Basic Proof Techniques

Joshua Wilde, revised by Isabel Tecu, Takeshi Suzuki and María José Boccardi

August 13, 2013

1 Basic Notation
The following is standard notation for proofs:

• A⇒B . A implies B.

• A ⇐ B . B implies A.
Note that A ⇒ B does not mean B ⇒ A. Example: If (A) a person eats two hot dogs, she also
(B ) eats one hot dog. However, if (B ) a person eats one hot dog, that does not imply that she
also (A) eats two hot dogs.

• A ⇔ B. A implies B and B implies A.

Another way of saying this is that A holds if and only if (i ) B holds, or that A is equivalent to
B.

• ¬A. Not A, or the negation of A.

Example: If A is the event that x ≤ 10, then ¬A is the event that x > 10.

It is common to use mathematical symbols for words while writing proofs in order to write faster.
The following are commonly used symbols:

∀ For all, for any

∃ There exists

∈ Is contained in, is an element of

3 Such that, contains as an element

⊂ Is a subset of

QED Latin for quod erat demonstandum, or which was to be proven. A common way to signal to
the reader that you have successfully concluded your proof.

2 Proofs
We seek for ways to prove that A ⇒ B.

1
2 Basic Proof Techiniques

2.1 Direct Proofs

2.1.1 Deductive Reasoning

A direct proof by deductive reasoning is a sequence of accepted axioms or theorems such that A0 ⇒
A1 ⇒ A2 ⇒ · · · ⇒ An−1 ⇒ An , where A = A0 and B = An . The diculty is nding a sequence of
theorems or axioms to ll the gaps.

Example: Prove the number three is an odd number.

Proof: A number q is odd if there exists an integer m such that q = 2m + 1. Let m = 1. Then
2m + 1 = 3. Therefore three is an odd number. QED

2.1.2 Contrapositive

A contrapositive proof is just a direct proof of the negation. It makes use of the fact that the statement
A⇒B is equivalent to the statement ¬B ⇒ ¬A. For example, if (A) all people with driver's licenses
are (B ) at least 16 years old, then if you are not (¬B ) 16 years old, then you do not (¬A) have a
driver's license. So proving A⇒B is really the same as proving ¬B ⇒ ¬A.

Example: Let x and y be two positive numbers. Prove that if xy > 9, then x>3 or y > 3.

Proof: Suppose that both x≤3 and y ≤ 3. Then xy ≤ 9. QED (Here A: xy > 9, B : x > 3 or y > 3.
In order to prove A⇒B we proved ¬B ⇒ ¬A.)

2.2 Indirect Proofs

2.2.1 Contradiction

Suppose that we are trying to prove a proposition A, and we cannot prove it directly. However, we can
show that all other alternatives to A are impossible. Then we have indirectly proved that A must be
true. Therefore, the we can prove A ⇒ B by rst assuming that A 6⇒ B and nding a contradiction.

In other words, we start o by assuming that A is true but B is not. If this leads to a contradiction,
then either B was actually true all along, or A was actually false. But since we assume A is true, then
it must be that B is true, and we have a proof by contradiction.
√
Example: Prove that 2 is an irrational number.
√
Proof: Suppose not. Then 2 is a rational number, so it can be expressed in the form pq , where p and
q are integers which are not both even. This implies that
p2
2= ⇒ 2q 2 = p2 ,
q2
which implies that p2 is even, which in turn implies that q2 is not even. The fact that p2 is even also
implies that p is even, so there exists a integer m such that 2m = p. This implies

4m2 = p2 = 2q 2 ⇒ q 2 = 2m2 ,

which means that q is even, a contradiction. QED

2.2.2 Induction

Induction can only be used for propositions about integers or indexed by integers. Consider a list
of statements indexed by the integers. Call the rst statement P (1), the second P (2), and the nth
Math Camp 3

statement P (n). If we can prove the following two statements about the sequence, then every statement
in the entire sequence must be true:

1. P (1) is true.

2. If P (k) is true, then P (k + 1) is true.

Induction works because by 1., P (1) is true. By 2., P (2) is true since P (1) is true. Then P (3) is true
by 2. again, and so is P (4) and P (5) and P (6), until we show that all the P 's are true. Notice that
the number of propositions need not be nite.

1
Example: Prove that the sum of the rst n natural numbers is
2 n(n + 1).
P1 Pk
Proof: Let n = 1. Then 21 ·1(1+1) = j=1 j = 1. Now let n = k , and assume that j=1 j = 12 k(k+1).
We add k + 1 to both sides to get
k+1  
X 1 1 1
j = k(k + 1) + k + 1 = k + 1 (k + 1) = (k + 1) ((k + 1) + 1) .
j=1
2 2 2

QED

2.3 Epsilon-Delta Arguments

A lot of denitions and proofs in real analysis use the " and δ " concept. For example, recall the
denition of the limit of a function: We write limx→p f (x) = q if for every  > 0 we can nd δ > 0
such that |f (x) − q| <  for all x for which |x − p| < δ .

It is important to get the quantiers correct: For every  there exists δ such that... This means that
δ will change with  - for some value 1 we'll be able to nd δ1 so that the statement holds, and for
some other value 2 we'll nd δ2 which may dierent from δ1 . The opposite would be "there exists δ
such that for every  it holds that...". Here there is only one δ which has to t all dierent values of
.
When we prove a statement involving an -δ denition, we start by ascribing a xed, but unknown,
value to  ("x  > 0"). Then we try to nd value of δ that makes the statement in question come
true. This δ will usually be a function of . This completes the proof since  could have been anything:
For every  we have found a δ such that the statement holds.

When we want to show that a certain -δ statement does not hold, we usually choose one particular
 for which the statement should not be true ("Let  = 0.5"). Then we proceed by contradiction: We
pretend there exists some δ that ts our  and show that this leads to a contradiction. Ergo, no δ can
t our particular . Therefore it is not true that for all  we can nd a tting δ .

Example 1 limx→0 x2 + 1 = 1.
Prove that
2 2
Proof: Fix  > 0. How small does δ have to be such that |(x + 1) − 1| = |x | <  for all x for which
√ √ 2
√ 2 2 2
√ 2
|x| < δ ? δ =  works: If |x| <  then |x| < ( ) . |x| = |x | and ( ) = , therefore |x2 | < .
QED.

(
1 if x ≥ 0
Example 2 Show that f (x) = is not continuous.
−1 if x < 0
Recall that f is continuous at a point p if for every  > 0 we can nd δ > 0 such that |f (x) − f (p)| < 
for all x for which |x − p| < δ . f is probably not continuous is at 0. We therefore take p = 0 and show
that the above denition of continuity does not hold. It suces to show that for one particular  we
4 Basic Proof Techiniques

cannot nd a tting δ. We prove this by contradiction.

Proof: Let  = 1. δ such that |f (x) − f (0)| <  for all x for which |x| < δ .
Suppose there exists
Plugging in, this means |f (x) − 1| < 1 for all x for which |x| < δ . Set x = − 2δ . Then |x| < δ but
|f (x) − 1| = | − 1 − 1| = 2 which is not < 1. This contradicts |f (x) − 1| < 1 for all x for which |x| < δ .
Therefore there exists not δ such that the denition of continuity becomes true. f is not continuous
at 0 and therefore not a continuous function. QED.

Exercises

1. Negate the denition of convergence for a sequence. (The denition is in Real Analysis part.)

2. Negate the denition of continuity.

(
1 if x 6= 0
3. Let f (x) = . Prove that f is not continuous at 0.
1.1 if x = 0
1
4. Prove that limx→0 x 6= K for any number K.
5. Let f and g be two continuous functions from R to R. Show that f +g is continuous.
Math Camp 5

3 Homework
Prove the following by direct proof.

1. n(n + 1) is an even number.

1
2. The sum of the rst n natural numbers is
2 n(n + 1).

3. If 6x + 9y = 101, then either x or y is not an integer.

Prove the following by contrapositive.

1. n(n + 1) is an even number.

2. If x + y > 100, then either x > 50 or y > 50.

Prove the following by contradiction.

1. n(n + 1) is an even number.

√
2. 3 is an irrational number.

3. There are innitely many prime numbers.

Prove the following by induction.

1. n(n + 1) is an even number.

2. 2n ≤ 2n .
Pn 2 1
3. i=1 i = 6 n (n + 1) (2n + 1).

4. The sum of the rst n odd integers is n2 (This is the rst known proof by mathematical induction,
attributed to Francesco Maurolico. Just in case you were interested.)

Find the error in the following argument, supposedly by induction:

If there is only one horse, then all the horses are of the same color. Now suppose that within any
set of n horses, they are all of the same color. Now look at any set of n + 1 horses. Number them
1, 2, 3, . . . , n, n + 1. Consider the sets {1, 2, 3, . . . , n} and {2, 3, 4, . . . , n + 1}. Each set is a set of n
horses, therefore they are all of the same color. But these sets overlap, therefore all horses are the
same color.

Prove the following (solution in Analysis solution sheet):

1. Let f and g be functions from Rk to Rm which are continuous at x. Then h = f −g is continuous

at x.

2. Let f and g be functions from Rk to Rm which are continuous at x. Then h = fg is continuous

at x.

In rst semester micro you will be introduced to preference relations. We say that x  y, (read x
is weakly preferred to y) if x is at least as good as y to the agent. From this, we can derive two
important relations:
6 Basic Proof Techiniques

• The strict preference relation, , dened by x  y ⇔ x  y but not y  x. The strict preference
relation is read  x is strictly preferred to y .
• The indierence relation, ∼, dened by x∼y⇔xy and y  x. The indierence relation is
read  x is indierent to y .

We say that a preference relation is rational if:

• ∀ x, y , either xy or y  x.

• ∀ x, y, z , if xy and y  z, then x  z.

Prove the following two statements given that preferences are rational:

1. If xy and y  z, then x  z.

2. If x∼y and y ∼ z, then x ∼ z.