Math 461 Homework 1 Solutions 9/11/2020

Exercises: (§2, 3, 6, 7)
1. Let f : A → B be a function.

(a) For A0 ⊂ A and B0 ⊂ B, show that A0 ⊂ f −1 (f (A0 )) and f (f −1 (B0 )) ⊂ B0 .

(b) Show that f is injective if and only if A0 = f −1 (f (A0 )) for all subsets A0 ⊂ A.
(c) Show that f is surjective if and only if f (f −1 (B0 )) = B0 for all subsets B0 ⊂ B.
2. Let C be a relation on a set A. For a subset A0 ⊂ A, the restriction of C to A0 is the relation defined
by the subset D := C ∩ (A0 × A0 ).
(a) For a, b ∈ A, show that aDb if and only if a, b ∈ A0 and aCb.
(b) Show that if C is an equivalence relation on A, then D is an equivalence relation on A0 .
(c) Show that if C is an order relation on A, then D is an order relation on A0 .
(d) Show that if C is a partial order relation on A, then D is a partial order relation on A0 .

3. Let A and B be non-empty sets.

(a) Prove that A × B is finite if and only if A and B are both finite.
(b) Let B A denote the set of functions f : A → B. Show that if A and B are finite, then so is B A .
(c) Suppose B A is finite and B has at least two elements. Show that A and B are finite.

4. We say two sets A and B have the same cardinality if there is a bijection of A with B. In this exercise,
you will prove the Schröder–Bernstein Theorem: if there exist injections f : A → B and g : B → A,
then A and B have the same cardinality.
(a) Suppose C ⊂ A and that there is an injection f : A → C. Define A1 := A, C1 := C, and for n > 1
recursively define An := f (An−1 ) and Cn := f (Cn−1 ). Show that

A1 ⊃ C1 ⊃ A2 ⊃ C2 ⊃ A3 ⊃ · · ·

and that f (An \ Cn ) = An+1 \ Cn+1 for all n ∈ N.

(b) Using the notation from the previous part, show that h : A → C defined by
(
f (x) if x ∈ An \ Cn for some n ∈ N
h(x) :=
x otherwise

is a bijection. [Hint: draw a picture.]

(c) Prove the Schröder–Bernstein Theorem.

5. Let {0, 1}N denote the set of functions f : N → {0, 1}.

(a) Show that {0, 1}N and P(N) have the same cardinality.
(b) Let C be the collection of countable subsets of {0, 1}N . Show that C and {0, 1}N have the same
cardinality. [Hint: first construct an injection from C to ({0, 1}N )N then use Exercise 4.]
———————————————————————————————————————————–

Solutions:
1. (a) Let a ∈ A0 . Then f (a) ∈ f (A0 ) and therefore a ∈ f −1 (f (A0 )). Since a ∈ A0 was arbitrary,
we have A0 ⊂ f −1 (f (A0 )). Next, let b ∈ f (f −1 (B0 )). Then there exists a ∈ f −1 (B0 ) such that
f (a) = b. But a ∈ f −1 (B0 ) implies b = f (a) ∈ B0 . Since b ∈ f (f −1 (B0 )) was arbitrary, we have
f (f −1 (B0 )) ⊂ B0 . 

1 c Brent Nelson 2020

Math 461 Homework 1 Solutions 9/11/2020

(b) (=⇒) : Suppose f is injective and let A0 ⊂ A. By the previous part, it suffices to show
f −1 (f (A0 )) ⊂ A0 . If a ∈ f −1 (f (A0 )), then f (x) ∈ f (A0 ) and so there is some a1 ∈ A0 with
f (a) = f (a1 ). Since f is injective, we must have a = a1 ∈ A0 . Thus f −1 (f (A0 )) ⊂ A0 .
(⇐=) : We will proceed by contrapositive. Suppose f is not injective. Then there exists a1 , a2 ∈ A
with a1 6= a2 and f (a1 ) = f (a2 ). Consider A0 := {a1 }. Then f ({a1 }) = {f (a1 )} and so
a1 , a2 ∈ f −1 (f ({a1 })). Consequently, {a1 } does not equal f −1 (f ({a1 })) (it is a strict subset). 
(c) (=⇒) : Suppose f is surjective and let B0 ⊂ B. By part (a) it suffices to show B0 ⊂ f (f −1 (B0 )).
Let b ∈ B0 . Since f is surjective, we can find some a ∈ A with f (a) = b. Consequently,
a ∈ f −1 (B0 ) and b = f (a) ∈ f (f −1 (B0 )). Thus B0 ⊂ f (f −1 (B0 )).
(⇐=) : We will again proceed by contrapositive. Suppose f is not surjective. Then there exists
b ∈ B so that f (a) 6= b for all a ∈ A. Consider B0 := {b}. Since nothing in A is mapped to b by
f , we have f −1 ({b}) = ∅. Thus f (f −1 ({b})) = ∅ =6 {b}. 
2. (a) If aDb, then this means (a, b) ∈ D = C ∩ (A0 × A0 ). In particular, (a, b) ∈ C so that aCb, and
(a, b) ∈ A0 × A0 so that a, b ∈ A0 . Conversely, if a, b ∈ A0 and aCb, then the former implies
(a, b) ∈ A0 × A0 and the latter implies (a, b) ∈ C. Thus (a, b) is in their intersection, which is D,
and consequently aDb. 
(b) Let C be an equivalence relation on A and let D be its restriction to a subset A0 ⊂ A. So C
satisfies reflexivity, symmetry, and transitivity and we must show D inherits these properties. For
a ∈ A0 , we have aCa by reflexivity and consequently aDa by part (a). Thus D is reflexive. For
a, b ∈ A0 , if aDb, then aCb by part (a). By symmetry of C we have bCa and since we still have
a, b ∈ A0 , we obtain bDa by part (a). Thus D is symmetric. Finally, for a, b, c ∈ A0 , if aDb and
bDc, then we have aCb and bCc, and so aCc by transitivity of C. Using part (a) again we obtain
aDc whence D is transitive. 
(c) Let C be an order relation on A and let D be its restriction to a subset A0 ⊂ A. So C satisfies
comparability, non-reflexivity, and transitivity and we must show D inherits these properties. Let
a, b ∈ A0 with a 6= b. Then aCb by comparability, and consequently aDb by part (a); that is, D
has comparability. Let a ∈ A0 . If aDa, then aCa by part (a), which contradicts non-reflexivity
of C. Thus aDa holds for no a ∈ A0 , which means D has non-reflexivity. Finally, the proof of
transitivity follows by exactly the same argument as in part (b). 
(d) Let C be a partial order relation on A and let D be its restriction to a subset A0 ⊂ A. So C
satisfies reflexivity, antisymmetry, and transitivity and we must show D inherits these properties.
Reflexivity and transitivity follows by the same arguments as in part (b), so it suffices show D is
antisymmetric. If a, b ∈ A0 satisfy aDb and bDa, then we have aCb and bCa by part (a). Since
C is antisymmetric, we must have a = b. Thus D is antisymmetric. 
3. (a) (=⇒) : Suppose A × B is finite. Then by Corollary 6.7, there is an injective function f : A × B →
{1, 2, . . . , n} for some n ∈ N. Let a0 ∈ A and b0 ∈ B (which exist since A and B are assumed to
be non-empty), and note that the maps

ιA : A 3 a 7→ (a, b0 ) ∈ A × B
ιB : B 3 b 7→ (a0 , b) ∈ A × B

are injective. Consequently, f ◦ ιA : A → {1, 2, . . . , 2} and f ◦ ιB : B → {1, 2, . . . , n} are injective

maps as compositions of injective maps. Thus A and B are finite by Corollary 6.7.
(⇐=) : Suppose A and B are finite. By Corollary 6.7, there are injective functions f : A →
{1, 2, . . . , n} and g : B → {1, 2, . . . , m}. Note that n, m ≥ 1 since A and B are both non-empty.
Observe that the map

h : {1, 2, . . . , n} × {1, 2, . . . , m} → {1, 2, . . . , nm}

(i, j) 7→ (i − 1)m + j

is injective. Indeed, if h(i, j) = h(i0 , j 0 ) then (i − i0 )m = j 0 − j, which implies j 0 − j is divisible

by m. Since j 0 − j ∈ {−m + 1, −m + 2, . . . , −1, 0, 1, . . . , m − 2, m − 1}, this is only possible if

2 c Brent Nelson 2020

Math 461 Homework 1 Solutions 9/11/2020

j 0 − j = 0 in which case (i − i0 )m = 0. Thus j = j 0 and i = i0 and h is injective. Consider the

map k : A × B → {1, 2, . . . , nm} defined by k(a, b) := h(f (a), g(b)). We claim this is injective,
in which case A × B is finite by Corollary 6.7. Suppose k(a, b) = k(a0 , b0 ). Then h(f (a), g(b)) =
h(f (a0 ), g(b0 )). Since h is injective, we must have (f (a), g(b)) = (f (a0 ), g(b0 )). So f (a) = f (a0 )
and g(b) = g(b0 ), but since each of these functions is injective we obtain a = a0 and b = b0 . Thus
k is injective. 
(b) Let n be the cardinality of A and m the cardinality of B. We will show that there is a bijection
between B A and B n , and then use the previous part (and induction) to show B n is finite. Since A
has cardinality n, there is a bijection σ : {1, 2, . . . , n} → A. So we can define a map φ : B A → B n
by φ(f ) := (f (σ(1)), . . . , f (σ(n))) for f ∈ B A . Suppose φ(f ) = φ(f 0 ) for f, f 0 ∈ B A . Then
f (σ(j)) = f 0 (σ(j)) for each j = 1, . . . , n. This implies f = f 0 because each a ∈ A occurs in the
set {σ(1), . . . , σ(n)}. Thus φ is injective. Also, given any (b1 , . . . , bn ) ∈ B n the function f ∈ B A
defined by f (a) := bσ−1 (a) satisfies φ(f ) = (b1 , . . . , bn ). Thus φ is also surjective. So it now suffices
to show B n is finite, and we will proceed by induction on n. If n = 1, then this is immediate from
the finiteness of B. So suppose we know B n−1 is finite. Then B n = B n−1 × B, and consequently
B n is finite by part (a). Induction then concludes the proof. 
(c) We will first show A is finite. Let b1 , b2 ∈ B be distinct elements. For a fixed a ∈ A, define
fa : A → B by fa (a) = b1 and fa (a0 ) = b2 for a0 6= a.Then a 7→ fa is an injection from A into B A .
Since B A is finite, there is an injection from B A to {1, . . . , n} for some n ∈ N. The composition
of these injections, along with Corollary 6.7 shows A is finite. Next, we show B is finite. For each
b ∈ B, define gb : A → B by gb (a) := b for all a ∈ B. Then b 7→ gb is an injection from B to B A .
By the same argument as with A, this implies B is finite. 

4. (a) We will establish this series of containments by proving “An ⊃ Cn ⊃ An+1 ” via induction on n.
For n = 1, we have A1 = A, C1 = C, and A2 = f (A). So the inclusion A1 ⊃ C1 follows from the
fact that C is a subset of A, and the inclusion C1 ⊃ A2 follows from the fact that the C is the range
of f . Now assume An−1 ⊃ Cn−1 ⊃ An . Then appllying f yields f (An−1 ) ⊃ f (Cn−1 ) ⊃ f (An ),
but this is precisely the series of inclusions An ⊃ Cn ⊃ An+1 . Thus the full series of inclusions
holds by induction.
Now, we must show f (An \Cn ) = An+1 \Cn+1 for each n ∈ N. Fix n ∈ N and let a ∈ An \Cn . Then
f (a) ∈ An+1 by definition of An+1 . We also cannot have f (a) ∈ Cn+1 because Cn+1 = f (Cn )
would imply that f (a) = f (c) for some c ∈ Cn and hence a = c ∈ Cn since f is injective, a
contradiction. Thus f (a) ∈ An+1 \ Cn+1 , and so f (An \ Cn ) ⊂ An+1 \ Cn+1 . Conversely, let
b ∈ An+1 \ Cn+1 . Then An+1 = f (An ) implies there is some a ∈ An with f (a) = b. We must also
have a 6∈ Cn because otherwise b = f (a) ∈ Cn+1 , a contradiction. Thus An+1 \ Cn+1 ⊂ f (An \ Cn )
and so the sets are equal.
(b) We first show h is injective. Suppose h(x) = h(y). If x ∈ An \ Cn for some n ∈ N, then
h(y) = h(x) = f (x) ∈ An+1 \ Cn+1 by part (a). We cannot have h(y) = y because this would
require (by definition of h) that y 6∈ An \ Cn for any n, and yet y = h(y) = f (x) ∈ An+1 \ Cn+1 .
Thus we must have h(y) = f (y), and so f (y) = f (x). Since f is injective, this implies x = y. If
x 6∈ An \ Cn for all n ∈ N, then h(x) = x by definition of h. By the same reasoning as above, we
cannot have y ∈ Am \ Cm for any m, and so we have y = h(y) = h(x) = x. Thus h is injective.
Next we show h is surjective. Let y ∈ C. If y 6∈ An \ Cn for any n ∈ N, then h(y) = y and so y is
in the image of h. If y ∈ An \ Cn for some n, then we must have n > 1 since y ∈ C = C1 . Thus,
by part (a), An \ Cn = f (An−1 \ Cn−1 ). So there is some x ∈ An−1 \ Cn1 with f (x) = y. Since
x ∈ An−1 \ Cn−1 , we have h(x) = f (x) = y. Thus h is surjective.
(c) Suppose f : A → B and g : B → A are injections. Consider C := g(B) ⊂ A and note that
g ◦ f : A → C is an injection. So by part (b), there is a bijection h : A → C. Since g is an
injection, by changing the range of g we get that g : B → g(B) = C is a bijection. Hence
g −1 ◦ h : A → B is a bijection and so A and B have the same cardinality. 

5. (a) Given f ∈ {0, 1}N , define a subset a subset of the natural numbers by Af := {n ∈ N | f (n) = 1}.
We claim that f 7→ Af is a bijection {0, 1}N → P(N). If Af = Af 0 , then for each n ∈ N we have

3 c Brent Nelson 2020

Math 461 Homework 1 Solutions 9/11/2020

f (n) = f 0 (n) = 1 if n ∈ Af = Af 0 and f (n) = f 0 (n) = 0 otherwise. Thus f 7→ Af is injective.

Given A ∈ P(N), define f : N → {0, 1} by f (n) = 1 if n ∈ A and f (n) = 0 otherwise. Then
Af = A and so the map is also surjective. Thus {0, 1}N and P(N) have the same cardinality. 
(b) We will show there are injections C → {0, 1}N and {0, 1}N → C and then use the Schröder–
Bernstein Theorem. The latter is easy to define: simply send f ∈ {0, 1}N to {f } ∈ C. For the the
former, we will actually define intermediate injections C → ({0, 1}N )N → {0, 1}N .
If C ∈ C, then by Theorem 7.1 there is a surjective function fC : N → C. Changing the range
of fC to all of {0, 1}N , we can view fC ∈ ({0, 1}N )N where C is the image of f . Then for
C, C 0 ∈ C, if fC = fC 0 then in particular the image of fC (which is C) equals the image of
fC 0 (which is C 0 ). Thus C 7→ fC is an injection C → ({0, 1}N )N . It remains to show there is
an injection ({0, 1}N )N → {0, 1}N . First recall that since N × N is countably infinite, there is a
bijection g : N → N → N. Now, given f ∈ ({0, 1}N )N , we view it as a function f : N → {0, 1}N .
That is, for each n ∈ N, f (n) ∈ {0, 1}N and so f (n) : N → {0, 1}. Thus (f (n))(m) ∈ {0, 1} for
each n, m ∈ N, which means we can view f as a function f : N × N → {0, 1}. Consequently,
f ◦ g : N → {0, 1}, or f ◦ g ∈ {0, 1}N . We claim f 7→ f ◦ g is the desired injection. Indeed, if
f ◦ g = f 0 ◦ g for f, f 0 ∈ ({0, 1}N )N , then for any (n, m) ∈ N × N let k = g −1 (n, ). We have
f (n, m) = f (g(k)) = f 0 (g(k)) = f 0 (n, m). Since (n, m) ∈ N × N was arbitrary, we obtain f = f 0
and so f 7→ f ◦ g is injective. 

4 c Brent Nelson 2020