The cardinality of the set of real numbers

Jailton C. Ferreira

Abstract
A proof that the set of real numbers is denumerable is given.
arXiv:math/0108119v7 [math.GM] 15 Jul 2013

1 Introduction
Cantor’s proof that the real set is uncountable is now included in the standard mathematics
curriculum. The objections and reservations presented by Poincaré, Kronecker, Weil and Brouwer,
for example, are silenced. Apparently we are in the Paradise that Cantor has created. This paper
seeks to show that this is not the case.
Each real number of the interval [0, 1] can be represented by an infinite path in a given binary
tree. In Section 2 the binary tree is projected on a grid N × N and it is shown that the set of
the infinite paths corresponds one-to-one to the set N . The Theorems 2.1 and 2.2 give the first
proof and the Theorem 2.3 provides a second proof.
Section 3 contains the 1891 Cantor’s proof and section 4 examines the the inconsistency of
this proof.

2 The proof of |F |= |N |
Theorem 2.1 Let B be a binary tree such that every node has two children and its depth is equal
to |N |. Let A be the set of the nodes of the binary tree B. The cardinality of A is less than or equal
to |N |.

Proof:
Let us now consider the Figure 1. The horizontal sequence of finite natural numbers, presented
in increasing order from left to right, contains all numbers of N ; to each natural number of the
sequence corresponds a vertical line. The vertical sequence of numbers, presented in increasing
order from top to bottom, contains all numbers of N ; to each natural number of the sequence
corresponds a horizontal line. To each node of the grid formed by the horizontal and vertical lines
corresponds a pair (m, n), where m belongs to the horizontal sequence of numbers and n belongs
to the vertical sequence of numbers.

1
 0 1 2 3 4 5 6 7 8 9

0
1
2
3
4
5
6
7
8
9

Figure 1

Let us project the binary tree B on the grid of the Figure 1 in the following way: to the root of
the tree corresponds the pair (0, 0); to the 2 nodes of the level 1 correspond the pairs (0, 1) and
(1, 0); to the 4 nodes of the level 2 correspond the pairs (0, 3), (1, 2), (2, 1) and (3, 0); to the 8
nodes of the level 3 correspond the pairs (0, 7), (1, 6), (2,5), (3, 4), (4, 3), (5, 2), (6, 1) and (7, 0)
and to the 2k nodes of the level k correspond the pairs

(0, 2k − 1), (1, 2k − 2), (2, 2k − 3), . . . , (2k − 3, 2), (2k − 2, 1) and (2k − 1, 0) (1)

The Figure 2 shows the binary tree B up to the depth 4. The Figure 3 shows the projection of the
binary tree up to the depth 3 on the grid of the Figure 1.

❤✭✭
✘✭
✘
❳ ✘✘ 0 ❤❤
✟ ❳❳ ✭
✟✟ 0 ❤✭
❳✭ ❤❤
✟
❍ 1
✭
0 ❍
❍ ❤✭
✘✘✭ ❤❤
✘❳
❍❳ ✘ 0
❳❳✭ ✭
1 ❤✭❤❤
1
❅ ✭ ✭✭
. ✘✘❤ ❤❤
❅ ✘❳✘
✟❳ 0
❳❳✭ ✭
❅ ✟✟ 0 ❤✭❤❤
❅❍
✟ 1
✭
1 ❍
❍ ✘✘❤✭
✭ ❤❤
✘❳
❍❳ ✘ 0
❳❳✭ ✭
1 ❤✭❤❤
1

1 2 3 4

Figure 2

2
 0 1 2 3 4 5 6 7 8 9

0 PP
❅ PP
1 ❅PP P
❅ ❅ PPP
2 ❅ ❇ ❅
3 ❇❅ ❅
❇ ❇ ❅ ❅
4 ❇ ❇
5 ❇ ❇
❇❇
6
7
8
9

Figure 3

It is known that the bijection f : N → N × N , where N is the set of finite natural numbers,
can be defined applying the diagonal method.

0 1 2 3 4 5 6 7 8 9

0
1
2
✒
3
4
5
6
7
8
9

Figure 4

The Figure 4 shows how starting from the pair (0, 0) and following the thick line we can establish
the one-to-one correspondence between the set of the pairs (m, n) and N , this is,

(0,0) (1,0) (0,1) (0,2) (1,1) (2,0) (3,0) ...

l l l l l l l (2)
0 1 2 3 4 5 6 ...

Considering that to any node of the binary tree corresponds a node of the grid and that there are
nodes in the grid without corresponding nodes in the binary tree, we conclude that

|A|≤ |N | (3)

Theorem 2.2 The cardinality of the set of real numbers of the interval [0, 1] is equal to |N |.

Proof:
Let us denote by F the set of real numbers of the interval [0, 1]. Any element of F can be
represented in the binary system by

3
 f1 × 2−1 + f2 × 2−2 + f3 × 2−3 + f4 × 2−4 + . . . (4)

where
fi ∈ {0, 1} and i∈N and i 6= 0 (5)

The representation (4) can be simplified to

.f1 f2 f3 f4 f5 . . . (6)

where the first character of the sequence is the point. Each fi of (6) is substituted by 0 or 1 to
represent a given number. The i-th 0 or 1 on the right of “.” corresponds to fi . The set F can
also be represented by a binary tree where each node has two children. Each infinite path on the
binary tree, .qτ 1 qτ 2 . . . qτ i . . . with depth equal to |N | , represents an element of F .
Let us denote by B the binary tree above mentioned; by Bi the binary tree of depth i; by P
the set of all paths starting from the root of B; by Pi the set of all paths from the root of Bi to
its leaves. Each element of Pi or P is the set of the nodes (except the root) that form the path.
Let the set

{P1 , P2 , . . . , Pi , . . .} such that i∈N (7)

The set {P1 , P2 , . . . , Pi } contains all paths from the root of the tree Bi and the endings of the
paths that exist in {P1 , P2 , . . . , Pi } are all the nodes of Bi , except the root. This indicates that
the set in (7) - which corresponds to all the tree B - contains all paths of B.
Let us consider the set of all infinite paths from the root of B

Q = {qα , qβ , qγ , . . . , qτ , . . .} (8)

The infinite path qτ can be represented by the set

{qτ 1 , qτ 2 , . . . , qτ i , . . .} (9)

where qτ i is the i-th node of qτ after the root. Let

Qτ = {{qτ 1}, {qτ 1 , qτ 2 }, . . . , {qτ 1 , qτ 2 , . . . , qτ i }, . . .} (10)

The union of all elements of Qτ is equal to qτ , that is,

qτ = {qτ 1 } ∪ {qτ 1 , qτ 2 } ∪ . . . ∪ {qτ 1 , qτ 2 , . . . , qτ i } ∪ . . . (11)

Considering (11) we can obtain from (8)

Q = {{qα1 } ∪ {qα1 , qα2 } ∪ . . . ∪ {qα1 , qα2 , . . . , qαi } ∪ . . . ,

{qβ1 } ∪ {qβ1 , qβ2 } ∪ . . . ∪ {qβ1 , qβ2 , . . . , qβi } ∪ . . . ,
........................................................
{qτ 1 } ∪ {qτ 1 , qτ 2 } ∪ . . . ∪ {qτ 1 , qτ 2 , . . . , qτ i } ∪ . . . ,
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .}
(12)

From (12) we have

4
 |Q|≤ |{qα1 } ∪ {qα1 , qα2 } ∪ . . . ∪ {qα1 , qα2 , . . . , qαi } ∪ . . .
∪ {qβ1 } ∪ {qβ1 , qβ2 } ∪ . . . ∪ {qβ1 , qβ2 , . . . , qβi } ∪ . . .
.........................................................
∪ {qτ 1 } ∪ {qτ 1 , qτ 2 } ∪ . . . ∪ {qτ 1 , qτ 2 , . . . , qτ i } ∪ . . .
.........................................................|
(13)

Since the sets on the right side of (13) with cardinality equal to i are the elements of Pi , that is,

Pi = {qα1 , qα2 , . . . , qαi } ∪ {qβ1 , qβ2 , . . . , qβi } ∪ {qτ 1 , qτ 2 , . . . , qτ i } ∪ . . . (14)

we conclude that

|Q|≤ |P1 | + |P2 | + |P3 | + . . . + |Pi | + . . . (15)

The set of the nodes of B has cardinality less than or equal to |N | (Theorem 2.1). Since the
number of nodes in the level i of the tree Bi is equal to |Pi |, we have

|N |≥ |P1 | + |P2 | + |P3 | + . . . + |Pi | + . . . (16)

Therefore

|Q|≤ |N | (17)

Considering the bijection f : F → Q and |F |≥ |N |, we conclude

|F |= |N | (18)

Theorem 2.3 The cardinal number of the set of infinite paths of the binary tree B is equal to |N |.

Proof:
Be the infinite path of Q

qτ 0 , qτ 1 , qτ 2 , . . . , qτ i , . . . (19)

where qτ i = 1 for all i ∈ N with i > 0, projected in the grid as shows the Figure 3. In the
coordinates of the grid, the path (19) corresponds to the pairs

(0, 0), (0, 1), (0, 3), (0, 7), (0, 15), . . . , (0, 2k − 1), . . . (20)

Let us notice that in (0, 0) all infinite paths begin. By the pair (0, 1) a part of all infinite paths
passes, by (0, 3) a part of the infinite paths that passed by (0, 1) passes, by (0, 7) a part of the
infinite paths that passed by (0, 3) passes and so forth. When an entire path is accomplished, the
path exists. In the examined case, (19) is the accomplished path.
Each pair (0, 2k − 1) of the sequence (20) belongs to the set of pairs given by (21)

(0, 2k − 1), (1, 2k − 2), (2, 2k − 3), . . . , (2k − 3, 2), (2k − 2, 1) and (2k − 1, 0) (21)

Let us denote by Gk the set of pairs given by (21) for k. For any k the cardinal number of the
set of infinite paths that passes by (0, 2k − 1) is equal to the cardinal number of the set of infinite
paths that pass by any other pair of (21). For any k the distance from (0, 0) to (0, 2k − 1) in the
grid is equal to |Gk |. When we examine the pairs of (20), from left to right, and we accomplished

5
the path (19), Gk becomes the set of the infinite path whose cardinality is equal to the cardinality
of the set of the nodes of the path (19).
The cardinality of the set of the infinite paths is |F | and the cardinality of the set of nodes of
any infinite path is |N |. Therefore

|F |= |N | (22)

3 The 1891 proof

Theorem. Let F be the set of real numbers of the interval [0, 1]. The set F is not
countable.
Proof. We assume that the set F is countable. This means, by the definition of
countable sets, that F is finite or denumerable. Let us notice that |F | is not less than
|N | where N the set of natural numbers. Be

F = {a1 , a2 , a3 , . . .}

where the cardinality of {a0 , a1 , a2 , . . .} is |N |. We can write their decimal expansions

as follows:

a1 = 0. d1,1 d1,2 d1,3 ...

a2 = 0. d2,1 d2,2 d2,3 ...
(23)
a3 = 0. d3,1 d3,2 d3,3 ...
... ...

where the d’s are binary characters 0 and 1. Now we define the number

x = 0.d1 d2 d3 . . .

by selecting d1 6= d1,1 , d2 6= d2,2 , d3 6= d3,3 , . . . . This gives a number not in the set
{a1 , a2 , a3 , . . .}, but x ∈ F . Therefore, F is not countable [1].

4 The inconsistency of the proof

When (23) is represented as (24) bellow, the proof can be called the written list form of the
Cantor’s argument of 1891.

1 ←→ 0. d1,1 d1,2 d1,3 ...

2 ←→ 0. d2,1 d2,2 d2,3 ...
(24)
3 ←→ 0. d3,1 d3,2 d3,3 ...
... ...

Let be the list

6
 n∈N ←→ x∈F

0 ←→ 0.000000 ...
1 ←→ 0.100000 ...
2 ←→ 0.010000 ...
3 ←→ 0.110000 ...
4 ←→ 0.001000 ...
5 ←→ 0.101000 ...
6 ←→ 0.011000 ...
7 ←→ 0.111000 ...
(25)
8 ←→ 0.000100 ...
9 ←→ 0.100100 ...
10 ←→ 0.010100 ...
11 ←→ 0.110100 ...
12 ←→ 0.001100 ...
13 ←→ 0.101100 ...
14 ←→ 0.011100 ...
15 ←→ 0.111100 ...
16 ←→ 0.000010 ...
... ...

Let us notice that (i) d1,1 , d3,1 , d5,1 , d7,1 , d9,1 , d11,1 , d13,1 , d15,1 , d17,1 ... are equal to 0 and d2,1 ,
d4,1 , d6,1 , d8,1 , d10,1 , d12,1 , d14,1 , d16,1 , d18,1 ... are equal to 1, (ii) d1,2 , d2,2 , d5,2 , d6,2 , d9,2 , d10,2 ,
d13,2 , d14,2 , d17,2 ... are equal to 0 and d3,2 , d4,2 , d7,2 , d8,2 , d11,2 , d12,2 , d15,2 , d16,2 , d19,2 ... are
equal to 1, (iii) d1,3 , d2,3 , d3,3 , d4,3 , d9,3 , d10,3 , d11,3 , d12,3 , d17,3 ... are equal to 0 and d5,3 , d6,3 ,
d7,3 , d8,3 , d13,3 , d14,3 , d15,3 , d16,3 , ... are equal to 1, (iv) ... and so on ...

d1,1 d1,2 d1,3 ...

d2,1 d2,2 d2,3 ...
(26)
d3,1 d3,2 d3,3 ...
...

More generally, considering the matrix (26) the procedure to obtain (25) is: the first column of
the matrix is filled from top to bottom with a succession of the pattern 01. The second column is
filled with a succession of the pattern 0011. The n-th column is filled from top to bottom with a
succession of the pattern constituted of 2n 0’s followed by 2n 1’s.
When the number of rows in (25) goes to infinity we have that a x equal to

limn→∞ (2−1 + 2−2 + 2−3 + ... + 2−n ) = 0.11111111... (27)

belongs to (25). There is no way out of this. All the reals in the interval [0,1] are included in
the list. However applying the Cantor’s argument to (25) we found that 0.11111111... does not
belongs to (25).

References
[1] G. Cantor, Über eine elementare Frage der Mannigfaltigkeitslehre, Jahresbericht der Deutschen
Mathematiker-Vereinigung, 1 (1891), 75-78.