Assignment 4 Solutions, MATH 240

prepared by Jonah Saks

Q1, (a) Let f (x, y) = (x + y, x − y, 2x + y). We claim this function is injective. Indeed, suppose there
exists some (x1 , y1 ), (x2 , y2 ) ∈ R × R such that f (x1 , y1 ) = f (x2 , y2 ). Then (x1 + y1 , x1 − y1 , 2x1 + y1 ) =
(x2 + y2 , x2 − y2 , 2x2 + y2 ). In particular, we see that x1 + y1 = x2 + y2 and x1 − y1 = x2 − y2 . Adding these
equations together, we deduce that (x1 + y1 ) + (x1 − y1 ) = (x2 + y2 ) + (x2 − y2 ), and upon simplifying, we
obtain 2x1 = 2x2 , which implies x1 = x2 . Then, combining this fact with the equation x1 + y1 = x2 + y2 ,
we see that y1 = y2 . Hence, (x1 , y1 ) = (x2 , y2 ), as desired. This prove that f is injective.

We claim that f is not surjective. Indeed, we claim that there does not exist a pair (x, y) ∈ R × R such
that f (x, y) = (1, 1, 1). We’ll prove this by contradiction. Suppose there existed such a pair (x, y). Then
(x + y, x − y, 2x + y) = (1, 1, 1) implies in particular that x + y = 1 = x − y. Subtracting x from both
sides of the equation yields y = −y, which implies y = 0. Replacing this into our initial equation, we get
(1, 1, 1) = (x + y, x − y, 2x + y) = (x, x, 2x), upon which we notice the contradiction that, on the one hand,
it must be that x = 1, while on the other hand, 2x = 1 implies x = 1/2, our desired contradiction. This
proves f is not surjective.

Q1, (b) We claim that g is not injective. Indeed, g({1, 2}) = {1} = g({2}). This shows that two distinct
elements of P(N) get mapped by g to the same element.

We claim that g is surjective. Let Y ∈ P(N). We’d like to show that there exists some X ∈ P(N) such that
g(X) = Y . We can write Y = {y1 , y2 , ...}, a possibly (countably) infinite set. Then, define X = {2y1 , 2y2 , ...}.
It remains to verify that g(X) = Y . By the definition of g, we compute g(X) = {m ∈ N : 2m ∈ {2y1 , 2y2 , ...}}.
Clearly, 2m ∈ {2y1 , 2y1 , ...} if and only if m ∈ {y1 , y2 , ...}. Hence, g(X) = {m ∈ N : 2m ∈ {2y1 , 2y2 , ...}} =
{m ∈ N : m ∈ Y } = Y , as desired. This proves that g is surjective.

Q1, (c) We claim that h is injective. Let (x1 , y1 ), (x2 , y2 ) ∈ C be such that h(x1 , y1 ) = h(x2 , y2 ). We’d like
to show that it necessarily follows that (x1 , y1 ) = (x2 , y2 ). By definition of h, we have that xy11 = xy22 , so that
x2 y 2
y1 x2 = y2 x1 . Squaring both sides yields y12 x22 = y22 x21 , so y12 = x1 2 2 . Now combining this with x21 + y12 = 1,
2
we get that:
x21 y22 y22 x22 + y22
     
2 2 2 2 1
1 = x1 + 2 = x1 1 + 2 = x1 = x1
x2 x2 x22 x22
That is, x21 = x22 . Since x1 , x2 > 0, this implies that x1 = x2 . Finally, combining this with x1 y2 = x2 y1 , we
conclude that y1 = y2 . Hence, (x1 , y1 ) = (x2 , y2 ), as desired.

We claim that h is surjective. Let r ∈ R. We’d like to show that we can construct some pair (x, y) ∈ C
such that h(x, y) = r. If we were able to do this, then our pair (x, y) would satisfy r = h(x, y) = xy , or
equivalently, y = rx. Since our pair (x, y) should be an element of C, it must also satisfy x2 + y 2 = 1.
Combining this with our previous equation, we get that 1 = x2 + y 2 = x2 + (rx)2 = x2 + r2 x2 = x2 (1 + r2 ).
Dividing the equation 1 = x2 (1 + r2 ) by (1 + r2 ) on both sides (which we can do since 2
q 1 + r > 0 for all
1 1 1
r ∈ R), we get that x2 = 1+r 2 . Finally, since 1+r 2 > 0, we are free to define x = 1+r 2 , the positive
squareqroot. Recall that our potential pair (x, y) would necessarily satisfy y = rx, so we can simply define
1 1 1 1+r 2
y=r 1+r 2 . It’s clear that (x, y) ∈ C since x2 + y 2 = 1+r 2 + r2 1+r 2 = 1+r 2 = 1, and x > 0 since we defined

1
 q
1
q r 1+r 2
1 y
it to be the positive square root 1+r 2 . Finally, h(x, y) = r since x = q
1
= r, as desired.
1+r 2

Q2, (a) We’d like to show that χA (x) = 1 − χA (x) for all x ∈ U . We verify this for the cases that either
x ∈ A or x ∈/ A. If x ∈ A, then x ∈/ A, so by definition of the characteristic function, we have χA (x) = 1,
χA (x) = 0, and the equation becomes χA (x) = 0 = 1 − 1 = 1 − χA (x), as desired. In the case that x ∈ / A,
then χA (x) = 0 and χA (x) = 1, and the equation becomes χA (x) = 1 = 1 − 0 = 1 − χA (x), as desired.
Hence, the equation holds in both cases.

Q2, (b) We’d like to show that χA∩B (x) = χA (x)χB (x) for all x ∈ U . This time, we will prove this fact
directly simply using the definition. The product χA (x)χB (x) is equal to 1 if and only if χA (x) = 1 = χB (x),
and it is equal to 0 if either of χA (x) or χB (x) equals 0. Hence,

χA (x)χB (x) = 1 ⇔ (χA (x) = 1) ∧ (χB (x) = 1) ⇔ (x ∈ A) ∧ (x ∈ B) ⇔ x ∈ A ∩ B,

and χA (x)χB (x) = 0 otherwise. This is exactly the definition of the function χA∩B .

Q2, (c) If A and B were disjoint, then the formula χA∪B = χA + χB would work, since either x ∈ A \ B,
x ∈ B \ A, or x ∈
/ A ∪ B, and in each of these three cases, the formula above is correct. However, if A ∩ B ̸= ∅,
then for any x ∈ A ∩ B, the formula above would yield χA∪B (x) = 1 on the left-hand side while, on the
right-hand side, we would get χA (x) + χB (x) = 1 + 1 = 2. Notice, however, that this formula is incorrect
exactly when x ∈ A ∩ B. Hence, we can correct our previous formula by inserting a term −χA∩B on the
right-hand side. The formula now becomes:
(∗)
χA∪B = χA + χB − χA∩B = χA + χB − χA · χB ,

where the (∗) equality follows from part (b). Now, it remains to prove that this formula is correct. We
can consider the cases x ∈ A \ B, x ∈ B \ A, x ∈ A ∩ B, and x ∈ / A ∪ B. In the first case, the left-hand
side produces a 1 while the right-hand side produces χA (x) + χB (x) − χA∩B (x) = 1 + 0 − 0 = 1, as de-
sired. Since the formula is symmetric in the variables A and B, the formula holds in the second case for
the same reason as in the first. For x ∈ A ∩ B, the left-hand side yields 1 while the right-hand side gives
χA (x) + χB (x) − χA∩B (x) = 1 + 1 − 1 = 1, as desired. Finally, in the last case, both sides give 0. Hence, we
have shown that the formula above holds for all x ∈ U .

Alternatively, one could derive the formula using De Morgan’s law, since A ∪ B = (A ∩ B), then it
follows that χA∪B = χ(A∩B) = 1 − χA∩B , where this second equality follows from part a). Then, by part
b), we have 1 − χA∩B = 1 − χA χB , and then applying the result again from part a), we get that this equals
1 − (1 − χA )(1 − χB ) = 1 − (1 − χA − χB + χA χB ) = χA + χB − χA χB , as desired.