Countable and Uncountable Sets

Rich Schwartz
November 12, 2007

The purpose of this handout is to explain the notions of countable and

uncountable sets.

1 Basic Definitions
A map f between sets S1 and S2 is called a bijection if f is one-to-one and
onto. In other words

• If f (a) = f (b) then a = b. This holds for all a, b ∈S1.

• For each b ∈S2, there is some a in S1 such that f (a) = b.

We write S1 ∼ S2 if there is a bijection f : S1 → S2. We say that S1 and S2 are
equivalent or have the same cardinality if S1 ∼ S 2. This notion of equivalence
has several basicproperties:

1. S ∼ S for any set S. The identity map serves as a bijection from S to

itself.

2. If S1 ∼ S2 then S2 ∼ S1. If f : S1 → S2 is a bijection then the inverse

map f −1 is a bijection from S2 to S 1.

3. If S1 ∼ S2 and S2 ∼ S3 then S1 ∼ S3. This boils down to the fact that the
composition of two bijections is also a bijection.
These three properties make ∼ into an equivalence relation.
Let N = {1, 2, 3...} denote the natural numbers. A set S is called count-
able is S ∼ T for some T ⊂ N . Here is a basic result about countable sets.

1
Lem m a 1.1 If S is both countable and infinite, then there is a bijection
between S and N itself.

Proof: For any s ∈S, we let f (s) denote the value of k such that s is the kth
smallest element of S. This map is well defined for any s, because there are
only finitely many natural numbers between 1 and s. It is impossible for two
different elements of S to both be the kth smallest element of S. Hence f is
one-to-one. Also, since S is infinite, f is onto. ♠

Lem m a 1.2 If S is countable and S ′ ⊂ S, then S is also countable.

Proof: Since S is countable, there is a bijection f : S → N . But then

f (S ′ ) = N ′ is a subset of N , and f is a bijection between S ′ and N ′ . ♠

A set is called uncountable if it is not countable. One of the things I will do

below is show the existence of uncountable sets.

Lem m a 1.3 If S ′ ⊂ S and S ′ is uncountable, then so is S.

Proof: This is an immediate consequence of the previous result. If S is

countable,then so is S ′ . But S ′ is uncountable. So, S is uncountable as well.
♠

2 Examples of C o u n t a b l e Sets
Finite sets are countable sets. In this section, I’ll concentrate on examples of
countably infinitesets.

2.1 T h e I n t e ge rs
The integers Z form a countable set. A bijection from Z to N is given by f
(k) + 2k if k ≥ 0 and f (k) = 2(−k) + 1 if k < 0. So, f maps 0, 1, 2, 3... to 0,
2, 4, 6... and f maps −1, −2, −3, −4... to 1, 3, 5, 7....

2
2. Th e Rational Numbers
I’ll give a different argument than the one I gave in class. Let Lq denote the
finite list of all rational numbers between −q and q that have denominator at
most q. There are at most q(2q + 1) elements of Lq . We can make the list L1,
L 2, L 3 , ... and throw out repeaters. This makes a list of all the rational
numbers. As above, we define f (p/q) to be the value of k such that p/q is the
kth fraction on our list.

3. T h e Algebraic N u m b e r s
A real number x is called algebraic if x is the root of a polynomial equation
√
c0 + c1x + ... + c n x n where all the c’s are integers. For instance, 2 is an
algebraic integer because it is a root of the equation x2 − 2 = 0. To show that
the set of algebraic numbers is countable, let L k denote the set of algebraic
numbers that satisfy polynomials of the form c0 +c 1 x+...+c n x n where n < k
and max(|cj |) < k. Note that there are at most k k polynomials of this form,
and each one has at most k roots. Hence L k is a finite set having at most kk+1
elements. As above, we make our list L 1, L 2 , L3 of all algebraic numbers and
weed out repeaters.

2.4 C o u n t a b l e Un i o ns of C o u n t a b l e Sets
Lemm a 2.1 Suppose that S1 , S2, ... ⊂ T are disjoint countable sets. Then
S
S = i S i is a countable set.

Proof: There are bijections f i : S i → N for each i. Let L k denote the set of
elements s ∈ S such that s lies in some S i for i < k, and f i (s) < k. Note that
L k is a finite set. It has at most k2 members. The list L 1 , L 2 , L3... contains
every element of S. Weeding out repeaters, as above, we see that we have listed
all the elements of S. Hence S is countable. ♠

The same result holds even if the sets S i are not disjoint. In the general
case, we would define
[ k−1
S k′ = S k− S ,i
i=1

and apply the above argument to the sets S ′ , S1′ ....2 The point is that S ′ is i
′ S S
countable, the various S sets are disjoint, and i S i = i S i . ′

3
3 Examples of U n c o u n t a b l e Sets
1. T h e S e t of B i n a r y Sequences
Let S denote the set of infinite binary sequences. Here is Cantor’s famous proof
that S is an uncountable set. Suppose that f : S → N is a bijection. We
form a new binary sequence A by declaring that the nth digit of A is the
opposite of the nth digit of f −1 (n). The idea here is that f −1 (n) is some
binary sequence and we can look at its nth digit and reverse it.
Supposedly, there is some N such that f (A) = N. But then the Nth digit
of A = f −1 (N) is the opposite of the Nth digit of A, and this is a
contradiction.

2. T h e Real Numbers
Let R denote the reals. Let R ′ denote the set of real numbers, between 0and 1,
having decimal expansions that only involve 3s and 7s. (This set R ′ is an
example of what is called a Cantor set.) There is a bijection between R ′ and
the set S of infinite binary sequences. For instance, the sequence 0101001... is
′
mapped to .3737337.... Hence R is uncountable. But then Lemma 1.3 says
that R is uncountable as well.

3. T h e Tr a n s c e n d e n t a l N u m b e r s
A real number x is called transcendental if x is not an algebraic number. Let
A denote the set of algebraic numbers and let T denote the set of tran-
scendental numbers. Note that R = A ∪ T and A is countable. If T were
countable then R would be the union of two countable sets. Since R is un-
countable, R is not the union of two countable sets. Hence T is uncountable.
The upshot of this argument is that there are many more transcendental
numbers than algebraic numbers.

4. Tail E n d s of B i n a r y Sequences
Let T denote the set of binary sequences. We say that two binary sequences A1
and A2 are equivalent if they have the same tail end. For instance 1001111...
and 111111... are equivalent.

4
Lem m a 3.1 For any binary sequence A, there are only countably many bi-
nary sequences equivalent to A.

Proof: Let L n denote the set of sequences that differ from A only in the first n
digits. Then L n is a finite set with at most 2n elements. Now we list L1, L2,
L3.... This gives a list of all the binary sequences equivalent to A. The rest of
the proof is as above. ♠

Say that a tail end is the collection of sequences all equivalent to a given
one. Note that T is the union of tail ends. Each tail end it a countable set, and
T is uncountable. Hence, there are uncountably many tail ends, by Lemma 2.1.

3.5 T h e P e n r o s e Tiles
To each Penrose tiling P we can associate a tail end τ (P ). Recall that there is
an infinite sequence P = P0, P1, P2, ... where Pn is the parent of Pn−1. In other
words, P n is obtained from Pn−1 by the grouping process discussed in class.
We say that the nth digit of τ (P ) is a 0 if x is contained in a kite of Pn
and a 1 if x is contained in a dart of Pn . We might need to move x slightly
to avoid choosing a point that lies right on a crack. If we replace x by x ′ , then
only the initial part of the sequence changes. So, τ (P ) is well defined.
By using the subdivision operation, we can produce a Penrose tiling P that
has any τ (P ) we like. Hence, there are uncountably many different Penrose
tilings.

4 A H eirarchy of Infinite Sets

For any set S let 2S denote the set of subsets ofS.

Lem m a 4.1 There is no bijection between S and 2S .

Proof: This is really a generalization of Cantor’s proof, given above. Sup-

pose that there really is a bijection f : S → 2S . We create a new set A as
follows. We say that A contains the element s ∈ S if and only if s is not a
member of f (s). This makes sense, because f (s) is a subset of S.

5
 Since A is a subset of S, we have A = f (a) for some a ∈S. If a ∈A then a ∈
f (a). But then, by definition, a is not a member of A. On the other hand, if a
ƒ∈A, then a ∈f (a). But, again, this is a contradiction. The only way out of the
contradiction is to realize that there can be no bijection f . ♠

We can start with S0 = N , and recursively define S n = 2 S n−1 . That is, S n

is the set of subsets of S n−1 . Then, the sets S0, S 1, S2, ... form an infinite
heirarchy of sets, each one so much larger than the previous one that there is no
bijection between it and the previous one.
The fun doesn’t stop there. We candefine
[∞
Σ0 = Sn.
n=0

Then, there is no bijection between Σ0 and S n for any n. The set Σ0 is larger
than all of the sets previously defined. One can now define Σ n = 2 Σn−1 . And so
on.