MAT 2125 – Homework 1 – Solutions

(due at midnight on January 25, in Brightspace)

1 Practicing LATEX
1. We could re-write the English sentence “For every real number, there exists a bigger real number” as
∀x ∈ R, ∃ y ∈ R : y > x. Do a similar translation of the English sentence “For every integer, there exists
a smaller integer.”
R 10
2. Compute 0 sin(x) cos(x)dx, showing at least two intermediate steps. Using the “align” environment or
otherwise, vertically align the “=” sign between steps.

Solution: For both of the following, take a look at the provided .tex source file if you ran into difficulties.
1. We write:

∀x ∈ N ∃y ∈ N : y < x.

2. We compute:
Z 10 Z 10
1
sin(x) cos(x)dx = sin(2x)dx
0 2 0
10
1
=− cos(2x)
4 0
1
≈ − (−0.839 − 1)
4
≈ 0.460.

2 Cardinality
Show that there exists a bijection between Z and Q. Hint: If you can find a surjection in both directions, then
you have shown that a bijection exists. This might be easier.

Proof: Write Q = { m n : m, n ∈ Z, n > 0, GCD(m, n) = 1}, where GCD(m, n) is the greatest common di-
visor of m, n. Define the map f : Q → Z by f ( m
n ) = m. To see that this is surjective, note that for all m ∈ Z,
m m
1 ∈ Q and f ( 1 ) = m.

Next, we define the map g : Z 7→ Q according to three cases:

1. For numbers of the form 2a 3b with a, b ∈ {0, 1, 2, . . .}, set g(2a 3b ) = ab .
2. For numbers of the form −2a 3b with a, b ∈ {0, 1, 2, . . .}, set g(−2a 3b ) = − ab .
3. For all other numbers n, set g(n) = 0.
We need to check that g is well-defined, and then that it is surjective. To see that it is well-defined, we note
that integers have unique prime decompositions, and 2, 3 are prime. This means that every number can have
at most one decomposition of the form ±2a 3b , so every number is in at most one case. It is also clear that
every number n must be in at least one case. Thus, every number belongs to exactly one case, so it is well-defined.

m
To check that g is surjective, we consider some n ∈ Q and again consider three cases:
m m
1. n > 0: g(2m 3n ) = n.
m m
2. n < 0: g(−2m 3n ) = n.
m m
3. n = 0: g(5) = n. 

1
3 Calculations with Axioms
1. Using only the field axioms, show that the multiplicative identiy is unique. That is, show that if a, b are
both multiplicative identities, then in fact a = b.

Proof: Let a, b be two multiplicative identities in a field. Since a is a multiplicative identity,

ab = b.
Since b is a multiplicative identity,
ab = a.
Combining these two equations,
b = ab = a.
This completes the proof. 
2. Using only the field axioms, show that (2x − 1)(2x + 1) = 4x2 − 1. Note: The field axioms don’t define 2
or 4 are. Please take these to be shorthands for 2 = 1 + 1 and 4 = 1 + 1 + 1 + 1.

Proof: Each equality is labeled with the field axiom used, as in the course notes:
AM1
(2x − 1)(2x + 1) = 2x(2x + 1) + (−1)(2x + 1)
AM1
= (2x)(2x) + (1)2x + (−1)(2x) + (−1)(1)
AM1
= (2x)(2x) + (1 + (−1))2x + (−1)(1)
A4
= (2x)(2x) + (−1)(1)
A3
= (2x)(2x) − 1
M1
= ((2)(2))(x2 ) − 1
= ((1 + 1)(1 + 1))(x2 ) − 1
AM1
= (1(1 + 1) + 1(1 + 1))x2 − 1
M3
= 4x2 − 1.
This completes the proof. 
3. Using only the order axioms, usual arithmetic manipulations, and inequalities between concrete numbers,
prove the following: If x ∈ R satisfies x <  for all  > 0, then x ≤ 0. Note: The order axioms in the
notes don’t give concrete inequalities such as e.g. 1 > 0, but we will show some of these in videos, class
or DGD. For the purposes of this question you can take obvious inequalities between specific integers as
given. That is, you could take 3 > 1 as given, but should justify x < 2x.

Proof: Assume first that x > 0. By O4 (and the fact that 0 < 21 < 1), this means
   
1 1
x> ·0=0
2 2
x
as well. By O3, since 2 > 0,
x x x
< + = x.
2 2 2
Putting together these two sequences of inequalities, we have
x
0 < < x.
2
But then we have found some number  = x2 > 0 so that x > ; this contradicts the original assumption.
Thus, we conclude that our original assumption x > 0 is false; by O1, we conclude x ≤ 0. 

2
4. Show that there exists some x ∈ R satisfying x2 + √
x = 5. Hint: Find an interval [a, b] for which a2 + a < 5
2
and b + b > 5, then try to adjust the proof that 2 exists.

Proof: Let’s follow the hint. Consider the interval I = [0, 10], define S = {x ∈ I : x2 + x < 5},
and define A = sup S. Note that for x ∈ [0, 1], x2 + x − 5 ≤ 12 + 1 − 5 = −3 < 0, so A ≥ 1. Similarly, for
x ∈ [9, 10], x2 + x − 5 ≥ 92 + 9 − 5 > 0, so A ≤ 9.

Claim: A2 + A = 5. This is shown in two parts: first we show that A2 + A ≤ 5, then we show
that A2 + A ≥ 5.

We show that A2 + A ≤ 5 by contradiction. Let’s assume A2 + A > 5. Then, by an earlier part of


this homework, there exists some 0 <  < 1 so that A2 + A > 5 + . But then for all 0 < δ < 100 , we have

(A − δ)2 + (A − δ) = A2 − 2Aδ + δ 2 + A − δ
≥ A2 − (2)(10)(δ) + A − δ
≥ A2 + A − 21δ
> A2 + A − 
> 5.

Furthermore, since A ≥ 1 and δ ≤ 0.01, we know that A − δ ∈ I. Thus, in this case A − 100 < A is also
an upper bound on S, contradicting the fact that A is defined to be the least upper bound on S. We
conclude that A2 + A ≤ 5.

Next, we show that A2 + A ≥ 5 by contradiction. Let’s assume A2 + A < 5. Then, by an earlier


part of this homework, there exists some 0 <  < 1 so that A2 + A < 5 − . But then for all 0 < δ < 100 ,
we have

(A + δ)2 + (A + δ) = A2 + A + (2A + 1 + δ)δ

≤ A2 + A + 22δ
< A2 + A − 
> 5.

Furthermore, since A ≤ 9 and δ ≤ 0.01, we know that A + δ ∈ I. Thus, in this case A + 100 ∈ S and

A + 100 > A, contradicting the fact that A is defined to be an upper bound on S. We conclude that
A2 + A ≤ 5.

Having shown A2 + A ≤ 5 and A2 + A ≥ 5, we conclude that A2 + A = 5. 

5. Let A, B ⊂ R and define C = {x − y : x ∈ A, y ∈ B}. Prove that inf(C) = inf(A) − sup(B).

Proof: We prove the statement in two parts: first we show that inf(C) ≤ inf(A) − sup(B), then we
show that inf(C) ≥ inf(A) − sup(B).

To prove the first part, fix  > 0. Then there exists some a ∈ A and b ∈ B so that a ≤ inf(A) + 2 ,
b ≥ sup(B) − 2 . Thus, a − b ∈ C and a − b ≤ inf(A) − sup(B) + . Thus, inf(C) ≤ inf(A) − sup(B) + .
Applying the previous question in the homework, we conclude that inf(C) ≤ inf(A) − sup(B).

To prove the second part, consider c ∈ C. Then there exists a ∈ A, b ∈ B so that c = a − b. By

the definition of the infimum and supremum, a ≥ inf(A) and b ≤ sup(B), so c = a − b ≥ inf(A) − sup(B).
We conclude that inf(C) ≥ inf(A) − sup(B).

Since we have shown inf(C) ≤ inf(A) − sup(B) and inf(C) ≥ inf(A) − sup(B), we conclude by the
first order axiom inf(C) = inf(A) − sup(B). 

3
 6. Consider a set S with 0 ≤ sup(S) = A < ∞ and A ∈ / S. Show that for all  > 0, S ∩ [A − , A] is nonempty.
Using this fact or otherwise, conclude that in fact S ∩ [A − , A] is infinite.

Proof: We prove the first claim by contradiction. Assume there exists some  > 0 so that S ∩ [A − , A]
is empty. Since A is an upper bound for S, we also know that S ∩ (A, ∞) is empty. Thus, S ∩ [A − , ∞)
is empty. But this means that A −  < A is an upper bound for s, contradicting the fact that A is the
least upper bound for S. We conclude that in fact S ∩ [A − , A] is not empty.

We also prove the second part by contradiction. Assume there exists some  > 0 so that S ∩ [A − , A]
is empty. Then we can enumerate its elements, {b1 , . . . , bn }. Let B = max(b1 , . . . , bn }. Since A ∈
/ S, we
know that b1 , . . . , bn < A. Since B is a maximum of finitely many elements, this means that B < A as
well. But then A > A − A−B 2 > B, so [A − A−B
2 , A] ∩ S is empty. But this is impossible, by the first part
of the question. This completes the proof. 

4 Induction
Somebody walks up to you with a proof by induction of the statement “For any integer N ∈ N, all collections
of N sheep are the same colour,” as follows:

• Notation: Let x1 , x2 , . . . , be the colours of all sheep in the world, put in some order.
• Base Case: Obviously the first sheep is a single colour, x1 .

• Inductive Case: Assume that the statement is true up to some integer n. By the inductive assumption,
the collection of the first n sheep {x1 , . . . , xn } are one colour (label this “colour 1”), and the collection of
the last n sheep {x2 , . . . , xn+1 } are also one colour (label this “colour 2” - note that we haven’t yet shown
it is the same colour as the first collection). Since {x2 , . . . , xn } are in both sets, we must have that “colour
1” and “colour 2” are the same, and so {x1 , . . . , xn+1 } are all one colour.

Explain why this purported proof fails by identifying and explaining a (significant) false statement. Note: we
are asking for an important, actually-false statement, not merely something like a typo or insufficiently-formal
justification for an assertion.

Solution: The critical error is in the following part of the argument, in the case n = 1:
“the collection of the first n sheep {x1 , . . . , xn } are one colour, and the collection of the last n sheep
{x2 , . . . , xn+1 } are also one (possibly different) colour. Since {x2 , . . . , xn } are in both sets, both sets
must in fact be the same colour, and so {x1 , . . . , xn+1 } are all one colour.”

Consider the case n = 1. Then the collection {x2 , . . . , xn } is actually empty, and so we can’t conclude that the
two sets {x1 }, {x2 } share any sheep, and so we can’t conclude that they are the same colour.