You are on page 1of 5

All proofs of 1=0.999...

are false
Abdeljalil Saghe

To cite this version:


Abdeljalil Saghe. All proofs of 1=0.999... are false. 2019. �hal-02085315�

HAL Id: hal-02085315


https://hal.science/hal-02085315
Preprint submitted on 30 Mar 2019

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est


archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
ALL PROOFS OF 1 = 0.9 ARE FALSE

SAGHE ABDELJALIL

Abstract. In this paper we propose to prove that 1 is different to 0.9.


To clarify our point of view, we present some examples and proofs to
see that all demonstrations of the other opinion are false.

Contents
1. Introduction 1
2. Why 0.9 is different to 1 1
3. The equality 10u = u + 9 is false 2
4. The second false proof 3
References 4

1. Introduction
In the sense to correct the demonstrations proposed by some mathemati-
cians to prove the equality 1 = 0.9. I propose in this article to give the
truth demonstration by using the properties of the infinitesimal numbers
presented by the Robinson’s theory (see [2]). In the last false proof we
correct it by using the Transfer principle(see [1]). Finally, we propose an
example to explain our opinion.

2. Why 0.9 is different to 1


Let x be a positive real number. The decimal representation of x is giving
by:
x0 , x1 x2 x3 · · · xn · · · ,
where x0 ∈ N , and 0 ≤ xi ≤ 9 for all integer i ≥ 1. The sequence (xi ) is
given by:

x0 = ⌊x⌋ ;
i i−1
xi = ⌊10 x⌋ − 10⌊10 x⌋ for all integer i ≥ 1.
Then, for the number u = 0.9, we have:

⌊u⌋ = 0 ;
(S) i i−1
⌊10 u⌋ − 10⌊10 u⌋ = 9 for all integer i ≥ 1.

2010 Mathematics Subject Classification. 26E35, 06A75 .


Key words and phrases. Nonstandard analysis, hyppereals, Internal Set Theory.
1
As 1 is not a solution of any equation of the above system, then u 6= 1.

3. The equality 10u = u + 9 is false


Consider the following demonstration:
Proof. we note
u = 0.999 · · · .
We have:
(3.1) u = 0.999 · · ·
(3.2) ⇒10u = 9.999 · · ·
(3.3) ⇒10u = 9 + u
by solving the equation, we find u = 1. 
This proof is false. In fact, let x = hxi i be the hyperreal defined by
the sequence (xi ). For all function f : R → R, the hyperreal f (x) is
given by f (x) = hf (xi )i. Let u be the hyperreal number defined by u =
h0.9, 0.99, 0.999, . . .i. For every integer n ≥ 1, we note:
gn (x) = ⌊10n x⌋ − 10⌊10n−1 x⌋.
We verify easily that ⌊0.9⌋ = ⌊0.99⌋ = ⌊0.999⌋ = · · · = 0. In addition, we
have gn (0.9) = gn (0.99) = gn (0.999) = · · · = 9 for all n ≥ 1 .
So, we obtain:

⌊u⌋ = h⌊0.9⌋, ⌊0.99⌋, ⌊0.999⌋, · · · i = 0 and ;
gn (u) = hgn (0.9), gn (0.99), gn (0.999), · · · i = 9 ∀n ≥ 1.
Then, the hyperreal u is a solution of the system (S). We note:
u = 0, 999 · · ·
Next, we can prove that the equality 10u = u + 9 is not true. In fact, we
have:
u = h0.9, 0.99, 0.999, . . .i.
Then
10u = h9, 9.9, 9.99, . . .i
In addition, we have:
u + 9 = h9.9, 9.99, 9.999, . . .i.
Then, we deduce that:
(3.4) u + 9 − 10u = h0.9, 0.09, 0.009, . . .i,
(3.5) = 9δ,
where δ = h0.1, 0.01, 0.001, . . .i is the infinitesimal number defined by the
sequence (10−n )n≥1 . As δ is a nonzero hyperreal number. Then, we deduce
that 10u 6= u + 9.
Another proof is given as follow.
For every integer n ≥ 1, we define: un = 1 − 10−n = 0.9 · · · 9.
We verify easily that:
(3.6) un + 9 − 10un = 0, 0 · · · 09,
(3.7) = 9δn ,
where δn = 10−n .
So, we find:
(3.8) u + 9 − 10u = hui + 9 − 10ui i,
(3.9) = h9δ1 , 9δ2 , · · · , 9δi , · · · i,
(3.10) = 9δ.
Remark 3.11. • The problem we may encounter if we use the notation
0.999 · · · , is the fact that the solution of the system (S) is not unique.
For example, Let u = hui i = h0.9, 0.99, 0.999, · · · i and v = hvi i =
h0.99, 0.999, 0.9999, · · · i two hyperreal numbers. We can verify easily
that u and v are solutions of (S). As ui 6= vi for all i, then, u 6= v
and u − v is an nonzero infinitesimal number.
• Since the hypereal δ = 1 − 0.999 · · · is infinitesimal. Then, 0.999 · · ·
is not a real number.
• We can find more difficulties to prove that the number 0.999 · · · is
an Omicran (see [3]).

4. The second false proof


Consider the following demonstration:
Proof. We have:
m
X
(4.1) 0.999 · · · = lim 9.10−n ,
m→+∞
n=1
+∞
X
(4.2) = 9. 10−n ,
n=1
9 1
(4.3) = . ,
10 1 − 10−1
(4.4) = 1.

The goal of this section, is to prove that this proof is not true. For that,
Let (sm ) be the series defined as:
m
X
sm = 9.10−n .
n=1
The sentence [lim∞ sm = 0.999 · · · ] is not true. The truth limit of this series
is 1, we can established it by using the following definition of limit:
∀εreal > 0, ∃ n0 ∀m ≥ n0 | sm − 1 |≤ ε.
By using the Transfer principle, we find:
∀ε > 0, ∃ n0 ∀m ≥ n0 | sm − 1 |≤ ε.
Then lim sm = 1. In addition, we can verify easily that we have:
∀εreal > 0, ∃ n0 ∀m ≥ n0 | sm − 0.999 · · · |≤ ε.
But the following proposition
∀ε > 0, ∃ n0 ∀m ≥ n0 | sm − 0.999 · · · |≤ ε,
is not true. Here we cannot apply the Transfer principle for the sentence
[∀εreal > 0, ∃ n0 ∀m ≥ n0 | sm − 0.999 · · · |≤ ε], because 0.999 · · · is
not a real number and the above proposition is not an internal formula.
Examples 4.5. Let (xn ) the sequence defined as xn = 1 + n1 . Let δ be an
infinitesimal number.
We can verify easily that:
∀εreal > 0, ∃ n0 ∀m ≥ n0 | xm − 1 |≤ ε,
and
∀εreal > 0, ∃ n0 ∀m ≥ n0 | xm − (1 + δ) |≤ ε.
Despite that, we have lim xn = 1 and lim xn 6= 1 + δ.

References
[1] E. Nelson, Internal Set Theory: a new approach to Nonstandard Analysis, Bull. A.M.S.,
83 (1977), pp. 1165-1198.
[2] A. Robinson, Non-standard Analysis, North-Holland, Amsterdam, 1966.
[3] A.Saghe, A New Approach to Nonstandard Analysis, Sahand Communications in Math-
ematical Analysis, 12.1 (2018), pp. 195-254.

Centre Bts Essaouira - Maroc.


E-mail address: saghe007@gmail.com

You might also like