How To Write Proofs

Part I: The Mechanics of Proofs

1. Introduction
2. Direct Proof
3. Proof by Contradiction
4. Proof by Contrapositive
5. If, and Only If
6. Proof by Mathematical Induction

Part II: Proof Strategies

1. Unwinding Definitions (Getting Started)
2. Constructive Versus Exestintial Proofs
3. Counter Examples
4. Proof by Exhaustion (Case by Case)
5. What Does "Well Defined" Mean?
6. The Pigeon Hole Principle
Part I: The Mechanics of Proofs
 Introduction to Proofs
Proofs are the heart of mathematics. If you are a math major, then you must come to
terms with proofs--you must be able to read, understand and write them. What is the
secret? What magic do you need to know? The short answer is: there is no secret, no
mystery, no magic. All that is needed is some common sense and a basic
understanding of a few trusted and easy to understand techniques.

The Structure of a Proof

The basic structure of a proof is easy: it is just a series of statements, each one being
either

• An assumption or
• A conclusion, clearly following from an assumption or previously proved
result.

And that is all. Occasionally there will be the clarifying remark, but this is just for the
reader and has no logical bearing on the structure of the proof.

A well written proof will flow. That is, the reader should feel as though they are being
taken on a ride that takes them directly and inevitably to the desired conclusion
without any distractions about irrelevant details. Each step should be clear or at least
clearly justified. A good proof is easy to follow.

When you are finished with a proof, apply the above simple test to every sentence: is
it clearly (a) an assumption or (b) a justified conclusion? If the sentence fails the test,
maybe it doesn't belong in the proof.

An Example: The Irrationality of the Square Root of 2

In order to write proofs, you must be able to read proofs. See if you can follow the
proof below. Don't worry about how you would have (or would not have) come up
with the idea for the proof. Read the proof with an eye towards the criteria listed
above. Is each sentence clearly an assumption or a conclusion? Does the proof flow?
Was the theorem in fact proved?

Before we begin the proof, let's recall a few definitions. A real number is called
rational if it can be expressed as the ratio of two integers: p/q. The ancient Greeks
thought that all numbers were rational. A number that is not rational would be called
irrational. You probably believe that π is irrational. (It might surprise you that this is
not easy to prove.) When the Greeks proved that the square root of 2 is not a rational
number, the very foundations of arithmetic were called into question. This is one of
the reasons that Greek geometry subsequently flourished--all numbers could be
treated geometrically without reference to rationality.

Another fact that we will need is the Fundamental Theorem of Arithmetic. This
exciting sounding theorem is nothing more than the fact that every positive integer
has a unique representation as a product of prime numbers. The technique of proof we
will use is proof by contradiction. You do not need any specialized knowledge to
understand what this means. It is very simple. We will assume that the square root of
2 is a rational number and then arrive at a contradiction. Make sure you understand
every line of the proof.

Theorem. The square root of 2 is an irrational number.

Proof. Let's represent the square root of 2 by s. Then, by definition, s satisfies the
equation

s2 = 2.

If s were a rational number, then we could write

s = p/q

where p and q are a pair of integers. In fact, by dividing out the common multiple if
necessary, we may even assume p and q have no common multiple (other than 1). If
we now substitute this into the first equation we obtain, after a little algebra, the
equation

p2 = 2 q2.

But now, by the Fundamental Theorem of Arithmetic, 2 must appear in the prime
factorization of the number p2 (since it appears in the same number 2 q2). Since 2
itself is a prime number, 2 must then appear in the prime factorization of the number
p. But then, 22 would appear in the prime factorization of p2, and hence in 2 q2. By
dividing out a 2, it then appears that 2 is in the prime factorization of q2. Like before
(with p2) we can now conclude 2 is a prime factor of q. But now we have p and q
sharing a prime factor, namely 2. This violates our assumption above (see if you can
find it) that p and q have no common multiple other than 1.

Direct Proofs
Let's start with an example.

Example: Divisibility is Transitive

If a and b are two natural numbers, we say that a divides b if there is another natural
number k such that b = a k. For example, 2917 divides 522143 because there is a
natural number k (namely k = 179) such that 522143 = 2917 k.

Theorem. If a divides b and b divides c then a divides c.

Proof. By our assumptions, and the definition of divisibility, there are natural
numbers k1 and k2 such that
 b = a k1 and c = b k2.

Consequently,

c = b k2 = a k1 k2.

Let k = k1 k2. Now k is a natural number and c = a k, so by the definition of

divisibility, a divides c.

If P, Then Q
Most theorems (homework or test problems) that you want to prove are either
explicitly or implicitly in the form "If P, Then Q". In the previous example, "P" was
"If a divides b and b divides c" and "Q" was "a divides c". This is the standard form of
a theorem (though it can be disguised). A direct poof should be thought of as a flow of
implications beginning with "P" and ending with "Q".
P -> ... -> Q

Most proofs are (and should be) direct proofs. Always try direct proof first, unless you
have a good reason not to.

It Seems Too Easy

If you find a simple proof, and you are convinced of its correctness, then don't be shy
about. Many times proofs are simple and short.

In the theorem below, a perfect square is meant to be an integer in the form a2 where
a itself is an integer and an odd integer is any integer in the form 2a+1 where a is an
integer.

Theorem. Every odd integer is the difference of two perfect squares.

Proof. Suppose 2a+1 is an odd integer, then

2a+1 = (a+1)2 - a2.

Where's the proof? It's there. It's just very short.

Another Simple Example

Recall that a natural number is called composite if it is the product of other natural
numbers all greater than 1. For example, the number 39481461 is composite since it is
the product of 15489 and 2549.

Theorem. The number 100...01 (with 3n-1 zeros where n is an integer larger then 0)
is composite.
Proof. We can rewrite our number as 100...01 = 103n + 1 where n is an integer larger
than 0. Now use the identity a3 + b3 = (a+b)(a2 - a b + b2) with a = 10n and b = 1, to
get

(10n)3 + 1 = (10n + 1)(102n - 10n + 1).

We will be done once we have shown that both factors (10n + 1) and (102n - 10n + 1)
are greater than 1. In the first case, this is clear since 10n > 0 when n > 0. In the
second case, 102n - 10n = 10n (10n - 1) > 0, when n > 0. This completes the proof.

Make sure you understand why it was necessary to discuss the two cases at the end.

One-to-One Functions
A function f:X->Y is called one-to-one if for any pair a, b in X such that f(a) = f(b)
then a = b. Also, if f:X->Y and g:Y->Z are two functions then the composition
gf:X->Z is the function defined by gf(a) = g(f(a)) for every a in X. Note that the
composition gf is only defined if the domain of f is contained in the range of g.

Theorem. If two one-to-one functions can be composed then their composition is

one-to-one.

Proof. Let a and b be in X and assume gf(a) = gf(b). Thus, g(f(a)) = g(f(b)), and since
g is one-to-one we may conclude that f(a) = f(b). Finally, since f is one-to-one, a = b.

Roots of Polynomials
A number r is called a root of the polynomial p(x) if p(r) = 0.

Theorem. If r1 and r2 are distinct roots of the polynomial p(x) = x2 + b x + c, then

r1 + r2 = - b and r1 r2 = c.

Proof. It follows from our assumptions that p(x) will factor

p(x) = (x - r1) (x - r2)

If we expand the right hand side we get

p(x) = x2 - (r1 + r2) x + r1 r2.

Compare the coefficients above with those of p(x) = x2 + b x + c to get r1 + r2 = - b

and r1 r2 = c.

Exercises
Prove each of the following.
 1. If a divides b and a divides c then a divides b + c. (Here a, b, and c are positive
natural numbers and the definition of divisibility is given above.)
2. If a is an integer, divisible by 4, then a is the difference of two perfect squares.
3. If a and b are real numbers, then a2 + b2 >= 2 a b.
4. The sum of two rational numbers is a rational number.
5. If two onto functions can be composed then their composition is onto. (A
function f:X->Y is called onto if for every b in Y there is an element a in X
such that f(a) = b. )
6. If r1, r2, r3 are three distinct (no two the same) roots of the polynomial
p(x) = x3 + b x2 + c x+ d, then r1 r2 + r1 r3 + r2 r3 = c.

Proof by Contradiction
In a proof by contradiction we assume, along with the hypotheses, the logical
negation of the result we wish to prove, and then reach some kind of contradiction.
That is, if we want to prove "If P, Then Q", we assume P and Not Q. The
contradiction we arrive at could be some conclusion contradicting one of our
assumptions, or something obviously untrue like 1 = 0. Read the proof of the
irrationality of the square root of 2 in the introduction for an example.

Here are a few more examples.

Infinitely Many Primes

One of the first proofs by contradiction is the following gem attributed to Euclid.

Theorem. There are infinitely many prime numbers.

Proof. Assume to the contrary that there are only finitely many prime numbers, and
all of them are listed as follows: p1, p2 ..., pn. Consider the number q = p1p2... pn + 1.
This number is not divisible by any of the listed primes since if we divided pi into q,
there would result a remainder of 1 for each i = 1, 2, ..., n. Well then, we must
conclude that q is a prime number, not among the primes listed above, contradicting
our assumption that all primes are in the list p1, p2 ..., pn.

Proof by contradiction is often used when you wish to prove the impossibility of
something. You assume it is possible, and then reach a contradiction. In the examples
below we use this idea to prove the impossibility of certain kinds of solutions to some
equations.

Example: A Diophantine Equation

A Diophantine equation is an equation for which you seek integer solutions. For
example, the so-called Pythagorean triples (x, y, z) are positive integer solutions to the
equation x2 + y2 = z2. Here is another.

Theorem. There are no positive integer solutions to the Diophantine equation

x2 - y2 = 1.
Proof. (Proof by Contradiction.) Assume to the contrary that there is a solution (x, y)
where x and y are positive integers. If this is the case, we can factor the left side: x2 -
y2 = (x-y)(x+y) = 1. Since x and y are integers, it follows that either x-y = 1 and
x+y = 1 or x-y = -1 and x+y = -1. In the first case we can add the two equations to get
x = 1 and y = 0, contradicting our assumption that x and y are positive. The second
case is similar, getting x = -1 and y = 0, again contradicting our assumption.

Example: Rational Roots

There is a formula for solving the general cubic equation a x3 + b 2 c x + d = 0, that is
more complicated than the quadratic equation. But in this example, we wish to prove
there is no rational root to a particular cubic equation without have to look at the
general cubic formula.

Theorem. There are no rational number solutions to the equation x3 + x + 1 = 0.

Proof. (Proof by Contradiction.) Assume to the contrary there is a rational number

p/q, in reduced form, with p not equal to zero, that satisfies the equation. Then, we
have p3/q3 + p/q+ 1 = 0. After multiplying each side of the equation by q3, we get the
equation

p3 + p q2 + q3 = 0

There are three cases to consider. (1) If p and q are both odd, then the left hand side of
the above equation is odd. But zero is not odd, which leaves us with a contradiction.
(2) If p is even and q is odd, then the left hand side is odd, again a contradiction. (3) If
p is odd and q is even, we get the same contradiction. The fourth case--p even and q
even--is not possible because we assumed that p/q is in reduced form. This completes
the proof.

The Converse of a Theorem

The Converse of "If P, Then Q" is the assertion "If Q, Then P". For example, the
converse of "If it is my car, it's red" is "If the car is red, then its mine." It should be
clear from this example that there is no guarantee that the converse of a true statement
is true.

Proof by Contradiction is often the most natural way to prove the converse of an
already proved theorem.

The Converse of the Pythagorean Theorem

The Pythagorean Theorem tells us that in a right triangle, there is a simple relation
between the two leg lengths (a and b) and the hypotenuse length, c, of a right triangle:
a2 + b2 = c2. Perhaps you don't know that the converse is also true.
The Converse of the Pythagorean Theorem. If the (nonzero) three side lengths of a
triangle--a, b and c--satisfy the relation a2 + b2 = c2, then the triangle is a right
triangle. (Assume the Pythagorean Theorem has already been proved.)

Proof. (Proof by Contradiction.) Suppose the triangle is not a right triangle. Label the
vertices A, B and C as pictured. (There are two possibilities for the measure of angle
C: less than 90 degrees (left picture) or greater than 90 degrees (right picture).)

Erect a perpendicular line segment CD as pictured below.

By the Pythagorean Theorem, BD2 = a2 + b2 = c2, and so BD = c. Thus we have

isosceles triangles ACD and ABD. It follows that we have congruent angles CDA =
CAD and BDA = DAB. But this contradicts the apparent inequalities (see picture)
BDA < CDA = CAD < DAB (left picture) or DAB < CAD = CDA < BDA (right
picture).

Exercises
Use the method of Proof by Contradiction to prove each of the following.

1. The cube root of 2 is irrational.

2. There are no positive integer solutions to the Diophantine equation x2 - y2 =
10.
3. There is no rational number solution to the equation x5 + x4 + x3+x2+ 1 = 0.
4. If a is a rational number and b is an irrational number, then a+b is an irrational
number.
 Proof by Contrapositive
Proof by contrapositive takes advantage of the logical equivalence between "P implies
Q" and "Not Q implies Not P". For example, the assertion "If it is my car, then it is
red" is equivalent to "If that car is not red, then it is not mine". So, to prove "If P,
Then Q" by the method of contrapositive means to prove "If Not Q, Then Not P".

Example: Parity
Here is a simple example that illustrates the method. The proof will use the following
definitions.

Definitions.

1. An integer x is called even (respectively odd) if there is another integer k for

which x = 2k (respectively 2k+1).
2. Two integers are said to have the same parity if they are both odd or both
even.

For the purpose of this example we will assume as proved that each integer is either
even or odd.

Theorem. If x and y are two integers for which x+y is even, then x and y have the
same parity.

Proof. The contrapositive version of this theorem is "If x and y are two integers with
opposite parity, then their sum must be odd." So we assume x and y have opposite
parity. Since one of these integers is even and the other odd, there is no loss of
generality to suppose x is even and y is odd. Thus, there are integers k and m for
which x = 2k and y = 2m+1. Now then, we compute the sum

x+y = 2k + 2m + 1 = 2(k+m) + 1,

which is an odd integer by definition.

How Is This Different From Proof by Contradiction?

The difference between the Contrapositive method and the Contradiction method is
subtle. Let's examine how the two methods work when trying to prove "If P, Then Q".

• Method of Contradiction: Assume P and Not Q and prove some sort of

contradiction.
• Method of Contrapositive: Assume Not Q and prove Not P.

The method of Contrapositive has the advantage that your goal is clear: Prove Not P.
In the method of Contradiction, your goal is to prove a contradiction, but it is not
always clear what the contradiction is going to be at the start.
A Test For Perfect Squares
In this example, we will need two notions. An integer n is called a perfect square if
there is another integer k such that n = k2. For example, 13689 is a perfect square
since 13689 = 1172.

The second idea is the remainder and modular arithmetic. For two integers m and n, n
mod(m) = r will be the remainder resulting when we divide m into n. This means that
there is an integer q such that n = mq + r. For example, 107 mod(29) = 11 since 29
will go into 107 4 times with a remainder of 11 (or, in other words, 107 = (4)(29) +
11). Determining whether or not a positve integer is a perfect square might be
difficult. For example, is 82,642,834,671 a perfect square? First we compute
82,642,834,671 mod(4) = 3. Then use this theorem:

Theorem. If n is a positive integer such that n mod(4) is 2 or 3, then n is not a perfect

square.

Proof. We will prove the contrapositive version: "If n is a perfect square then n
mod(4) must be 0 or 1." (Do you understand why this is the contrapositive version?)
Suppose n = k2. There are four cases to consider.

1. If k mod(4) = 0, then k = 4q, for some integer q. Then, n = k2 = 16 q2 = 4(4 q2)

, i.e. n mod(4) = 0.
2. If k mod(4) = 1, then k = 4q + 1, for some integer q. Then, n = k2 = 16 q2 + 8 q
+ 1= 4(4 q2 + 2 q) + 1, i.e. n mod(4) = 1.
3. If k mod(4) = 2, then k = 4q + 2, for some integer q. Then, n = k2 = 16 q2 + 16
q + 4 = 4(4 q2 + 4 q + 1), i.e. n mod(4) = 0.
4. If k mod(4) = 3, then k = 4q + 3, for some integer q. Then, n = k2 = 16 q2 + 24
q + 9 = 4(4 q2 + 6 q + 2) + 1, i.e. n mod(4) = 1.

Exercises
Prove each of the following by the contrapositive method.

1. If x and y are two integers whose product is even, then at least one of the two
must be even.
2. If x and y are two integers whose product is odd, then both must be odd.
3. If n is a positive integer such that n mod(3) = 2, then n is not a perfect square.
4. If a and b a real numbers such that the product a b is an irrational number, then
either a or b must be an irrational number.

If, and Only If

Many theorems are stated in the form "P, if, and only if, Q". Another way to say the
same things is: "Q is necessary and sufficient for P". This means two things: "If P,
Then Q" and "If Q, Then P". So to prove an "If, and Only If" theorem, you must prove
two implications.
Example: Division
In this example we will use a very useful fact about integers, the so called Division
Algorithm: If n and m are integers, then there are two other integers’ q and r, where 0
<= r < m, and such that n = qm + r. For example, if n = 103 and m = 15, then 103 =
(6)(15) + 13. (That is, if we divide15 into 103, we get a quotient of q = 6, with a
remainder of r = 13.)

Theorem. If a is an integer, then a is not evenly divisible by 3 if, and only if, a2 -1 is
evenly divisible by 3.

Proof. Since this is an "If, and Only If" theorem, we must prove two implications.

("If") We must prove "a is not evenly divisible by 3 if a2 -1 is evenly divisible by 3".
So we assume that 3 evenly divides a2 -1 = (a-1)(a+1). Since 3 is a prime number, 3
must evenly divide either a-1 or a+1. In either case, it should be apparent that 3 cannot
evenly divide a.

("Only If"). We must prove "a is not evenly divisible by 3 only if a2 -1 is evenly
divisible by 3." This means "If a is not evenly divisible by 3, then a2 -1 is evenly
divisible by 3". This is where we use the division algorithm stated above. We can
write a = 3q + r, where r = 0, 1 or 2. Our assumption that a is not divisible by 3
implies r cannot be 0. If r =1, then a-1 = 3q and so 3 evenly divides a2 -1 = (a-1)(a+1).
A similar argument works if r = 2.

Sometimes you can prove an "If, and Only If" assertion without explicitly dividing the
proof into two parts. The next example illustrates how this might be done.

Example: A Division Rule

You probably learned in school that a positive integer n is evenly divisible by 3 if the
sum of the digits of n is divisible by 3. For example, 2620461 is evenly divisible by 3
since 2 + 6 + 2 + 0 + 4 + 6 + 1 = 21 = (3)(7). In fact, 2620461 = (3)(873487). This
condition is really necessary and sufficient.

Theorem. A positive integer n is evenly divisible by 3 if, and only if, the sum of the
digits of n is divisible by 3.

Proof. Suppose n is a positive integer whose digit representation is a0a1...ak. This

means, n = a0 + a1 10 + ...+ ak 10k. The digit sum is s = a0 + a1 + ... + ak.

Now, n -s = (a0 + a1 10 + ...+ ak 10k) - (a0 + a1 + ... + ak) = a1 9 + a2 99 + ... + ak

(99...9) (where the last term has k nines). So, clearly, n - s is divisible by 3. It follows
that n is divisible by 3 if, and only if, s is divisible by 3.

Exercises
Prove each of the following.
 1. If a is an integer, then a is not evenly divisible by 5 if, and only if, a4 -1 is
evenly divisible by 5.
2. For two integers a and b, a+b is odd if, and only if, exactly one of the integers,
a or b, is odd.
3. For two integers a and b, the product ab is even if and only if at least one of
the integers, a or b, is even.
4. A positive integer n is evenly divisible by 9 if, and only if, the sum of the
digits of n is divisible by 9.
5. A positive integer n is evenly divisible by 11 if, and only if, the difference of
the sums of the digits in the even and odd positions in n is divisible by 11.

Mathematical Induction
Let's begin with an example.

Example: A Sum Formula

Theorem. For any positive integer n, 1 + 2 + ... + n = n(n+1)/2.

Proof. (Proof by Mathematical Induction) Let's let P(n) be the statement

"1 + 2 + ... + n = (n (n+1)/2." (The idea is that P(n) should be an assertion that for any
n is verifiably either true or false.) The proof will now proceed in two steps: the
initial step and the inductive step.

Initial Step. We must verify that P(1) is True. P(1) asserts "1 = 1(2)/2", which is
clearly true. So we are done with the initial step.

Inductive Step. Here we must prove the following assertion: "If there is a k such that
P(k) is true, then (for this same k) P(k+1) is true." Thus, we assume there is a k such
that 1 + 2 + ... + k = k (k+1)/2. (We call this the inductive assumption.) We must
prove, for this same k, the formula 1 + 2 + ... + k + (k+1) = (k+1)(k+2)/2.

This is not too hard:

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

The first equality is a consequence of the inductive assumption.

The Math Induction Strategy

Mathematical Induction works like this: Suppose you want to prove a theorem in the
form "For all integers n greater than equal to a, P(n) is true". P(n) must be an assertion
that we wish to be true for all n = a, a+1, ...; like a formula. You first verify the initial
step. That is, you must verify that P(a) is true. Next comes the inductive step. Here
you must prove "If there is a k, greater than or equal to a, for which P(k) is true, then
for this same k, P(k+1) is true."
Since you have verified P(a), it follows from the inductive step that P(a+1) is true, and
hence, P(a+2) is true, and hence P(a+3) is true, and so on. In this way the theorem has
been proved.

Example: A Recurrence Formula

Math induction is of no use for deriving formulas. But it is a good way to prove the
validity of a formula that you might think is true. Recurrence formulas are notoriously
difficult to derive, but easy to prove valid once you have them. For example, consider
the sequence a0, a1, a2, ... defined by a0 = 1/4 and an+1 = 2 an(1-an) for n > 0.

Theorem. A formula for the sequence an defined above, is an = (1 - 1/22n)/2 for all n
greater than or equal to 0.

Proof. (By Mathematical Induction.)

Initial Step. When n = 0, the formula gives us (1 - 1/22n)/2 = (1 - 1/2)/2 = 1/4 = a0. So
the closed form formula gives us the correct answer when n = 0.

Inductive Step. Our inductive assumption is: Assume there is a k, greater than or
equal to zero, such that ak = (1 - 1/22k)/2. We must prove the formula is true for n =
k+1.

First we appeal to the recursive definition of ak+1 = 2 ak(1-ak). Next, we invoke the
inductive assumption, for this k, to get

ak+1 = 2 (1 - 1/22k)/2 (1 - (1 - 1/22k)/2) = (1 - 1/22k)(1 + 1/22k)/2 = (1 - 1/22k+1)/2. This

completes the inductive step.

Exercises
Prove each of the following by Mathematical Induction.

1. For all positive integers n, 12 + 22 + ... + n2 = (n)(n+1)(2n+1)/6.

2. Define a sequence a0, a1, a2 by the recursive formula an+1 = 2 an - an2. Then, a
closed form formula for an is an = 1 - (1 - a0)2n for all n = 0, 1, 2, ....
Part II: Proof Strategies
Unwinding Definitions (Getting Started)
One of the most often asked questions of students that are new to proofs is "How do I
get started?" The answer is usually simple: Unwind the definitions. First, look at
what you are being asked to prove. Does it involve a term that has been defined (in
the lecture or in the text or in the problem)? Write out the definition. What about the
assumptions? Do they involve definitions? If so, write those out. Sometimes there are
Theorems that are relevant to your problem. If so, write those out. Do not be afraid to
jot down everything you know about what you are trying to prove.

Example: The Greatest Common Divisor

The greatest common divisor of two positive integers a and b is the number
d = gcd(a, b) that satisfies two properties: (1) d evenly divides a and b and (2) if d' is
any other positive integer that evenly divides a and b, then d > d'. We can think of the
gcd as a binary operation.

Theorem. The binary operation gcd is associative, that is, for any three positive
integers a, b and c,

gcd(gcd(a, b), c) = gcd(a, gcd(b, c)).

Strategy. What do we have to prove? Two gcd's are the same. Where do we start? Let
d be one of these gcd's. Let's let d = gcd(gcd(a, b), c). What does this mean? It means
(1) d evenly divides gcd(a, b) and c and (2) if d' is any other positive integer that
evenly divides gcd(a, b) and c, then d>d'. We must prove d = gcd(a, gcd(b, c)). What
does this mean? We must prove two things: (1) d evenly divides a and gcd(b, c) and
(2) if d' is some other positive integer that evenly divides a and gcd(b, c), then d>d'.

(1) Since d divides gcd(a, b), d must divide and b. We know d divides c, so d must
divide gcd(b, c). So the first part is easy.

(2) Now we suppose d' divides a and gcd(b, c). Then, d' divides b and c, so d' must
divide gcd(a, b) too. But then by our assumption, d>d'. And this is all we needed to
prove.

Proof. Let d = gcd(gcd(a, b), c). Then d divides a, b and c, and hence divides a and
gcd(b, c). If d' divides a and gcd(b, c), then d' must divide gcd(a, b) and c, and hence
d>d'. Thus d = gcd(a, gcd(b, c)).

Example: Algebraic Numbers

A real number a is called an algebraic number if there is a nonzero polynomial p(x)
whose coefficients are all rational numbers such that p(a) = 0. An example would be
the sqrt(2), the square root of 2. If we let p(x) = x2 - 2, then p(sqrt(2)) = 0, so sqrt(2) is
an algebraic number. Pretty clearly any root of a rational number is an algebraic
number.
Theorem. If a is an algebraic number and r is a rational number, then a + r is an
algebraic number.

Strategy. What do we have to prove? a + r is an algebraic number. What does this

mean? We must prove that there is a polynomial p(x) with rational number
coefficients such that p(a+r) = 0. What is our assumption? We assume (1) a is an
algebraic number, that is, there is a polynomial q(x) with rational number coefficients
such that q(a) = 0 and (2) r = s/t where s and t are integers. Where do we start? Start
with the polynomial we have, q(x), for which q(a) = 0. Can we modify q(x) into a
polynomial, p(x), that does what we want: p(a+r) = 0? Yes! Let p(x) = q(x-r). Then
p(a+r) = q(a) = 0.

Proof. Let q(x) be the nonzero polynomial with rational coefficients for which
q(a) = 0. Then p(x) = q(x-r) is also a polynomial with rational coefficients (since r is a
rational number) and p(a+r) = 0. Hence a + r is an algebraic number.

Exercises
Prove each of the following.

1. The least common multiple of two positive integers a and b, lcm(a, b), is the
positive integer m that satisfies the two conditions: (1) a and b evenly divide m
and (2) if m' is another positive integer for which a and b divide m', then m <
m'. The least common multiple is associative, i.e. for any three positive
integers, lcm(lcm(a, b), c) = lcm(a, lcm(b, c)).
2. If a is an algebraic number and r is a rational number, then the product ra is an
algebraic number.

Constructive Versus Existential Proofs

Constructive Proofs
How would you prove 299 + 1 is a composite number? You would exhibit a
factorization:
299 + 1 = (233)3 + 1 =(233 + 1)((233)2 - 233 + 1).

In other words, to prove 299 + 1 is composite we constructed a factorization. Not

surprisingly, we call such a proof constructive.

Example: Pythagorean Triples

A Pythagorean triple is a triple of positive integers (a, b, c) that satisfies the equation
a2 + b2 = c2. For example, (3, 4, 5) is a Pythagorean triple since 32 + 42 = 52. Are there
more? Yes, there are infinitely more, just take multiples of (3,4, 5): (3k, 4k, 5k) where
k can be any positive integer. We call something like (3k, 4k, 5k) a one parameter
family of solutions. There is one parameter, namely k. Are there more solutions? Yes.
Theorem. There is a two parameter family of Pythagorean triples.

Proof. (We construct the solution.) Let a = u2 - v2 and b = 2 u v where u and v are
positive integers with u > v. Then a2 + b2 = (u2-v2)2 + (2uv)2 = u4 - 2 u2v2 + v4 + 4 u2v2
= u4 + 2u2v2 + v4 = (u2 + v2)2. Thus, (u2 - v2, 2 u v, u2 + v2), for u > v, is two
parameter family of Pythagorean triples.

Example: Rational Number

Theorem. There is a rational number that lies strictly between the square root of 10100
and the square root of 10100+1.

Proof. The square root of 10 100 is 10 50. After a little bit of trial and error, we let x =
10 50 + 10 -51, which is clearly a rational number bigger than the square root of 10 100.
To prove that x is less than the square root of 10100+1, we compute

x2 = (10 50 + 10 -51)2 = 10 100 + (2) 10 -1 + 10 -102

which is clearly less than 10100+1?

Existential Proofs
Sometimes it is possible to prove the existence of something mathematical without
actually constructing it. Why would you want to do this? Well, it could be that you
just cannot think of a constructive proof, or that a constructive proof is very long and
tedious. In any case, existential proofs are another valuable technique in proofs. Let's
look at a familiar example from the calculus.

An Example from Calculus

First let's recall the Intermediate Value and Mean Value Theorems:

Intermediate Value Theorem. If a real valued function f is continuous on the closed

interval [a, b] and if N is a number strictly between f(a) and f(b), then there exists a
number c in (a, b) such that f(c) = N.

Mean Value Theorem. If a real valued function f is continuous on the closed interval
[a, b] and f is differentiable on the open interval (a, b) then there is a number c in (a,
b) such that f'(c) = (f(b) - f(a))/(b-a). We can use the Mean Value Theorem to prove
that certain polynomials do not have more than one real root. (A root of a polynomial
p(x) is a number c such that p(c) = 0.)

Theorem. The polynomial p(x) = x3 + x - 1 has exactly one real root.

Proof. The proof is in two parts.

Part 1. (Direct Existential Proof.) First we will prove p(x) has one real root. We
appeal to the Intermediate Value Theorem with a = 0 and b = 1: p(0) = - 1 < 0 and
p(1) = 1 > 0. Since 0 (N = 0) is between -1 (=p(0)) and 1 (=p(1)), we may conclude
that there is a real number c, between 0 and 1, for which p(c) = 0.

Part 2. (Proof by Contradiction.) Now we will prove that p(x) has only one root.
Assume to the contrary that p(x) has more than one root. Let's suppose two distinct
roots c1 and c2, so p(c1) = p(c2) = 0. Then by appealing to the Mean Value Theorem,
there must be a number c between c1 and c2 for which p'(c) = (p(c2) - p(c1))/(c2 - c1) =
0. But a direct calculation shows that p'(x) = 3x2 + 1, which can never be zero since x2
>= 0 for all real numbers x. The contradiction completes the proof.

Example: Continuous Motion

Here is an example where we appeal to the Mean Value Theorem to obtain the
existence of something.

Theorem. If an object is traveling in a straight line with a differentiable position

function s(t), where t denotes the time variable, for t between a and b, then there is a
time t0, between a and b where the instantaneous velocity at t = to is equal to the
average velocity over the entire path.

Proof. The velocity function is the derivative of the position function v(t) = s'(t).
According to the Mean Value Theorem, there is a value t = t0 between a and b where
v(t0) = s'(t0) = (s(b) - s(a))/(b-a) = average velocity over the path.

Exercises
Prove each of the following.

1. 425650925 - 1 is a composite number. (Constructive proof.)

2. There is a rational number that lies strictly between 19 100 - 1 and 19 100.
(Constructive proof.)
3. sqrt(2) + sqrt(3) is an algebraic number. (See the section on Unwinding
Definitions (Getting Started) for the definition of an algebraic number.) (Hint:
First compute (sqrt(2) + sqrt(3))2.) (Constructive proof.)
4. Suppose f(x) is a real valued differentiable function (of real numbers) with f(1)
= -1 and f(2) = -2, then there is a value of x = x0, between 1 and -2 such that
the y-intercept of the tangent line to the curve at x = x0 is equal to f(t0) + t0.
(Existential proof--use the Mean Value Theorem).

Counter Examples
Counter examples play an important role in mathematics. Whereas a complicated
proof may be the only way to demonstrate the validity of a particular theorem, a
single counter example is all that is need to refute the validity of a proposed theorem.
For example, numbers in the form 22n + 1, where n is a positive integer, were once
thought to be prime. These numbers are prime for n = 1, 2, 3 and 4. But when n = 5,
we get
 225 + 1 = 4294967297 = (641)(6700417)

a composite number. Conclusion: When faced with a number in the form 22n + 1, we
are not allowed to assume it is either prime or composite, unless we know for sure for
some other reason.

A natural place for counter examples to occur is when the converse of a known
theorem comes into question. The converse of an assertion in the form "If P, Then Q"
is the assertion "If Q, Then P".

Example: From Calculus

In Calculus you learn that if a function is differentiable at a point, then it is continuous
at that point. What would the converse assert? It would say that if a function is
continuous at a point, then it is differentiable at that point. But you know this is false.
The counter example is f(x) = |x|. This function is continuous at x = 0, but it is not
differentiable at x=0. This one counter example is all we need to refute the converse.

Example: Rational & Irrational Numbers

If a and b are rational numbers, then so is a+b. The proof is very simple. By definition
of a rational number, a = p/q and b = s/t for some quadruple of integers p, q, s, and t
and such that q and t are nonzero. The sum a+b = p/q + s/t = (p t + q s)/(q t), a rational
number by definition. What would the converse say? It would assert "If a and b are
real numbers such that a + b is a rational number, then a and b are rational numbers."
But this is false. Just let a = sqrt(2) + 1, where sqrt means the square root, and
b = - sqrt(2). Neither a nor b are rational numbers, but a + b = 1, which is rational.

Exercises
1. State the converse of "If a and b are even integers then a+b is an even integer".
Show that the converse is not true by producing a counter example.
2. State the converse of "If a, b and c are real numbers such that a + b = c, then
(a+b)2 = c2". Show that the converse is not true by producing a counter
example.
3. State the converse of "If a, b and c are integers such that a divides b, then a
divides the product bc." Show that the converse is not true by producing a
counter example.
4. State the converse of "If a and b are rational numbers, then so is the product
ab". Show that the converse is not true by producing a counter example.

Proof by Exhaustion (Case by Case)

Sometimes the most straight forward, if not the most elegant, way to construct a proof
is by checking cases.
Example: Divisibility
Theorem. If n is a positive integer then n7 - n is divisible by 7.

Proof. First we factor n7 - n = n(n6 - 1) = n(n3 - 1)(n3 + 1) = n(n-1)(n2 + n + 1)(n+1)(n2

- n + 1). Now there are 7 cases to consider, depending on n = 7 q + r where r = 0, 1, 2,
3, 4, 5, 6, 7.

Case 1: n = 7q. Then n7 - n has the factor n, which is divisible by 7.

Case 2: n = 7q + 1. Then n7 - n has the factor n-1 = 7q.

Case 3: n = 7q + 2. Then the factor n2 + n + 1 = (7q + 2)2 + (7q+2) + 1 = 49 q2 + 35 q

+ 7 is clearly divisible by 7.

Case 4: n = 7q + 3. Then the factor n2 - n + 1 = (7q + 3)2 - (7q+3) + 1 = 49 q2 + 35 q +

7 is clearly divisible by 7.

Case 5: n = 7q + 4. Then the factor n2 + n + 1 = (7q + 4)2 + (7q+4) + 1 = 49 q2 + 63 q

+ 21 is clearly divisible by 7.

Case 6: n = 7q + 5. Then the factor n2 - n + 1 = (7q + 5)2 - (7q+5) + 1 = 49 q2 + 63 q +

21 is clearly divisible by 7.

Case 7: n = 7q + 6. Then the factor n + 1 = 7q +7 is clearly divisible by 7.

Exercises
Prove each of the following using a case by case analysis.

1. The "Triangle Inequality" for real numbers, |a + b| is less than or equal to |a| +
|b|. (The cases correspond to the signs (plus or minus) of a and b.)

What does "Well Defined" Mean?

Sooner or later you will have to prove that something is "well defined". So what does
this mean? Let's look at an example.

Example: Congruence Arithmetic

For integers a and b and a positive integer m, we say that a is congruent to b modulo
m, written a = b mod(m), if a - b is evenly divisible by m. Another way of saying this
is there is another integer k such that a = b + k m. For example, 1749 = 15 mod(17)
because 1749 = 15 + (102)(17). We like to think of a and b as representing the same
"number" modulo m. So there are m "numbers" in this system and they are 0, 1, 2, ...,
m-1. For example, m = 0 mod(m) so we won’t have to list 0.
It turns out that we can do arithmetic with these new "numbers". For example, we can
define addition modulo m by standard addition. But there is a potential problem. What
if a = b mod(m) and c = d mod(m), we should get the same result, modulo m, by
adding a + b or c + d.

Theorem. Addition is well defined modulo m, that is, if a = b mod(m) and c = d

mod(m), then (a+c) = (b+d) mod(m).

Strategy. What do we have to prove? (a+c) = (b+d) mod(m). What does this mean? It
means we must show there is an integer k such that a+c = (b+d) + k m. What are we
assuming? a = b mod(m) and c = d mod(m). This means there are integers k1 and k2
such that a = b + k1 m and c = d + k2m. What do we do? We can add these last two
equations together to get: (a+c) = (b+d) + (k1+k2)m. So if we were to let k = k1 + k2,
we would have what we want. Now we se what to do. So we write the proof.

Proof. By our assumption, there are integers k1 and k2 such that a = b + k1 m and c =
d + k2m. Adding these two equations together gives us (a+c) = (b+d) + (k1+k2)m,
which, by definition, means (a+c) = (b+d) mod(m).

So "well defined" means that the definition being made has no internal inconsistencies
and is free of contradictions. To better understand this idea, let's look at an example
where a definition turns out not to be well defined.

Example: Not Well Defined

Using modular arithmetic, consider division. For integers it makes since to talk about
x/2 when x is even. Does this make since modulo 2? For example, let x = 2. In mod
(2) arithmetic, the "number" 2/2 should be the unique solution (y) to the equation 2y =
2 mod (2). But, as you can see, any integer y will satisfy this equation. That is, x/2 is
not well defined.

Example:Functions Modulo m
In the previous two examples, we looked at the "numbers" modulo m. In this system
there are only m "numbers", represented by 0, 1, ..., m-1. It is traditional to call this
set Zm. For example, Z4 has 4 elements, represented by 0, 1, 2, 3. Remember, all other
integers are just other names for these 4. For example, 13 = 1 mod(4) and - 13 = 3
mod(4).

Theorem. The function f:Z4 -> Z4, given by f(x) = 2x+1 is well defined.

Strategy. It is easy to see that f(0) = 1, f(1) = 3, f(2) = 5 = 1 mod(4) and f(3) = 7 = 3
mod(4). What do we need to prove? We need to prove that f(a) = f(b) mod (4). That
is, f(a) - f(b) is evenly divisible by 4, that is, (2a+1) - (2b+1) = 2(a - b) is evenly
divisible by 4. What is our assumption? We are assuming a = b mod(4). What does
this mean? It means that a - b is evenly divisible by 4. We can see immediately that
our assumption implies 2(a - b) is divisible by 4, which is what we wanted.
Proof. If a = b mod(m), then (a - b) is divisble by 4. Hence so is (2a + 1) - (2b + 1),
that is f(a) = f(b) mod (m).

Exercises
Prove each of the following.

1. Multiplication is well defined in arithmetic modulo m. That is, if a = b

mod(m) and c = d mod(m), then (a c) = (b d) mod(m).
2. The function f:Zm -> Zm given by f(x) = x2 + x is well defined.

The Pigeon Hole Principle

The so called pigeon hole principle is nothing more than the obvious remark: if you
have fewer pigeon holes than pigeons and you put every pigeon in a pigeon hole, then
there must result at least one pigeon hole with more than one pigeon. It is surprising
how useful this can be as a proof strategy.

Example
Theorem. Among any N positive integers, there exist 2 whose difference is divisible
by N-1.

Proof. Let a1, a2, ..., aN be the numbers. For each ai, let ri be the remainder that results
from dividing ai by N - 1. (So ri = ai mod(N-1) and ri can take on only the values 0, 1,
..., N-2.) There are N-1 possible values for each ri, but there are N ri's. Thus, by the
pigeon hole principle, there must be two of the ri's that are the same, rj = rk for some
pair j and k But then, the corresponding ai's have the same remainder when divided by
N-1, and so their difference aj - ak is evenly divisible by N-1.

Example
Theorem. For any N positive integers, the sum of some of these integers (perhaps one
of the numbers itself) is divisible by N.

Proof. Consider the N numbers b1 = (a1) mod(N), b2 =(a1 + a2) mod(N),

b3 = (a1+a2+a3) mod(N), ..., bN = (a1 + ... + aN) mod(N). If one of these numbers is
zero, then we are done. Otherwise, only the N-1 numbers 1, 2, ..., N-1 are represented
in this list, and so two of them must be the same, bi = bj (say i < j). This would then
imply that (ai+1 + ... + aj) mod(N) = 0, proving our claim.

Exercises
Prove each of the following using the pigeon hole principle.

1. If a city has 10,000 different telephone lines numbered by 4-digit numbers and
more than half of the telephone lines are in the downtown, then there are two
 telephone numbers in the downtown whose sum is again the number of a
downtown telephone line.
2. If there are 6 people at a party, then either 3 of them knew each other before
the party or 3 of them were complete strangers before the party.