1

Joseph Liouvilles construction of a transcendental number

Topmatter:
Some ideas pulled from exercises 6.6 6.8 in Galois Theory
(Second Edition) by Ian Stewart, page 69, some from
elsewhere.

Here, present a (probably modified) version of the work of the

French mathematician Liouville that explicitly made a
transcendental number, and some work that was done after his
of 1844. The number here produced is Liouvilles Constant.
The later work of the German mathematician Cantor made no
transcendental numbers, but had significant ideas of cardinality. My beatup text

Joseph Liouville
There are countless web references to transcendental numbers, a Wikipedia
page. There are biographies of Joseph Liouville, discussions of transcendental numbers, and even a page
on the fifteen most famous transcendental numbers. But could not find a good construction, though this
page has some discussion on a higher level than this one, as does this link.

Credit:
Slowly typeset by Dean Moore, October December 2010, Boulder, Colorado, USA, tweaks afterward.
Did some notable fixes in September 2016, improving explanations, fixing confusing math and mistakes,
plus bad notation, a couple of holes, general nonsense. Also did fixes, June 2017.

In March 2009, worked the proof for pure entertainment, later typeset to slowly piece through, to
understand the logic, and fill in gaps. Parts are my revenge against math texts, where the passage Thus it
clearly follows was found, and hours of work was required to fill in the abyss between clearly and
follows. Why in graduate school at CU Boulder a mathematician studying analysis bought a book on
Galois Theorywhich lands in the basket of abstract algebrais a hidden mystery, though years later
worked through it. Fascinating material, Galois Theory.

Here, aim at a simplistic explanation with many references (probably too many)
that could be understood by an undergrad in math with some understanding of
real analysis, number theory, rings and fields, a little set theory. You are
expected to have familiarity with elementary algebra as exponents. Some
understanding of limits of sequences and infinite sums is a must. Understanding
of some mathematical notation is assumed; one reference is here. A few
relevant terms:

means For all, Basic understanding of limits of sequences is

means There exists, a must, as is some understanding of series.
means Is an element of, and
 2

0, is a set notation, here means The set of all integers from 0 to , inclusively.
)* /
Pass to a subsequence means, of say, a sequence like to choose a new sequence
+* ,-.
/
)* / )*
0
by throwing out a subset of . This is written and note, 1 refers to a subset
+* ,-. +*
0 1-.
)* ) )* )
0
of all , now indexed by . Notably, when lim = , it is also true that lim = .
,/ +* + 1/ +* 0
+

Images tend to be Wikipedia images. If by some weirdness someone finds it useful, great. Send
me an email, dean at deanlm dot com, or if made some typo, flub or poor logic.
Legal information:
Creative Commons AttributionNonCommercialShareAlike 3.0 License
You are free:
to Share to copy, distribute, display, and perform the work
to Remix to make derivative works
Under the following conditions:
Attribution. You must attribute the work in the manner specified by the author or licensor.
Noncommercial. You may not use this work for commercial purposes.
Share Alike. If you alter, transform, or build upon this work, you may distribute the resulting work only under a license identical to this one.
For any reuse or distribution, you must make clear to others the license terms of this work.
Any of these conditions can be waived if you get permission from the copyright holder.

Background:
Rational numbers, the ancient Greeks:
Ancient Greek mathematicians first assumed all numbers rational, i.e., ratios of integers
)
as , with , , 0. Easy examples are
+
5 13 32 7 127 7853
> = , . = = 13, or E = = 3.14159265
7 1 28 57

The Greeks knew of the square root of two, 2, from obvious considerations from the
Pythagorean Theorem (diagram, right: Note 2 is the hypotenuse of the right triangle
with both short legs of length 1, as 1M + 1M = 2). Initially, they assumed 2
rational. Eventually, a Greek proved the square root of two was irrational, i.e.,
cannot be expressed as a fraction of two integers. This has been proved, in many
ways.

So for , , 0 to write

= 2

is impossible. As rational numbers are dense in the real line, one may use rational numbers to get
arbitrarily close (i.e., limit), but equality cannot happen.
 3

Notably for the polynomial = M 2, the numbers M = 2 and

E = 2 satisfy M = E = 0. So 2 are the roots of a
polynomial in rational (here integer) coefficients.

So we have an simple way to include numbers like 2 in our number

system and in terms of simple familiar things: roots of polynomials in
integer coefficients, ideas studied since at least the ancient Egyptians and
Babylonians.

Algebraic numbers:
Babylonian mathematics on .
Thus algebraic numbers: all numbersreal or complex, but this paper
stays in the world of real numbersthat are roots of some polynomial
T

= , , = > + . + M M + + TU. TU. + T T ,

,->

; , , 0,

in rational coefficients (equivalently, integer: since is set

equal zero one may multiply by a least common multiple to make
all coefficients integers). That these form an algebraically closed
field is proven elsewhere. This result is in various books; some
discussion here. As an example, a number as

..
E Z
3+ 7.Y
ZM
Y . YE
5M 2[ +1 1

is algebraic, a root of some integercoefficient polynomial. From

The French mathematician
variste Galois, at roughly the
age of 15. He had about five
years to live, but before dying
solved a problem that had been
floating around for hundreds of
years. Galois was brilliant,
changed mathematics forever,
but lived a sad, rather tragic life.
Galois Theory, roots of polynomials of degree five and higher
cannot, in general, be expressed as combinations of radicals of
integers or rational numbers, as in general, such polynomials are
not solvable. But Galois Theory gets a bit away from the subject
matter of this paper.
Mainly as its impossible, no one proved all numbers are
algebraic. After Liouvilles work of 1844, in 1874 Georg Cantor
proved the algebraic numbers a field of cardinality 0, actually
small in a settheoretic sense, the first infinity. The real
 4

numbersor the complexesare fields of cardinality 2_ . Note 2_ is

sometimes equated to 1, but this is set theory and not currently
decidable, and gets afield for this paper. Here 2_ means the second
infinity, which is strictly larger than > . But cardinality waited for
Cantor, whose work postdated Liouville.

The transcendental numbers:

Algebraic numbers led to the idea of the transcendental number: a
numberreal or complexthat is not the root of any polynomial with
rational (equivalent, integer) coefficients or algebraic coefficientsthe
algebraic numbers are algebraically closed, so algebraic coefficients is
the same as to integer coefficients. Algebraically closed fields also get a
bit afield, for the paper.
Georg Cantor, whose theorems
That is to say, a transcendental number equals any numberreal or are now at the foundation of
complexthat is not algebraic. It is easy to talk about such an idea, as a mathematics, but who in his life
saw his ideas experience
theoretical construct. ferocious opposition, notably
from Leopold Kronecker and
However, no one explicitly constructed one, though in 1677 the Henri Poincar
mathematician James Gregory attempted to prove transcendental (this
history here and here). The term transcendental goes back to at least Leibniz in 1673, if more
modern formulations appear to trace to Euler in 1748.

Genesis: Transcendental numbers were first proven to exist in 1844 by the French mathematician
Joseph Liouville, though he did not then construct an explicit decimal number but a continued
fraction. The first decimal proven transcendental was the Liouville constant which Liouville
proved transcendental in 1850, not 1844 as stated in some web references. It belongs to a class of
numbers, a Liouville number, is a bit odd, and never occurs in physics. The first naturally
occurring transcendental numbers were later proven to be e (Hermite, 1873) and (Lindemann,
1882), neither of which are Liouville numbers.

As an aside, when in 1882 the number was proven transcendental, it proved that by the
methods of ancient Greek geometers it is impossible to square the circle. This answered a
question had been open for thousands of years.

After Liouville, Cantor discovered no transcendental numbers but showed they had to exist, and
his work had significant implications of cardinality: Via his diagonal argument, Cantor showed
the real numbers comprise a big uncountable 2_ set. Cantor also showed the algebraic
numbers are a countable set, that is, are of the cardinality called countable infinity. As

the algebraic numbers are countable, and no transcendental number is algebraic, and
the real numbers are uncountable,
 5

it follows that transcendental numbers must comprise an uncountable set, that is, a big set of
cardinality 20 , not a small set of cardinality > . In terms of describing cardinality with
differentsized blobs, the following Venn diagram is inaccurate, as the measure of the algebraic
numbers equals zero, thus algebraic numbers should be an invisible dot. But it gives some idea:

The real numbers, including

The algebraic
numbers, of transcendental numbers, of
cardinality >
cardinality .

A foretaste of what is to come:

Let > be an irrational algebraic number, a root of a polynomial () [] of degree .
In this foretaste, we fix > . Liouville took , and bounds of the derivative of near > ,
and set a speed limit on how fast a sequence of rational numbers can limit on an
irrational algebraic number. We will set up this speed limit.

Using that speed limit, Liouville then showed only a finite set of rational numbers could
beat a stronger speed limit to > . We will set up this speed limit.

He defined a new irrational number we will define. He set up a sequence of rational

numbers limiting to this new irrational number. He then proved an infinite subset of the
sequence broke the strong speed limit, when he had proved that only a finite set could
break his speed limit. So the limit could not be an algebraic number. Hence, the limit
number was transcendental.

Construction of a transcendental number:

To do in detail, the construction is long and tedious with a few proofs. First we define a few
things.

Definition 1
Let [] be polynomials, with all coefficients in . For this entire paper, all polynomials have
coefficients in .
 6

Definition 2
Suppose we have an ef
degree polynomial () not
identical zero (i.e., 1). Notably
Lemma 7 and all else below depends
on the polynomial having coefficients
in , equivalent to . We henceforth
assume this, so unless so identified,
our polynomial is not assumed monic; it Graph of the fifth order polynomial () = . The lone real root at .
is a real algebraic; the other four roots are complex algebraic numbers. By Galois Theory,
may look like = 3 Z + 19 M this polynomial is not solvable (mentioned here), so no roots may be obtained via
81, not monic like algebra and extraction of roots on integers/rational numbers.

19
= Z M + 27
3
Our polynomial follows:
T
= ,-> ,
,
= > + . + M M + + TU. TU. + T T ,

; , , 0, ; T 0

Definition 3
Suppose > , and we have > = 0, that is, > equals an irrational algebraic number.
We fix the number > for the remainder of this paper.

Note we keep fixed until Definition 17, will be referred to without redefinition.

Note when = 1, > is rational, the single zero of a polynomial like = 2 9, and every
rational number is of degree 1; see Definition 4 direct below for a definition of degree when
applied to a number. As > is irrational, deg 2.

Definition 4
When > equals an algebraic number and is the minimal polynomial,the monic polynomial
of rational coefficients, of smallest degree of which > is a root is called the degree of > ; see
also this link for a definition. Polynomials () of arbitrarily large degree satisfy > = 0, but
using a minimal polynomial can give better, larger bounds; see Lemma 7 below. As an
n n
example, 2o equals a root of = M 2, so is second degree. But 2p equals a root
n
of = Z 2 and no polynomial in rational coefficients of lower degree than 5, so 2p is
fifth degree.

Henceforth can simply ignore the word minimal, and all in this paper will work.
 7

Definition 5

Note = deg , the degree of , will come up again.

Definition 6

In this entire paper, , , 0, so and are always integers, and is always rational, as

)*
is . In any fraction , assume 1, as we can always move a negative sign to .
+*

Lemma 7.
Let polynomial [] of be of degree . If , , 0 (assume 1, safe; see
) ) .
Definition 6) and 0, it follows that is always true.
+ + +y

Proof.
) )
As 0 assume = 0. Note is easily rational as is a polynomial
+ +
over . So, {0}. Stating this,
T
,
= , =

,->

Taking absolute values of both sides we have:

T
,
= , =

,->

Now multiply both sides by T , which is positive, so it filters through absolute value brackets:

T = T =

T T
,
T
, = , , TU,

,-> ,->

or,
T

T = , , TU,

,->
 8

Note in bottom right sum that as 0 is always true, and always equals an integer.
And as , , , and , , all terms of aboves bottom right sum are integers. So T
equals some natural number, is in . In short, T 1.

Thus,

T 1

) .
Dividing by T we get .
+ +y

Example 8

As an example of the last, = 2 satisfies = M 2. While 2 = 1.41421 is

)
irrational, its digits never ending with no known pattern, we can approximate it with =
+
.[.[M. Y MMY
= (coprime form). So = 10000. Note
.>>>> .>>>>

M
7 89 227
= 2 1.00759 10UZ
10000

1 1
Note here = 2 (the degree of = M 2), and 2 = = 1010 , and as
100002
1
1.00759 10UZ = 10U
10000M
) .
that seems obvious, by about three orders of magnitude.
+ +y

Calculus Result 9

By the extreme value theorem of calculus, on a bounded interval the derivative of the


polynomial is continuous, thus is bounded. Thus such that [0 1, 0 + 1] we
have .

Discussion 10
)* / )*
We will be looking at rational sequences with lim = > . As the rational numbers are
+* ,-. ,/ +*
dense in the real line, such a sequence exists. So after a finite point all terms of such a sequence
are in the interval [> 1, > + 1], and we can use Calculus Result 9.

Comment 11
 9

)
Our next proposition will establish our first speed limit. Of a rational number , remember that
+

for all irrational numbers > that 0 is always true.

Convention 12
)
For > an irrational root of the polynomial [], assume is close to > in the
+
) )
sense of > 1 < < > + 1, and that is closer to > than any to other root of . In a soso
+ +
graph:

() > + 1

> 1
>

) )
In short, we want 0, why is closer to > than to any other root of .
+ +

We further comment, if is a minimal polynomialsee Proposition 13 direct belowthis

)
convention is not needed. This is as a rational cant be a factor of a minimal which has an
+
)
irrational algebraic root > , as if is a root of = > + . + T>. TU. + T T , we can
+
factor, and for , a degree 1 polynomial, we could get = () ( ), and
for > we would have a new minimal polynomial , of lower degree; see Definition 4.

Proposition 13.

Using Calculus Result 9, we maintain that for any rational number which by Convention 12 is
near our earlierdefined irrational 0 , a root of the ef degree minimal polynomial
() [], that there exists an such that
1
>
T

Proof.
In this proof it should be clear why having () be a minimal polynomial (see Definition 4)
. 1
of > works best, as the exponent on is smallest, and will be largest. All this may work,
+y
 10

for a 100thorder polynomial, but if we use the minimal polynomial which is, say, 5thorder, we
will get the largest bound.

We also point out, (> ) 0: If > = 0, that means > is at least a double root of (),
which for a minimal polynomial doesnt make sense: Could differentiate (), and have a
polynomial of one less degree of which > equals a root. So > > 0. In short, there is
no limiting out of this propositions conclusion.

To work: Let () be the minimal polynomial of > . Further let > 0 be as in Calculus Result
9, that is, such that [> 1, > + 1] we have .
) )
Note > is irrational, so > must be true. By Convention 6, is nearer to > than to any other
+ +
) )
root of , so = 0 is impossible. That 0 will be of some importance.
+ +

)
By the Mean Value Theorem, between and > the derivative of assumes a mean value,
+
) )
that is, min , > , max , > such that
+ +

(> )

= ()
>

As > = 0, we may restate the left side,

> 0

=
> >

Simplifying, and equating to the right side:

=
>

Taking absolute values, from our chosen value of in Calculus Result 9 this gives



= ()
>

But from Proposition 13 we know that

 11

1
.
T
) )
Aside: The last is where 0 matters, a hypothesis of Proposition 13, and why we choose
+ +
)
closer to > than to any other root of , as this renders = 0 impossible. All that truly
+
)
matters is that 0. It is also key that > = 0, a hypothesis of this proposition.
+

) .
Joining with () , and doing a little algebra
+ +y

=
>

1 1 1
= T
> >

Restating the top left and bottom right terms:

1 1

T >


We divide by and multiply by 0 to derive

1
>
T

This verifies our inequality, and establishes our first speed limit.

Discussion 14
)* / )*
We will be looking at rational sequences with lim = 1 ; we will later define . .
+* ,-. ,/ +*

For an irrational algebraic > for some large = > we have

)* /
The sequence > 1, > + 1 , > , and
+* ,-,
_

some positive number will be a bound of on [> 1, > + 1], and

 12

Convention 12 will be true > , so

Proposition 13 will be true, > .
)
Proposition 13 showed > has to be bigger than something. The next proposition shows that
+
)
for only a finite set of rational numbers can > be smaller than a stronger bound, and
+
matters in Theorem 18 that proves the existence of a transcendental number.

We note the ThueSiegelRoth theorem has stronger bounds than the next. Its proof is not easy,
and led to a Fields Medal. What is relevant is that > = deg ().

Proposition 15.
Let [] be the minimal polynomial of an irrational root > . Fix a real > 0. Fix a natural
number such that

> = ()


Let
be an infinite set of unique rational numbers, that is, whenever , we have
=1
, 1

, 1
For some index set there exists an at most finite set of unique rational numbers,

,
= ; , , , , , 0
,
such that

,
> <
, ,

That is, the index set is finite, hence is a finite set. We assume { 1, }; see Definition
6.

Proof.
We do a proof by contradiction. Assume for an infinite index set (as the rational numbers form
an 0 set, we choose = ) there exists a set
,
= ; , , , , ,
,
 13

)*
with each unique within the index set , such that for all and for our fixed > 0 and
+*
> that we have
,
> <
, ,
/
)* )* )*
0
By > < we know that lim = > . Pass to a subsequence and assume
+* +* ,/ +* +*
0 1-.
/
,0
(> 1, > + 1),
,0
1-.
)*
0
and that each is closer to > than to any other root of (). So Proposition 13 holds.
+*
0

Now we show ,0 is an unbounded set.

If the denominators ,0 are restricted to a bounded set (i.e., live within a finite interval, so
there can only be a finite number of ,0 ) we get a contradiction: as the set is infinite and


each is unique within the set , we have lim ,0 = forced. As the ,0 live in a bounded
1/

interval and lim ,0 = , after some finite point the inequality

1/

> <

,0

)*
0
must be proven false: as lim ,0 = , the term > , but < and 1, so
1/ +*
0

is bounded, by at most.
+*
0

Thus lim ,0 = is forced. And, an important inequality: by passing to a subsequence, we

1/
assume that for all 1 that

>

,0
is true, for all 1 .
 14

As > , we have 1. Now apply Proposition 13: On the interval [> 1, > + 1] by

some > 0 bound the absolute value of the derivative of . As > is irrational, note the

polynomial is of degree at least 2; see Definition 3, so is nonzero.

Now we derive our contradiction. Where equals a bound of in the range [> 1, > + 1],

and remembering > is an irrational root of our degree polynomial , and applying
Proposition 13 we get:

,0 1
> > T >0
,0 ,0
,0

Tossing the middle, and reversing the inequality,

1
T <
,0 ,0

Doing algebra, the following is always true:

UT
,0 <

But as lim ,0 = , and 1, for all but an at most finite set of we have forced
1/

,UT
0
>

UT
violating the inequality ,0 < . Thus our initial assumption was wrong, and any set
as in the propositions statement is necessarily finite this will be key to proving a number
transcendental.

This verifies the proposition, and is our second speed limit.

Discussion 16
Putting together Proposition 13 and 15 with = 1, let > be an irrational algebraic number, and
assume () the minimal polynomial of 0 , as the minimal polynomial gives the largest bound.
Let = deg . Fix = + 1, and a bound as in Calculus Result 4.
 15

)* / )*
In the next equation, for any sequence of unique rational numbers with lim = > ,
+* ,-. ,/ +*
)
whenever * is close to > (i.e., after some finite point), for a bound > 0 on the absolute
+*
value of the derivative of on [> 1, > + 1] we have

. )
* > (always true)
+*y +*

For the next we have a best a finite set of , :

) . )*
* > < (a finite set of )
+* +* +*

Summarize: ignoring finite sets, in general for any rational sequence of unique rational
)* / )* )
numbers with lim = > , whenever * is close to > (i.e., after some finite point > , for
+* ,-. ,/ +* +*

all > ), for the above bound > 0 and > we have two speed limits,

)* .
> (always true)
+* +*y

and,

)*
> < only true for a finite subset of ,
+* +* =1

/
) *0
so we can pass to a subsequence to make it never true
+ *0
1-.

The second of these facts is extremely important in Theorem 18, where we prove a number
transcendental.

Now we create a transcendental number, i.e., a number . such that . is not the root of any
polynomial with () [].

Definition 17
We define a new number . in a special way. For now assume it an algebraic number, that is, the
root of a minimal polynomial of some degree , and, the number will play a role.
Let
 16

. = 10U!
-.

In the exponential we use factorial function, as in 5! = 5 4 3 2 1 = 120.

Apply a standard calculus test, the ratio test:

/
In a generic series -. , we have = 10U! . Apply the ratio test:

. 10U(.)!
= =
10U!

10! 10!
= =
10 . ! 10 . !
10!

10! .
It is easy that this converges to zero as , so the conditions of the root test are satisfied. We
!
have 1 = =1 10 is an absolutely convergent series, so, is convergent.

The number 1 is easily an infinite sum of rational numbers, and at each point we define a term
)* /
of an infinite sequence of rational numbers as
+* ,-.
,
, 1 1 1 1 1
= 10U! = + + + + +
, 10.! 10M! 10E! 10 ,U. ! 10,!
-.

Algebra, to get a common denominator:

, 10,!U.! + 10,!UM! + 10,!UE! + 10,!U ,U. !

+ 10,!U,!
=
, 10,
. , ,!U!
So as formulas, , = , and , = -. 10 . So,
.>*!

,
, 10,! + 10,!U.! + 10,!UM! + 10,!U ,U. !
+1 -. 10
,!U!
= =
, 10,! 10,!
 17

)*
and this converges to . , so we have lim = . .
,/ +*

We need to verify that Proposition 15 holds. It is easy to assign the denominator at each finite
point as , = 10,! . And, it is easy that each , is unique. If , note that

, = 10,! + 10,!U.! + 10,!UM! + + 1

and

1 = 101! + 101!U.! + 101!UM! + + 1

, ,!U! 1 1!U!
satisfy , 1 = -. 10 -. 10 , which as is nonzeroif > , it
,!U.!
possesses 10 which cannot cancel, and , and 1 are consequently different numbers. Also,
)
for any , note always has a final digit of 1, so in no canceling is possible, as equals a
+
power of 10.

)* /
So, is a set of unique rational numbers, and Proposition 15 holds, that is, for any
+* ,-.
fixed > and any fixed , there exists an at most finite set , indexed by some finite
set = {1, 2, 3 },

,
= ; , , , , , 0
,
such that

,
. <
, ,
Note . is called the Liouville Constant, and . = 0.110001000000000000000001000
with the 1s getting progressively farther and farther apart. As digits of . never repeat, it is
easily irrational. This means deg 2; see Definition 3.

Now we have some tools, our big theorem follows, and is to what we have been building up.

Theorem 18.

There does not exist a polynomial of which . is a root.

Proof.
 18

Assume the contrary, i.e., that the number . equals a root of some
minimal polynomial , and is of some (possibly
redefined) degree ; the degree matters, and we will refer to it.
The proof assumes a known number, but this proof works, for any
finite . Remember, as . is irrational, must be at least 2; see
Definition 3.
The geometric constant is
From Definition 17 we have a sequence of unique rational transcendental, but this was
proven in 1882, over thirty years
)* /
numbers with lim = 1 . after Liouvilles proof. This
+* ,-. meant squaring the circle is
impossible, solving a problem
Remember, the denominator at each finite point is , = 10,! . that had been open for
thousands of years.

First we draw a bound. Fix = + 1.

Go back to our definition of . . The next line requires small reordering, but thats easy, inside
absolute value brackets. Note that at any finite point that

, /
,
. = 10U! 10U! =
,
-. -.

/
1 1
10U! = ,. !
+ ,M !
+
10 10
-,.

Or, simpler,

) . .
* . = *n ! + + (Equation 1)
+* .> .> *o !

Now draw bounds: First let = + 1 = + 2. As in Definition 17 we pointed out that the
degree of () equals at least 2, we must have 4; see Definition 3; this will be mentioned,
later. Thus as = 1

+ 1 ! = + 1 ! >

1 ! = !

. .
In short, + 1 ! > !, and just as 3 > 2 leads to < , we derive
E M
 19

1 1
+1 !
<
10 10!
So, referring back to Equation (1) above, and adding some parentheses,

) . . .
* . < *! + *o ! + + (Equation 2)
+* .> .> .> * !

For our fixed and , for some we may easily write the right side as

. . .
= *! + *o ! + + (Equation 3)
.>*! .> .> .> * !

.
Finding is solving = + for , and

1 1 1
= 10,! !
+ ,M !
+ ,E !
+
10 10 10
Or,

. .
=1+ *o !*! + * !*! + (Equation 4)
.> .>

On last, note that for 2, the number 10 ,1 !U,! is a positive power of ten: As we have here
fixed = + 2, and as 2, we have 4; see Definition 3. As

+ 2 = + 4, and
+2> =+1

we have + ! ! is an increasingly large exponent, and the rightside terms of Equation

4 get small, extremely fast. As an example, if = 2, we have = 4, = 3, and set out a few
.
exponents on :
.>

1. = 4 + 2 ! ! = 6! 3 4! = 720 72 = 648
2. = 5 + 2 ! ! = 7! 3 5! = 5040 360 = 4680
3. = 6 + 2 ! ! = 8! 3 6! = 40320 2160 = 38160

Also note is known, so = + 1 = + 2 are known, and equals a value that can be
determined. As the inequality is strict, it is easy to be a bit sloppy.

For = + 2 and larger values of , the term

 20

1 1
,E !U!
+ ,[ !U!
+
10 10
only gets smaller. So in the bound, the same value of will work, regardless of how big is.

So combining Equations 2 and 3, we draw an important bound, for all + 1:

/
,
. = 10UT! <
, 10,!
-,.

Or, simpler,

,
. < ,!
, 10
)* / .
Note is an infinite sequence of unique rational numbers, and remember, , = , .
+* ,- .>*!
For each > > :
,
. < ,! =
, 10

=
10,!
,
In short,

,
. < , >
, ,
)* /
But this is true for all > = + 1, and is an infinite set, a direct contradiction to
+* ,-.

Proposition 15 and impossible if . is an algebraic number. Thus . is not an algebraic number.

Final conclusion 13
!
By Theorem 12 the number 1 = =1 10 cannot be the root of any polynomial with
coefficients in (equivalent, ), hence . is not an algebraic number. That is to say, . equals a
transcendental number.

Final comment 14
Again, just in the world of real numbers, as Cantor proved
 21

the algebraic numbers are countable, of cardinality > and

the reals are uncountable, of cardinality 2_ ,

it follows that transcendental numbers must be an uncountable set, that is, of cardinality 2_ . For
such a large set, as of June 2017 precious few numbers have been proven transcendental.