
f 1 (Y ) = { a A | f (a) Y }.
9 see

the difference between the codomain and the range

42
c
University
of Birmingham 2014

In various mathematical books you may find a slightly different notation for the image and pre-image
of the function; for the image of the function f : A B, some authors use f (A) instead of our f (A),
and the pre-image is denoted by f (B) instead of f 1 (B). Although the notation used by the author
of this paper is more common and should be sufficient for students, Bloch [1] in his book calls such
writing style an abuse of notation, which is a technically incorrect way of writing something that
everyone understands what it means anyway, and tends not to cause problems. He argues that the
notation f 1 (B) does not imply that the function f 1 exists, which sends us to the topic of inverse
functions, discussed later in this document.

Examples
Let us give some examples on what we have discussed so far.
The first, classical example when talking about the images and preimages is the function f : R
R given by f (x) = x2 . Notice that the image of 3 is 9, but the preimage of 9 is {-3,3}.
Given the function g : {1, 2, 3, 4} {red,white,blue} defined by
(
red
if x = 1, 2,
g(x) =
white otherwise.
Then the image of the function is the set {red, white}. The preimage of red is g 1 ({red}) = {1, 2}
and g 1 ({red, blue}) = {1, 2} as well. Notice that g 1 ({blue}) = .
Let h : R R, h(x) = ex . Here the codomain is defined to be R, but the image of the function
is the set (0, ).

11.2

Composition of the functions

We can combine the functions in many different ways using the composition of the functions.
Definition 11.4. Let f : A B and g : B C be functions. The composition of f and g is the
function g f : A C defined by
(g f )(x) = g(f (x)) for all x A.
This means that for any element of the domain (a A), we can apply the function f and then
the function g, to get an output ((g f )(a) C) of the function g f . To be able to combine two
functions, we must have that the codomain of the first one is equal to the domain of the second one.
Be careful not to mix the order of applying the functions, as the commutativity does not hold in
this case. The following example illustrates this fact.
Example 11.1. Let f : R R and g : R R be functions given by f (x) = x3 and g(x) = x + 5.
Then (g f )(2) = g(f (2)) = g(23 ) = g(8) = 13,
but (f g)(2) = f (g(2)) = f (2 + 5) = f (7) = 73 = 243.
Notice that in the above example we are able to form both g f and f g, since the domains and
codomains are equal. In the situation when
f : A B, g : B C,
it is impossible to construct f g, unless A = C!

11.3

Special functions

Among many different functions we can distinguish a special function, which will always return the
same value that was used as an input. We call such function the identity function defined by
f : A A such that f (x) = x and usually denoted by idA .
43
c
University
of Birmingham 2014

A constant function is another specific function. It is of the form f (x) = c, where c is the constant
returned from the function, regardless of the input value.

11.4

Injectivity, surjectivity, bijectivity

Definition 11.5 (Injectivity, one-to-one function). This is a function that maps distinct points of the
domain to distinct points of the codomain.
Let f be a function whose domain is D. If f is injective, then a, b D, a 6= b f (a) 6= f (b).
You can often judge whether the function is injective or not by just looking by its graph. We should
see that every element from the domain is assigned to exactly one member of the codomain. Remember
that although the picture is a helpful tool, it does not replace the formal argument!
3

g(x) = x2

f (x) = x3
2

1
2
2

Figure 6: Not injective

Figure 5: Injective
To show the injectivity (or non-injectivity) of a function f , we use the definition. We let x1 , x2 R
and start by assuming that f (x1 ) = f (x2 ). If the function is injective, we should arrive at the statement
x1 = x2 . Otherwise, the it is not injective.
Let f : R R, f (x) = x3 ,

Let g(x) : R R, g(x) = x2

Assume that f (x1 ) = f (x2 )

=
=

x31
x31

= x32
x32

Assume that g(x1 ) = g(x2 )

= x21 = x22
= x21 x22 = 0

=0

= (x1 x2 ) (x1 + x1 x2 + x22 ) = 0

|
{z
}
always > 0

= x1 = x2

= x1 = x2
Hence g is not injective.

Hence this function is injective.

Note that it is possible to get g(x) = x2

injective, by restricting the domain to R+
(positive real numbers).

44
c
University
of Birmingham 2014

= (x1 x2 )(x1 + x2 ) = 0

Definition 11.6 (Surjectivity, onto function). Let f be a function f : A B. Then f is surjective

for any element b B, there exists an element a A such that f (a) = b.
Note that f may send more than one element from A to the same element in B.
The image of the surjective function is equal to its codomain, so every element of B has at least
one element of A assigned to it. From Figure 5 we can see that the function f (x) = x3 is surjective,
however g(x) = x2 is not. To make g surjective, we need to change the codomain to R+ .

Definition 11.7 (Bijectivity, one-to-one correspondence10 ). A bijective function f : A B is mapping between two sets A and B, which is both injective and surjective. Each element from the set A
is paired with exactly one element from set B and each element from set B is paired with exactly one
element from set A.

Exercise 11.2. Study the examples below. Decide whether they are surjective, injective or bijective. If not, how can you restrict domain or codomain to make them bijective?
1. f : R R, f (x) = sin x,
2. g : Z Z, g(x) = ex ,
3. h : R+ R, h(x) = x2 5.

11.5

Inverse function

For any bijective function f , we can find an inverse function (f 1 ), which reverses the process.
While the function f takes an input x and returns the output f (x) = y, the inverse function will take
y back to x, i.e.
f 1 (y) = x.
Notice also that the composition of the function with its inverse gives the identity function11 ,
f 1 (f (x)) = x = f (f 1 (x)).
Notice that the function f (x) = x2 , will have an inverse (or not) depending on the domain (will
need to be restricted so that f is bijective).

11.6

Even and odd functions

Definition 11.8. The function f : X Y is even if it is true that

if x X, then x X, and
f (x) = f (x) for all x X.

10 Note that the one-to-one function and one-to-one correspondence are two different concepts and should not be
confused.
11 See the theorem at the end of the chapter

45
c
University
of Birmingham 2014

Even functions are symmetric with respect to the yaxis. You can check that the function
g(x) = x2 (fig.6) is an even function, but f (x) = x3 (fig.5) is not.
Definition 11.9. The function f : X Y is odd if it is true that
if x X, then x X, and
f (x) = f (x) for all x X.
The odd functions are symmetric with respect to the origin, which means that they remain
unchanged when reflected across both x and yaxis. From the graph you can see that the function
in fig.5 is odd, but the one in fig.6 is not.

11.7

Exercises

Theorem 11.3. Let f : A B, g : B C and h : C D be functions.

1. If f has an inverse, then the inverse is unique;
2. If f and g are injective, then g f is also injective;
3. If f and g are surjective, then g f is also surjective;
4. If f and g are bijective, then g f is also bijective;
5. (h g) f = h (g f ) (Associative Law);
6. f idA = f and idB f = f (Identity Law).
The above statements are a good exercise to practise proofs on functions and are left for the reader
to try.

46
c
University
of Birmingham 2014

References
[1] Ethan D. Bloch, Proofs and Fundamentals. A First Course in Abstract Mathematics,
Birkh
auser, Boston, 2000.
[2] Kevin Houston, How to Think Like a Mathematician. A Companion to Undergraduate Mathematics, Cambridge University Press, Cambridge, 2009.
[3] Martin Liebeck, A Concise Introduction to Pure Mathematics, CRC Press, Taylor & Francis
Group, Boca Raton, 2011.
[4] Richard Bornat, Proof and Disproof in Formal Logic, Oxford University Press, New York, 2005.
[5] Kam-Tim Leung, Doris Lai-Chue Chen, Elementary Set Theory, Hong Kong University Press,
Hong Kong, 1967.
[6] Martin Aigner, G
unter M. Ziegler, Proofs from THE BOOK, Springer-Verlag, Berlin, 2001.

47
c
University
of Birmingham 2014

12

Appendix

Theorem 12.1. card(R)= card(R2 )

Notes. Here the question is: how is it actually possible to correspond an element from the real line
to an element on the plane? We will present the idea of the proof, considering an interval I = [0, 1]
and the unit square Q = [0, 1]2 instead of the whole real line and the plane.
Proof. Consider a point in the unit sqare (x, y) Q and write the decimal expansion of x and y,
x = 0. x1 x2 x3 x4 x5 ...
and
y = 0. y1 y2 y3 y4 y5 ...
Now, to find a point on I corresponding to (x, y), we use the above expansion of x and y, creating
a point z, as shown below.
z = 0. x1 y1 x2 y2 x3 y3 ...
Note, that this approach would work well if not for the fact that 0.a b c 0 9 9 9... = 0.a b c 1 0 0 0....
So we can represent the point (x, y) Q using two different decimal expansions of x and y, which
would result in getting two distinct points z and z 0 (z, z 0 I). Hence the mapping is not one-to-one.
To avoid this situation, we do not allow the decimal expansions with the infinite number of nines, (for
example 0.04999... is forbidden).
Now we can be sure that the point z I is unique for each pair (x, y).

Theorem 12.2. Suppose that {v1 , v2 , v3 } is the linearly independent set of vectors. Then {v1 , v3 } is
also linearly independent set.
Notes. The proof is not complete and the reader is encouraged to fill in the gaps. First, recall that finitely many vectors v1 , v2 , . . . vr are linearly dependent if there exist scalars
(not all zero) such that a1 v1 + a2 v2 + + ar vr = 0. If the vectors are not linearly
dependent, then we say they are linearly independent.
Now, the statement is fairly simple but how to prove it? If we remember the relation
(A B) (not B not A), then the second one is easy to show.
Click here to see a video example
Proof. Assume that {v1 , v3 } are . . .
(assume the opposite)

So we can write
. . . v1 + . . . v3 = . . .
for some scalars . . . , . . . R.
(fill in the dots using your knowledge about the dependent vectors)

Now notice that

. . . v1 + . . . v2 + . . . v3 = 0,
hence we have that v1 , v2 and v3 are . . . .
So we proved that. . .
(write the conclusion)

hence by contrapositive we have that . . . .

48
c
University
of Birmingham 2014

Theorem 12.3. ln x x 1 for x > 0.

Proof. We will use the expansion of the exponential function (proof of the expansion is not provided,
as it is commonly used in the first year of studies):
ex = 1 + x +

x2
x3
x4
+
+
+ ...
2!
3!
4!

We first show that ex 1 + x for all x R.

From the expansion above (*), this is clear for x 0.
Also from (*), if 0 x < 1,
x3
x2
+
+ ...
2
3!
1 + x + x2 + x3 + . . .
1
=
1x

ex = 1 + x +

Therefore if 1 < x < 0,

ex

1
1
=
1 (x)
1+x

and so ex 1 + x for 1 < x < 0.

Finally, for x 1, ex is positive and 1 + x is negative. So ex 1 + x for x 1.
Therefore, ex 1 + x for all x R.
Now, let x = ln u (where u > 0). Then
eln u 1 + ln u
u 1 + ln u
So ln u 1 u for all n > 0.

49
c
University
of Birmingham 2014

(*)

c
University
of Birmingham 2014