5 The Beginning of Transcendental Numbers

We have dened a transcendental number (see Denition 3 of the Introduction), but so far we have
only established that certain numbers are irrational. We now turn to the beginnings of transcen-
dental numbers. Our rst theorem is
Theorem 11. Transcendental numbers exist.
Like many of our results so far, this will of course be a consequence of later results. The rst
proof that there exist transcendental numbers was given by Liouville. Before we give his proof,
we give a proof due to Cantor.
Proof 1. The essence of this proof is that the real algebraic numbers are countable whereas the set
of all real numbers is uncountable, so there must exist real transcendental numbers. Dene
P(n) =
_
f(x) =
n

j=0
a
j
x
j
Z[x] : 1
n

j=0
|a
j
| n
_
.
Observe that P(n) is nite. Also, every non-zero polynomial in Z[x] belongs to some P(n). By
considering the real roots of polynomials in P(1), P(2), . . . (at the kth stage, consider the real roots
of polynomials in P(k) which have not occurred as a root of a polynomial in P(j) for j < k), we
can order the algebraic numbers; hence, they are countable. Next, give the usual proof that the real
numbers are uncountable.
For Liouvilles proof, we dene
Denition 4. A real number is a Liouville number if for every positive integer n, there are
integers a and b with b > 1 such that
0 <

a
b

<
1
b
n
.
In a moment, we will show that Liouville numbers exist. The second proof of Theorem 11 will
then follow from our next result.
Theorem 12. All Liouville numbers are transcendental.
Lemma 1. Let be an irrational number which is a root of f(x) =

n
j=0
a
j
x
j
Z[x] with
f(x) 0. Then there is a constant A = A() > 0 such that if a and b are integers with b > 0,
then

a
b

>
A
b
n
. (6)
Proof. Let M be the maximum value of |f

(x)| on [1, +1]. Let

a
b

A
b
n
A < min{1, |
1
|, |
2
|, . . . , |
m
|}.
14
Hence,
a
b
[ 1, + 1] and
a
b
{
1
, . . . ,
m
}.
By the Mean Value Theorem, there is an x
0
between a/b and such that
f() f(a/b) = ( a/b)f

(x
0
)
so that

a
b

f(a/b) f()
f

(x
0
)

=
|f(a/b)|
|f

(x
0
)|
.
Since f(a/b) = 0, we deduce that
|f(a/b)| =

j=0
a
j
a
j
b
nj

/b
n
1/b
n
.
Thus, since |f

(x
0
)| M, we obtain

a
b

1
Mb
n
>
A
b
n

a
b

,
giving a contradiction. Thus, the lemma follows.
Proof of Theorem 12. Let be a Liouville number. First, we show that must be irrational.
Assume = c/d for some integers c and d with d > 0. Let n be a positive integer with 2
n1
> d.
Then for any integers a and b with b > 1 and a/b = c/d, we have that

a
b

c
d

a
b

1
bd
>
1
2
n1
b

1
b
n
.
It follows from the denition of a Liouville number that is not a Liouville number, giving a
contradiction. Thus, is irrational.
Now, assume is an irrational algebraic number. By the lemma, there exist a real number
A > 0 and a positive integer n such that (6) holds for all integers a and b with b > 0. Let r be a
positive integer for which 2
r
1/A. Since is a Liouville number, there are integers a and b with
b > 1 such that

a
b

<
1
b
n+r

1
2
r
b
n

A
b
n
.
This contradicts (6) and, hence, establishes that is transcendental.
Example: We show that =

j=0
1/2
j!
is a Liouville number. First, observe that the binary
expansion of has arbitrarily long strings of 0s and so it cannot be rational. Fix a positive
integer n and consider
a
b
=

n
j=0
1
2
j!
with a and b = 2
n!
> 1 integers. Then
0 <

a
b

j=n+1
1
2
j!
<

j=(n+1)!
1
2
j
=
1
2
(n+1)!1

1
2
n(n!)
=
1
b
n
.
This proves that is Liouville and also gives a second proof of Theorem 11.
15
There are stronger versions of Theorem 12. In particular, we note (without proof) that
Theorem 13 (Thu e-Siegel-Roth). Let be an algebraic number with Q. Let > 0. Then
there are at most nitely many pairs of integers (a, b) with b > 0 such that

a
b

<
1
b
2+
.
It is not known whether or not the right-hand side above can be replaced by A/b
2
where A =
A() is a positive constant depending only on .
Theorem 14. The set of Liouville numbers in [0, 1] has measure 0.
Proof. Let > 0. It sufces to show that the (Lebesgue) measure of the Liouville numbers in [0, 1]
is < . Let n be a positive integer for which

b=2
4/b
n1
< (observe that this is possible since

b=2
4/b
n1
< (4/2
n3
)

b=2
1/b
2
). If is a Liouville number in [0, 1], then there are integers a
and b with b > 1 such that

a
b

<
1
b
n
.
Since the right-hand side is 1/2 and [0, 1], we deduce that a/b (1/2, 3/2) so that
b/2 < a < 3b/2 (a is in an open interval of length 2b). In particular, for a given integer b > 1,
there are 2b possible values of a for which the above inequality can hold. For each b > 1, we get
that must be in one of 2b intervals of length 2/b
n
. Thus, the measure of the Liouville numbers
must be

b=2
4b
b
n
< ,
giving the desired result.
Corollary 1. There are transcendental numbers which are not Liouville numbers.
We now turn to a related discussion. Suppose that we have an sequence of positive integers
{a
k
}

k=1
with a
1
< a
2
< . When will the argument given in the example above lead to a proof
that

k=0
1/2
a
k
is transcendental? It is not too difcult to see that what one essentially wants is
that a
k
increases fast enough where fast enough means that
lim inf
k
a
k+1
a
k
= .
On the other hand, it can be shown that sequences which increase much slower also lead to tran-
scendental numbers, an observation apparently rst noticed by Erd os. The next theorem illustrates
the basic idea by showing a certain number of this form is transcendental; it can be shown also that
this number is not a Liouville number.
Theorem 15. The number

k=0
1/2
2
k
is transcendental.
To prove Theorem 15, for positive integers k and m, we dene c(k, m) to be the number of
mtuples (j
1
, . . . , j
m
) of non-negative integers for which
k = 2
j
1
+ 2
j
2
+ + 2
j
m
.
16
Lemma 1. With the notation above, c(k, m) m
2m
.
Proof. We do induction on m. For m = 1, c(k, m) {0, 1} for all k, so the result is clear. Suppose
m > 1. If k has more than m non-zero binary digits, then it is not too hard to see that c(k, m) = 0
(note that the sum of two like powers of 2 is a power of 2). If k has exactly m non-zero binary
digits, then c(k, m) = m! m
m
m
2m
(j
1
can correspond to any one of the m non-zero digits,
j
2
to any one of the remaining m 1 non-zero digits, and so on). If k has fewer than m non-zero
binary digits, then for some integers r and s with 1 r < s m, j
r
= j
s
. Since in this case,
2
j
r
+ 2
j
s
= 2
j
r
+1
, we deduce that
c(k, m)
_
m
2
_
c(k, m1) m
2
(m1)
2m2
m
2m
,
establishing the lemma.
Lemma 2. Let t and m be positive integers. Then c(k, m) = 0 for every integer k (2
t+1
+ +
2
t+m
, 2
t+m+1
).
Proof. This follows since each such k has > m non-zero binary digits (see the proof of Lemma
1).
Proof of Theorem 15. Let =

k=0
1/2
2
k
. Let m be a positive integer. Then the denition of
c(k, m) implies that

m
=
_

k=0
1
2
2
k
_
m
=

k=1
c(k, m)
2
k
.
Let t be a positive integer. Using Lemma 2 and then Lemma 1, we obtain
_
2
2
t+1
++2
t+m

m
_
2
2
t+1
++2
t+m

k=2
t+1
++2
t+m
+1
c(k, m)
2
k
2
2
t+1
++2
t+m

k=2
t+m+1
c(k, m)
2
k
2
2
t+1
++2
t+m
m
2m

k=2
t+m+1
1
2
k
2
2
t+1
++2
t+m
m
2m
2
2
t+m+1
+1
= 2
12
t+1
m
2m
.
If we view m as being xed and let t approach innity, we see that the binary expansion of
m
has
arbitrarily long strings of 0s. By the same reasoning, we see that more generally if b
0
, b
1
, . . . , b
m
are non-negative integers, then the binary expansion of b
0
+b
1
+ +b
m

m
has arbitrarily long
strings of 0s. In fact, if b
0
, b
1
, . . . , b
m
and c
0
, c
1
, . . . , c
m
are non-negative integers and if we have
the following binary expansions
{b
0
+ b
1
+ + b
m

m
} = (0.d
1
d
2
d
3
. . . )
2
and
{c
0
+ c
1
+ + c
m

m
} = (0.d

1
d

2
d

3
. . . )
2
,
17
then for any positive integer N, there is a positive integer j for which
d
j+1
= d
j+2
= = d
j+N
= d

j+1
= d

j+2
= = d

j+N
= 0.
Furthermore, we can take N and j so that
N = 2
t
and j = 2
t+1
+ + 2
t+m
with t an integer as large (but perhaps not as small) as we wish.
Now, assume is a root of f(x) =

n
j=0
a
j
x
j
Z[x] with a
n
> 0 (which would be possible
if were algebraic). Then we can write f(x) in the form
f(x) =
n

j=0
b
j
x
j

j=0
c
j
x
j
,
where the b
j
and c
j
are non-negative integers with b
j
c
j
= 0 for each j. In particular, b
n
= a
n
> 0
and c
n
= 0. Take m = n1, and let N = 2
t
where t is a positive integer to be chosen momentarily
and j = j(t) is as above. Then
2
j+2
t2
n1

i=0
b
i

i
and 2
j+2
t2
n1

i=0
c
i

i
both differ from an integer by 1/2
N2
t2
= 1/2
2
t1
+2
t2
. If we write
2
j+2
t2
n1

i=0
b
i

i
= m
1
+
1
and 2
j+2
t2
n1

i=0
c
i

i
= m
2
+
2
,
where m
1
and m
2
are the greatest integers in the above expressions, then we see that since f() =
0,
2
j+2
t2
b
n

n
= m
3
+
3
with m
3
Z and |
3
| = |
1

2
|
1
2
2
t1
+2
t2
. (7)
On the other hand, since j has the form given above, we get from the denition of c(k, m) that
c(j + 2
t1
, n) 1 and c(k, n) = 0 for all k (j + 2
t2
, j + 2
t1
). For t sufciently large, we have
that
2
j+2
t2
b
n

k=j+2
t2
+1
c(k, n)
2
k
= 2
j+2
t2
b
n

k=j+2
t1
c(k, n)
2
k
2
j+2
t2
b
n
n
2n
2
j2
t1
+1
= 2
12
t2
b
n
n
2n
<
1
2
2
t3
and
2
j+2
t2
b
n

k=j+2
t2
+1
c(k, n)
2
k
2
j+2
t2
b
n
2
j2
t1
= 2
2
t2
b
n

1
2
2
t2
.
It follows that (7) cannot hold, giving a contradiction and completing the proof.
Homework:
1. Prove that the set of Liouville numbers in [0, 1] is uncountable.
18