Cardinality

Shameek Paul

Let A and B be sets. Then B A denotes the set of all functions from A to B.

Proposition 1. Let X be a set. Then we have that 2X ' P(X).

Proof. Define ϕ : 2X → P(X) as ϕ(f ) = Sf where Sf = {x ∈ X : f (x) = 1}.

We call the set Sf the support of f . Define ψ : P(X) → 2X as ψ(A) = CA
where CA : X → 2 is given by CA (x) = 1 ⇐⇒ x ∈ A. We call the function CA
the characteristic function of the set A. We claim that ϕ and ψ are inverses of
each other.
Let f ∈ 2X . Then (ψ ◦ ϕ)(f ) = ψ(Sf ) = CSf . Let x ∈ X. Then CSf (x) =
1 ⇐⇒ x ∈ Sf ⇐⇒ f (x) = 1. So for every x ∈ X we see that CSf (x) = f (x)
and hence CSf = f . Thus, it follows that ψ ◦ ϕ is the identity. Let A ∈ P(X).
Then (ϕ ◦ ψ)(A) = ϕ(CA ) = SCA = {x ∈ X : CA (x) = 1} = {x ∈ X : x ∈ A} =
A = IP(X) (A). Thus, it follows that ϕ ◦ ψ is the identity.

Definition 1. We say that a set A is countable if A ≤ N, i.e., there exists a

subset B ⊆ N such that A ' B. This is equivalent to saying that there exists
an injective function from A to N.

Any subset of N is countable. If A is a finite set. Then there exists n ∈ N

such that A ' n. As n ⊆ N, we see that A is countable.
The map f : N × N → N given by f (r1 , r2 ) = 2r1 3r2 is an injective function.
So we see that N × N is countable.
Let p1 , . . . , pk be distinct primes. The map g : N × · · · × N → N given by
g(r1 , . . . , rk ) = pr11 . . . prkk is injective. So we see that N × · · · × N is countable.

Proposition 2. Suppose f : A → B is an injective function and B is countable.

Then A is countable.

Proof. As B is countable, there exists an injective function g : B → N. Let

h = g ◦ f . Then h is an injective function from A → N. So we see that A is
countable.

1
 In particular, we see that every subset of a countable set is countable. Show
that if A ⊆ B and if A is not countable, then B is also not countable.

Proposition 3. Suppose A and B are countable. Then A × B is countable.

Proof. There exist injective functions f : A → N and g : B → N. Define

h : A × B → N × N as h(a, b) = f (a), g(b) . We claim that h is injective.
Suppose h(a, b) = h(a0 , b0 ). Then f (a), g(b) = f (a0 ), g(b0 ) . So f (a) = f (a0 )

and g(b) = g(b0 ). As f and g are injective, we see that a = b and a0 = b0 . Thus,
we get that (a, b) = (a0 , b0 ). As N × N is countable, by Proposition 2 we see that
A × B is countable.

Corollary 1. A finite product of countable sets is countable.

Proof. Use induction to prove the following statement: Suppose A1 , . . . , Ak are

countable. Then A1 × · · · × Ak is countable.

Remark 1. Let S be a collection (set) of non-empty sets. Then the axiom of

choice says that there exists a function f : S → ∪ S such that for every A ∈ S,
we have that f (A) ∈ A. Such a function is called a choice function.

The axiom of choice is equivalent to Zorn’s lemma. We will not prove this.

Definition 2. A collection S of sets is said to be indexed by a set J if there

exists an onto map J → S. For every j ∈ J we denote the image of j by Aj . So
we see that S = {Aj : j ∈ J}.

The indexed version of the axiom of choice says that if S = {Aj : j ∈ J} is an

indexed collection of non-empty sets, then there exists a function f : J → ∪ S
such that for every j ∈ J we have that f (j) ∈ Aj .

Definition 3. Let g : Y → X be an onto function. A function f : X → Y is

said to be a section of g if g ◦ f = IX .

If f : X → Y and g : Y → Z are two functions such that g ◦ f is injective,

show that f is injective. It follows that every section is injective.

Theorem 1. Let A and B be sets and A 6= ∅. Then A ≤ B if and only if there

exists an onto function g : B → A.

Proof. Suppose A ≤ B. Then there exists f : A → B such that f is injective.

Let C = f (A). Then A ' C. Let h : C → A be a bijection and a0 ∈ A. Define
g : B → A as g(b) = h(b) if b ∈ C and g(b) = a0 if b ∈
/ C. As g(C) = h(C) = A,
we see that g is onto.

2
 Suppose there exists an onto function g : B → A. As g is onto, for every
a ∈ A we see that g −1 (a) 6= ∅. By the axiom of choice, it follows that there exists
a function f : A → B such that for every a ∈ A we have that f (a) ∈ g −1 (a).
Then g ◦ f = IA and so f is a section of g. So it follows that f is injective.

Corollary 2. Let A be a set. Then A is countable if and only if there exists an

onto function g : N → A.

Prove Corollary 2 without using the axiom of choice. Use the fact that N is
a well-ordered set.

Proposition 4. Let A be countable and f : A → B be an onto function. Then

B is countable.

Proof. There exists an onto function g : N → A. Let h = f ◦ g. As f is onto,

we see that h : N → B is an onto function. So we see that B is countable.

Theorem 2. Let S be a countable collection of countable sets. Then ∪ S is

countable.

Proof. There is an onto function h : N → S. For every A ∈ S there exists an

onto function fA : N → A. Define g : N × N → ∪ S as g(m, n) = fh(m) (n). We
claim that g is onto. Let x ∈ ∪ S. Then there exists A ∈ S such that x ∈ A. As
h is onto, there exists m ∈ N such that h(m) = A. As fA is onto, there exists
n ∈ N such that fA (n) = x. Thus, we see that g(m, n) = fh(m) (n) = fA (n) = x.
This proves our claim.

We denote the set of positive natural numbers by N+ .

Proposition 5. Z and Q are countable.

Proof. The map N × N → Z given by f (m, n) = m − n is an onto map. By

Proposition 3 we see that N × N is countable. So by Proposition 4 it follows
that Z is countable.
The map f : Z × N+ → Q given by f (m, n) = m/n is an onto map. By
Proposition 3 we see that Z × N+ is countable. So by Proposition 4 it follows
that Q is countable.

Corollary 3. The polynomial ring Q[X] \ {0} is countable.

Proof. For every k ∈ N let Ak be the set of polynomials in Q[X] whose degree
is k. Then Ak ' Q∗ × Qk . So by Proposition 1 we see that Ak is countable. So
S
by Theorem 2 we see that Q[X] \ {0} = {Ak : k ∈ N } is countable.

3
 By using Theorem 2, we see that Q[X] is countable. Show that for every
commutative ring R, we have that R[X] is countable if and only if R is countable.

Definition 4. We say that a ∈ C is algebraic if there exists f ∈ Q[X] \ {0} such

that f (a) = 0. An element a ∈ C is called transcendental if a is not algebraic.

Theorem 3. Let A be the set of all algebraic numbers. Then A is countable.

Proof. Let S = Q[X] \ {0}. By Corollary 3 we see that S is countable. For

every f ∈ S let Zf be the finite set {x ∈ C : f (x) = 0}. Then it follows that
S
A = { Zf : f ∈ S }. So by Theorem 2 we see that A is countable.

We will soon show that R is not countable. Assuming this fact, show that
the set C \ A of transcendental numbers is not countable.

Theorem 4. N is the smallest infinite set. This means that if A is any infinite
set, then we have that N ≤ A.

Proof. Let A be an infinite set. We will define a function f : N → A by

induction. Let S = { n ∈ N : f (n) is defined }. If A = ∅, then A will be finite.
Thus, we see that A 6= ∅. Let a0 ∈ A. We define f (0) = a0 and so 0 ∈ S.
Suppose n ∈ N and {0, . . . , n} ⊆ S. If A = { f (0), . . . , f (n) }, then A will be
finite. Thus, we see that A * { f (0), . . . , f (n) } and so there exists an+1 ∈ A
such that an+1 ∈
/ { f (0), . . . , f (n) }. We define f (n+1) = an+1 and so n+1 ∈ S.
By the construction of the function f we see that if m, n ∈ N such that
m < n, then f (m) 6= f (n). Thus, it follows that f is injective and so N ≤ A.

Show that an infinite product of countable sets can be countable. We will

now show that an infinite product of countable sets may not be countable.

Theorem 5. The set 2N is not countable (by Cantor’s diagonal argument).

Proof. We will show that there doesn’t exist any onto function ϕ : N → 2N . Let
ϕ : N → 2N be a function. For every k ∈ N, let ϕ(k) = fk . Define g : N → 2
as g(n) = 1 − fn (n) for every n ∈ N. Suppose g ∈ Image ϕ. Then there exists
k ∈ N such that fk = g and so we get that fk (k) = g(k). Then g(k) = 1 − g(k)
and so 2g(k) = 1. Thus, we get the contradiction that g(k) ∈/ {0, 1}. So we see
that g is not in the image of ϕ. This shows that ϕ is not onto.

Definition 5. Let X and Y be sets. We say that X < Y if X ≤ Y and X 6' Y .

By Proposition 1 and Theorem 5 we see that P(N) is not countable and so

N 6' P(N). The map N → P(N) given by n 7→ {n} is injective and so N ≤ P(N).
Thus, it follows that N < P(N). A similar argument is used in the next result.

4
Theorem 6. For any set X we have that X < P(X).

Proof. We have an injective function X → P(X) given by x 7→ {x} and so

X ≤ P(X). To show that X 6' P(X), we will show that there doesn’t exist
any onto function ϕ : X → P(X). Let ϕ : X → P(X) be a function. For every
x ∈ X, let ϕ(x) = Ax . Let B = {x ∈ X : x ∈ / Ax }. Suppose there exists x0 ∈ X
0
such that B = ϕ(x ). Then we have that B = Ax0 . If x0 ∈ B, then x0 ∈ Ax0 . As
x0 does not satisfy the condition to belong to B, we get the contradiction that
x0 ∈
/ B. If x0 ∈
/ B, then x0 ∈/ Ax0 . As x0 satisfies the condition to belong to B,
we get the contradiction that x0 ∈ B. So we see that B is not in the image of ϕ
and hence ϕ is not onto.

Let f : A → A be an injective map. Then we have that A ' f (A). Suppose

B is a set such that f (A) ⊆ B ⊆ A. Then we see that A ≤ B and B ≤ A.

Lemma 1. Let f : X → Y be injective. Then for every A, B ⊆ X we have that

f (A \ B) = f (A) \ f (B).

Proof. Let y ∈ f (A \ B). Then there exists x ∈ A \ B such that y = f (x). As

x ∈ A, we see that y ∈ f (A). Suppose y ∈ f (B). Then there exists x0 ∈ B
such that y = f (x0 ). As f is injective and f (x) = f (x0 ), we see that x = x0 .
So we get the contradiction that x ∈ B. Thus, we see that y ∈ / f (B) and so
y ∈ f (A) \ f (B).
Let y ∈ f (A) \ f (B). As y ∈ f (A), there exists x ∈ A such that y = f (x).
Suppose x ∈ B. Then we get the contradiction that y ∈ f (B). So we see that
x ∈ A \ B and hence y ∈ f (A \ B).

Lemma 2. Let A0 ⊇ A1 ⊇ A2 ⊇ A3 ⊇ . . . ⊇ An ⊇ An+1 ⊇ . . . be a decreasing

sequence of sets and let A0 = ∩ {An : n ∈ N}. Then we get the following
partition of A = A0 :

A = A0 ∪ (A0 \ A1 ) ∪ (A1 \ A2 ) ∪ (A2 \ A3 ) ∪ . . . ∪ (An \ An+1 ) ∪ . . .

Proof. Let x ∈ A. Suppose x ∈ / A0 . Then x ∈ / ∩ {An : n ∈ N} and so there

exists m ∈ N such that x ∈ / Am . Suppose x ∈ An . If n ≥ m, then An ⊆ Am
and so we get the contradiction that x ∈ Am . Thus, we see that m is an upper
bound of the set S = {n ∈ N : x ∈ An }. As S ⊆ m, it follows that S is a finite
set and so it has a maximum. If k = max S, it follows that x ∈ Ak \ Ak+1 .
Suppose there exist k, l ∈ N such that k < l and (Ak \Ak+1 ) ∩ (Al \Al+1 ) 6= ∅.
Let a ∈ A such that a ∈ (Ak \ Ak+1 ) ∩ (Al \ Al+1 ). As k < l, we have that
k + 1 ≤ l and so Al ⊆ Ak+1 . As a ∈ Al we get the contradiction that a ∈ / Ak+1 .

5
Suppose there exists k ∈ N such that x ∈ (Ak \ Ak+1 ) ∩ A0 . As x ∈ A0 , we get
the contradiction that x ∈ Ak+1 .

Theorem 7. Let f : A → A be an injective map. Suppose B is a set such that

f (A) ⊆ B ⊆ A. Then we have that A ' B.

Proof. We have that

A ⊇ B ⊇ f (A) ⊇ f (B) ⊇ f 2 (A) ⊇ f 2 (B) ⊇ f 3 (A) ⊇ f 3 (B) ⊇ f 4 (A) ⊇ . . .

Let
S = S0 ∪ S1 ∪ S2 ∪ . . .

where for every n ∈ N we have that

Sn = f n (A) \ f n (B).

By Lemma 1 for every n ∈ N we have that f (Sn ) = Sn+1 . So it follows that

for every n ∈ N the restriction of f to Sn gives a bijection from Sn to Sn+1 .
Hence, we see that the restriction of f to S gives a bijection from S to S \ S0 .
Define g : A → B as
(
f (x) if x ∈ S
g(x) =
x if x ∈
/S

Let us check that g is a bijection.

As A is a disjoint union of S and A \ S and the corresponding restrictions
of g give bijections from S to S and from A \ S to A \ S, it follows that g is
injective.
Let S 0 = S \ S0 . As B is a disjoint union of S 0 and B \ S 0 and the corre-
sponding restrictions of g give onto maps from S to S 0 and from B \ S 0 to B \ S 0 ,
it follows that g is onto.

Corollary 4. Let A ≤ B and B ≤ A. Then we have that A ' B.

Proof. Let h : A → B and g : B → A be injective functions and f = g ◦ h. As

h(A) ⊆ B, we see that f (A) = g(h(A)) ⊆ g(B). Thus, we have that f : A → A
is injective and f (A) ⊆ g(B) ⊆ A. So by Theorem 7 it follows that A ' g(B).
As g is injective we see that g(B) ' B. Hence, we get that A ' B.

We have seen that Q is countable and so Q ≤ N. As N ⊆ Q we have that

N ≤ Q. So by Corollary 4 we get that Q ' N.

Proposition 6. R ' 2N .

6
Proof. Define ϕ : R → P(Q) as ϕ(x) = {a ∈ Q : a < x}. Then we see that ϕ is
injective and so R ≤ P(Q). As Q ' N, we have that P(Q) ' P(N). Also, we
have that P(N) ' 2N . Thus, it follows that R ≤ 2N .
Define ψ : 2N → R as follows:
For every f : N → {0, 1}, we let ψ(f ) be the real number whose decimal ex-
pansion is f (0).f (1)f (2)f (3) . . . Then ψ is injective and so 2N ≤ R. Hence, by
Corollary 4 we see that R ' 2N .

Observation 1. As 2N is not countable, we see that R is not countable.

Show that if A ' B then P(A) ' P(B).

Show that if A ' B then for every set C we have that A × C ' B × C,
AC ' B C and C A ' C B .
Show that if A ≤ B then for every set C we have that A × C ≤ B × C,
AC ≤ B C and C A ≤ C B .
Show that for every A, B, C we have that (AB )C ' AB×C .

Observation 2. Suppose we have that A ≤ B1 ≤ B2 ≤ . . . ≤ Bk−1 ≤ Bk ≤ A.

Then for every i ∈ [1, k] we have that A ≤ Bi and Bi ≤ A. So by Corollary 4
we get that A ' Bi for every i ∈ [1, k].

We have that R ' 1 × R ≤ R × R ' R2 ' (2N )2 ' 2N×2 ≤ 2N ' R and so by
Observation 2 we get that R2 ' R.

Observation 3. We have that 2N ≤ NN ≤ RN ' (2N )N ' 2N×N ≤ 2N as

N × N ≤ N. So by Observation 2 we get that 2N ' RN . Thus, we see that the
set of all binary sequences is in bijection with the set of all sequences of real
numbers. By Proposition 6 we see that both of these sets are isomorphic to R.

As 2 < R and 2N ' RN , we see that if A < B we cannot say that AC < B C .
Show that 2R ' RR .

Theorem 8. The set A of all continuous functions from R to R is in bijection

with R.

Proof. Let ϕ : A → RQ be defined as follows: ϕ(f ) is the restriction of f to Q.

Suppose ϕ(f ) = ϕ(g). Let h − f − g. Then h is a continuous function whose
restriction to Q is zero. As Q is dense in R it follows that h = 0 and so f = g.
Thus, we see that ϕ is injective and so A ≤ RQ . As we have that RQ ' RN ' R,
we get that A ≤ R. Let ψ : R → A be defined as follows: ψ(c) is the constant
function c. Then ψ is injective and so R ≤ A. Hence, by Corollary 4 it follows
that A ' R.

7
Remark 2. Does there exist a set A such that N < A < R? If we assume this
statement, it can be shown that we will not get any contradiction. This was
proved by Paul Cohen in 1963. If we assume that there does not exist such a set
A, it can again be shown that we will not get any contradiction. This was shown
by Kurt Godel in 1940. So it follows that we can neither prove this statement
nor can we prove its negation. Such a statement is said to be an independent
statement.