How to Prove It – A Structured Approach

Introduction
College mathematics deals with a wider variety of questions, involving not only numbers, but also sets, functions, and other mathematical
objects. What ties them together is the use of deductive reasoning to find answers to questions. When you solve an equation for x you are using
the information given by the equation to deduce what the value of x must be. Similarly, when mathematicians solve other kinds of mathematical
problems, they always justify their conclusions with deductive reasoning.

Deductive reasoning in mathematics is usually presented in the form of a proof.

An integer large than 1 is said to be prime if it cannot be written as a product of two smaller positive integers. If it can be written as a product of
two smaller positive integers, then it is composite.

Consider the table in Figure I.1. For each integer n from 2 to 10, the table shows whether or not both n and 2𝑛 − 1 are prime, and a surprising
pattern emerges.

n Is n prime? 2𝑛 − 1 Is 2𝑛 − 1 prime?
2 yes 3 yes
3 yes 7 yes
4 no: 4 = 2 * 2 15 no: 15 = 3 * 5
5 yes 31 yes
6 no: 6 = 2 * 3 63 no: 63 = 7 * 9
7 yes 127 Yes
8 no: 8 = 2 * 4 255 no: 255 = 15 * 17
9 no: 9 = 3 * 3 511 no: 511 = 7 * 73
10 no: 10 = 2 * 5 1023 no: 1023 = 31 * 33
Figure I.1.

Will this pattern continue? It is tempting to guess that it will, but this is only a guess. Mathematicians call such guesses conjectures.

Conjecture 1. Suppose n is an integer larger than 1 and n is prime. Then 2𝑛 − 1 is prime.

Conjecture 2. Suppose n is an integer larger than 1 and n is not prime. Then 2𝑛 − 1 is not prime.

Unfortunately, if we continue the table in Figure I.1, we immediately find that Conjecture 1 is incorrect. It is easy to check that 11 is prime, but
211 − 1 = 2047 = 23 * 89, so 211 − 1 is composite. Thus, 11 is a counterexample to Conjecture 1. The existence of even one counterexample
establishes that the conjecture is incorrect.

No matter how many examples we check, there is always the possibility that the next one will be the first counterexample. The only way we can
be sure that Conjecture 3 is correct is to prove it. In fact, Conjecture 2 is correct. Here is a proof of the conjecture:

Proof of Conjecture 2. Since n is not prime, there are positive integers a and b such that a < n, b < n and n = ab. Let 𝑥 = 2𝑏 − 1 and 𝑦 = 1 +
2𝑏 + 22𝑏 + ⋯ + 2(𝑎−1)𝑏 . Then:

𝑥𝑦 = (2𝑏 − 1)(1 + 2𝑏 + 22𝑏 + ⋯ + 2(𝑎−1)𝑏 )

𝑥𝑦 = (2𝑏 )(1 + 2𝑏 + 22𝑏 + ⋯ + 2(𝑎−1)𝑏 ) − (1 + 2𝑏 + 22𝑏 + ⋯ + 2(𝑎−1)𝑏 )
𝑥𝑦 = (2𝑏 + 22𝑏 + 23𝑏 + ⋯ + 2𝑎𝑏 )(1 + 2𝑏 + 22𝑏 + ⋯ + 2(𝑎−1)𝑏 )
𝑥𝑦 = 2𝑎𝑏 − 1
𝑥𝑦 = 2𝑛 − 1

Since 𝑏 < 𝑛, we can conclude that 𝑥 = 2𝑏 − 1 < 2𝑛 − 1. Also, since 𝑎𝑏 = 𝑛 > 𝑎, it follows that 𝑏 > 1. Therefore 𝑥 = 2𝑏 − 1 > 21 − 1 = 1, so
𝑦 < 𝑥𝑦 = 2𝑛 − 1. Thus, we have shown 2𝑛 − 1 be written as the product of two positive integers x and y, both of which are smaller than 2𝑛 −
1, so is not 2𝑛 − 1 prime.

Now that the conjecture has been proven, we call it a theorem.

 Theorem 3. There are infinitely many prime numbers.

Proof of Theorem 3. Suppose there are only finitely many prime numbers. Let p1, p2, …, pn be a list of all the prime numbers. Let 𝑚 =p1p2…
pn+1. Note that m is not divisible by p1, since dividing m by p1 gives a quotient of p2p3… pn and a remainder of 1. Similarly, m is not divisible by
any of the p2, p3 …, pn.

We now use the fact that every integer larger than 1 is either prime or can be written as a product of two or more primes. Clearly m is larger
that 1, so m is either prime or a product of primes. Suppose first that m is prime. Note that m is larger than all of the numbers in the list p1, p2
…, pn, so we’ve found a prime number not in this list. But this contradicts our assumption that this was a list of all prime numbers.

Now suppose m is a product of primes. Let q be one of the primes on this product. Then m is divisible by q. But we’ve already seen than m is not
divisible by any of the numbers in the list p1, p2 …, pn, so once again we have a contradiction with the assumption that this list included all prime
numbers.

Since the assumption that there are finitely many prime numbers has led to a contradiction, there must be infinitely many prime numbers.

Prime numbers of the form 2𝑛 − 1 are called Mersenne primes, after Father Marin Mersenne (1588-1648), a french monk and scholar who
studied these numbers. Mersenne primes are related to perfect numbers, the subject of another famous unsolved problem of mathematics. A
positive integer n is said to be a perfect number if n Is equal to the sum of all positive integers smaller than n that divide n. (For any two integers
m and n, we say that m divides n if n is divisible by m; in other words, if there is an integer q such that 𝑛 = 𝑞𝑚.) For example, the only positive
integers smaller than 6 that divide 6 are 1, 2 and 3, and 1 + 2 + 3 = 6. Thus, 6 is a perfect number. The next smallest perfect number is 28.

Euclid prove that if 2𝑛 − 1 is prime, then (2𝑛−1 )(2𝑛 − 1) is perfect. Thus, every Mersenne prime gives rise to a perfect number. Furthermore,
about 2000 years after Euclid’s proof, the Swiss mathematician Leonhard Euler (1707-1783), proved that every even perfect number arises in
this way. Because it is not known if there are infinitely many Mersenne primes, it is also not known if there are infinitely many even perfect
numbers. It is also not known if there are any odd perfect numbers.

Theorem 4. For every positive integer n, there is a sequence of n consecutive positive integers containing no primes.

Proof of Theorem 4. Suppose n is a positive integer. Let 𝑥 = (𝑛 + 1)! + 2. We will show that non of the numbers 𝑥, 𝑥 + 1, 𝑥 + 2, …, 𝑥 + (𝑛 + 1)
is prime. Since this is a sequence of n consecutive positive integers, this will prove the theorem. To see that 𝑥 is not prime, note that:

𝑥 = 1 ∗ 2 ∗ 3 ∗ 4 ∗ … ∗ (𝑛 + 1) + 2
𝑥 = (2){[1 ∗ 3 ∗ 4 ∗ … ∗ (𝑛 + 1)] + 1} a

Thus, 𝑥 can be written as a product of two smaller positive integers, so 𝑥 is not prime. Similarly, we have:

𝑥 + 1 = 1 ∗ 2 ∗ 3 ∗ 4 ∗ … ∗ (𝑛 + 1) + 3
𝑥 + 1 = (3){[1 ∗ 2 ∗ 4 ∗ … ∗ (𝑛 + 1)] + 1}

So 𝑥 + 1 is also not prime. In general, consider any number 𝑥 + 𝑖, where 0 ≤ 𝑖 ≤ 𝑛 − 1. Then we have:

𝑥 + 𝑖 = 1 ∗ 2 ∗ 3 ∗ 4 ∗ … ∗ (𝑛 + 1) + (𝑖 + 2) a
𝑥 + 𝑖 = (𝑖 + 2){[1 ∗ 2 ∗ 3 ∗ 4 ∗ … ∗ (𝑛 + 1)] + 1}

So 𝑥 + 𝑖 is not prime.

Theorem 4 shows that there are sometimes long stretches between one prime and the next prime. But primers also sometimes occur close
together. Since 2 is the only even prime number, the only pair of consecutive integers that are both prime is 2 and 3. But there are lots of pairs
of primes that differ by only two, for example, 5 and 7, 29 and 31, and 7949 and 7951. Such pairs of primes are called twin primes. It is not
known wether there are infinitely many twin primes.
Chapter 1 - Sentential Logic
1.1 Deductive Reasoning and Logical Connectives
We begin our study of mathematical reasoning and proofs by examining how deductive reasoning works:

Example 1.1.1. Here are three examples of deductive reasoning:

1. It will rain or snow tomorrow.
It’s too warm for snow.
Therefore, it will rain.

2. If today is Sunday, then I don’t have to go to work today.

Today is Sunday.
Therefore, I don’t have to go to work.

3. I will go to work either tomorrow or today.

I’m going to stay home today.
Therefore, I will go to work tomorrow.

In each case, we have arrived at a conclusion from the assumption that some other statements, called premises, are true. When there are two
premises, and they are both true, we can be sure that the conclusion is also true. This is the sense in which the conclusion is forced on us by
premises, and this is the standard we will use to judge the correctness of deductive reasoning. We will say that an argument is valid if the
premises cannot all be true without the conclusion being true as well.

Here’s an example of an invalid deductive argument:

Either the butler is guilty or the maid is guilty.

Either the maid is guilty or the cook is guilty.
Therefore, either the butler is guilty or the cook is guilty.

The argument is invalid because the conclusion could be false even if both premises are true. For example, if the maid were guilty, but the butler
and the cook were both innocent, then both premises would be true and the conclusion would be false.

In Example 1.1.1. arguments 1 and 3 have the form:

𝑃 𝑜𝑟 𝑄.
𝑁𝑜𝑡 𝑄.
𝑇ℎ𝑒𝑟𝑒𝑓𝑜𝑟𝑒, 𝑃.

It is this form, and not the subject matter, that makes these arguments valid. Replacing certain statements in each argument with letter has two
advantages. First, it keeps us from being distracted by aspects of the arguments that don’t affect their validity.
Perhaps more important, our analysis of the forms of arguments 1 and 3 makes clear what is important in determining their validity: the words
or and not.
In this chapter. we’ll concentrate on words used to combine statements to form more complex statements. We’ll use letters to stand for
statements, but only for ambiguous statements that are either true or false. Questions, exclamations and vague statements will not be allowed.
It will also be useful to use symbols, sometimes called connective symbols. Here are our first connective symbols and the words they stand for:

∨ or
∧ and
¬ not

The statements 𝑃 ∨ Q is sometimes called the disjunction of 𝑃 and 𝑄, 𝑃 ∧ Q is called the conjunction of 𝑃 and 𝑄, and ¬ P is called the
negation of 𝑃.

Example 1.1.2. Analyze the logical forms of the following statements:

1. Either John went to the store, or we’re out of eggs.
𝑃: “John went to the store.”
𝑄: “We’re out of eggs.”
𝑃∨Q
2. Joe is going to leave home and not come back.
𝑃: “Joe is going to leave home.”
𝑄: “Joe is going to come back.”
𝑃∧¬Q

3. Either Bill is at work and Jane isn’t, or Jane is at work and Bill isn’t.
𝑃: “Bill is at work.”
𝑄: “Jane is at work.”
(𝑃 ∧ ¬ Q) ∨ (𝑄 ∧ ¬ 𝑃)
Example 1.1.3. What English sentences are represented by the following expressions?

𝑆: “John is smart.”
𝐿: “John is lucky.”

1. ( ¬ S ∧ L ) ∨ S: Either John is not smart and he is lucky, or he is smart.

2. ¬ S ∧ (L ∨ S): John is not smart, and either he is lucky or he is smart.
3. ¬ ( S ∧ L ) ∨ S: Either John is not both smart and lucky, or he is smart.

It is important to keep in mind that the symbols ∧, ∨ and ¬ don’t really correspond to all uses of the words and, or, and not in English. For
example, the symbol ∧ could not be used to represent the use of the word and in the sentence “John and Bill are friends.” because in this
sentence the word and is not being used to combine two statements. The symbols ∧ and ∨ can only be used between two statements, to
form their conjunction or disjunction, and the symbol ¬ can only be used before a statement, to negate it. This means that certain strings
of letters and symbols are simply meaningless. For example, 𝑃 ¬ ∧ Q, 𝑃 ∧ ∨ Q, and 𝑃 ¬ Q are all “ungrammatical” expressions in the
language of logic. “Grammatical” examples, such as those in Example 1.1.2 and 1.1.3, are sometimes called well-formed formulas of just
formulas.
Sometimes, words other than and, or and not are used to express the meanings represented by ∧, ∨ and ¬. For example, consider the first
statement in Example 1.1.3. Although we gave the English translation “Either John is nor smart and he is lucky, or he is smart.” an
alternative way of conveying the same information would be to say “Either John is not smart but he is lucky, or he is smart.”. Often, the
word but is used in English to mean and, specially when there is some contrast or conflict between the statements being combined. For a
more striking example, imagine a weather forecaster ending his forecast with the statement “Rain and snow are the only two possibilities
for tomorrow’s weather.”. This is just a roundabout way of saying that it will either rain or snow tomorrow. Thus, even though the
forecaster has used the word and, the meaning expressed by his statement is a disjunction. The lesson of these examples is that to
determine the logical form of a statement you must think about what the statement means, rather than just translating word by word into
symbols.
Sometimes logical words are hidden within mathematical notation. For example, consider the statement 3 ≤ 𝜋. Although it appears to be
a simple statement that contains no words of logic, if you read it out loud you will hear the word or.

𝑃: “3 < 𝜋”
𝑄: “3 = 𝜋”
𝑃 ∨ 𝑄: 3 ≤ 𝜋

For a slightly more complicated example, consider the statement 3 ≤ 𝜋 < 4. This statement means 3 ≤ 𝜋 and 𝜋 < 4, so once again a word of
logic has been hidden in mathematical notation.

𝑃: “3 < 𝜋”
𝑄: “3 = 𝜋”
𝑅: 𝜋 < 4
(𝑃 ∨ 𝑄)∧R

1.2 Truth Tables

An argument is valid if the premises cannot all be true without the conclusion being true as well. Thus, to understand how words such as and, or
and not affect the validity of arguments, we must see how they contribute to the truth or falsity of statements containing them.

When we evaluate the truth or falsity of a statement, we assign to it one of the labels true or false, and this label is called its truth table. Here
we present the truth table for 𝑃 ∧ Q and for ¬ P:

P Q 𝑃∧Q
F F F
F T F
T F F
T T T

𝑃 ¬S
F T
T F

The truth table for 𝑃 ∨ 𝑄 is a little trickier. The first three lines should certainly be filled in as shown in the next truth table:

P Q 𝑃 ∨Q
F F F
F T T
T F T
T T ?
But there may be some question about the last line. Should 𝑃 ∨ 𝑄 be true or false in the case in which 𝑃 and 𝑄 are both true? In other words,
does 𝑃 ∨ 𝑄 mean “𝑃 or 𝑄, or both” or does it mean “𝑃 or 𝑄 but not both”? The first way of interpreting the word or is called inclusive or
(because it includes the possibility of both statements being true), and the second is called the exclusive or. In mathematics, or always means
inclusive or, unless specified otherwise, so we will interpret ∨ as inclusive or. We therefore complete the truth table for 𝑃 ∨ 𝑄.

P Q 𝑃 ∨Q
F F F
F T T
T F T
T T T

Using the rules summarized in these truth tables, we can now work out truth tables for more complex formulas. All we have to do is work out
the truth values of the component parts of a formula, starting with the individual letters and working up to more complex formulas a step at a
time.

Example 1.2.1. Make a truth table for the formula ¬ ( P ∧ Q ) ∨ ¬ R:

P Q R ¬ (P ∧ Q) ∨ ¬ R
F F F T F F F T T F
F F T T F F F T F T
F T F T F F T T T F
F T T T F F T T F T
T F F T T F F T T F
T F T T T F F T F T
T T F F T T T T T F
T T T F T T T F F T

The truth values added in the last step give the truth value for the entire formula, so we will call the symbol under which they are listed (The ∨
symbol in this case) the main connective for the formula. Now that we know how to make truth table for complex formulas, we’re ready to
return to the analysis of the validity of arguments. Consider again our first example of deductive argument:

It will either rain or snow tomorrow.

It’s too warm for snow.
Therefore, it will rai.

As we have seen, if we state:

P: “It will rain tomorrow.”

Q: “It will snow tomorrow.”

Then we can represent the argument symbolically as follows:

P∨Q
¬Q
________
∴P (The symbol ∴ means therfore)

P Q P ∨ Q ¬ Q P
F F F F F T F F
F T F T T F T F
T F T T F T F T
T T T T T F T T

Looking at the truth table we see that the only row of the table in which both premises come out true is row three, and in this row the
conclusion is also true. Thus, the truth table confirms that if the premises are all true, the conclusion must also be true, so the argument is valid.

Example 1.2.2. Determine whether the following argument is valid:

Either John isn’t smart and he is lucky, or he’s smart.

John is smart.
Therefore, John isn’t lucky.

If we state:

S: “John is smart.”
 L: “John is lucky.”

Then we can represent the argument symbolically as follows:

(¬ S ∧ L) ∨ S
S
________
∴¬L

S L (¬ S ∧ L) ∨ S S ¬ L
F F T F F F F F F T F
F T T F T T T F F F T
T F F T F F T T T T F
T T F T F T T T T F T

Both premises are true in lines three and four of this table. The conclusion is also true in line three, but it is false in line four. Thus, it is possible
for both premises to be true and the conclusion false, so the argument is invalid. If fact, the table show us exactly why the argument is invalid.
The problem occurs in the fourth line of the table in which S and L are both true - in other words, John is both smart and lucky. Thus, if John is
both smart and lucky, then both premises will be true but the conclusion will be false, so it would be a mistake to infer that the conclusion must
be true from the assumption that the premises are true.
According to our truth table, this formula is false if S and L are both false, and true otherwise. But notice that this is exactly the same as the the
truth table for the simple formula L ∨ S! Because of this, we say that the formulas (¬ S ∧ L) ∨ S and L ∨ S are equivalent. Equivalent formulas
always have the same truth value no matter what statements the letters in them stand for and no matter what the truth values of those
statements are. Translating the formula L ∨ S back into English, we see that the first premise could have been stated more simply as: “John is
either lucky or smart (or both).”. But from this premise and the second premise (that John is smart), it clearly doesn’t follow that he’s not lucky,
because he might be both smart and lucky.
In analyzing deductive arguments and the statements that occur in them it is helpful to be familiar with a number of equivalences that come up
often.

De Morgan’s laws
• ¬ (P ∧ Q) is equivalent to ¬ P ∨ ¬ Q
• ¬ (P ∨ Q) is equivalent to ¬ P ∧ ¬ Q

Commutative laws
• P ∧ Q is equivalent to Q ∧ P
• P ∨ Q is equivalent to Q ∨ P

Associative laws
• P ∧ (Q ∧ R) is equivalent to (P ∧ Q) ∧ R
• P ∨ (Q ∨ R) is equivalent to (P ∨ Q) ∨ R

Idempotent laws
• P ∧ P is equivalent to P
• P ∨ P is equivalent to P

Distributive laws
• P ∧ (Q ∨ R) is equivalent to (P ∧ Q) ∨ (P ∧ R)
• P ∨ (Q ∧ R) is equivalent to (P ∨ Q) ∧ (P ∨ R)

Absorption laws
• P ∨ (P ∧ Q) is equivalent to P
• P ∧ (P ∨ Q) is equivalent to P

Double negation laws

• ¬ ¬ P is equivalent to P

Notice that because of the associative laws we can leave out parentheses in formulas of the forms P ∧ Q ∧ R and P ∨ Q ∨ R without worrying
that the resulting formula will be ambiguous, because the two possible ways of filling in the parentheses lead to equivalent formulas.

Example 1.2.3. Find a simpler formula equivalent to this formula:

¬ (Q ∧ ¬ P) ∨ P

Simplifying:
 ¬Q∨P∨P
¬Q∨P

Some equivalences are based on the fact that certain formulas are either always true or always false. For example, you can verify by making a
truth table that the formula Q ∧ (P ∨ ¬ P) is equivalent to just Q. But even before you make the truth table, you can possibly see why they are
equivalent in every line of the truth table, P ∨ ¬ P will come out true, and therefore Q ∧ (P ∨ ¬ P) will come out true when Q is also true, and
false when Q is false. Formulas that are always true, such as P ∨ ¬ P, are called tautologies. Similarly, formulas that are always false are called
contradictions. For example, P ∧ ¬ P is a contradiction.

We can now state useful laws involving tautologies and contradictions:

Tautology laws
• P ∧ (a tautology) is equivalent to P
• P ∨ (a tautology) is a tautology
• ¬ (a tautology) is a contradiction

Contradiction laws
• P ∧ (a contradiction) is a contradiction
• P ∨ (a contradiction) is equivalent to P
• ¬ (a contradiction) is a tautology

Example 1.2.4. Find a simpler formula equivalent to this formula:

¬ [P ∨ (Q ∧ ¬ R)] ∧ Q

Simplifying:

¬ [(P ∨ Q) ∧ (P ∨ ¬ R)] ∧ Q
[¬ (P ∨ Q) ∨ ¬ (P ∨ ¬ R)] ∧ Q
[(¬ P ∧ ¬ Q) ∨ (¬ P ∧ R)] ∧ Q
¬ P ∧ (¬ Q ∨ R) ∧ Q
¬ P ∧ [(Q ∧ ¬ Q) ∨ (Q ∧ R)]
¬P∧Q∧R