Infinite Sets and Infinite Sizes

Gary Hardegree
Department of Philosophy
University of Massachusetts
Amherst, MA 01003

1. Review .............................................................................................................................................1
2. Infinite Sets ......................................................................................................................................2
3. Comparing Infinite Sets ...................................................................................................................4
4. Denumerable Sets ............................................................................................................................6
5. Uncountable Sets..............................................................................................................................8
6. There are Infinitely-Many Infinite Sizes..........................................................................................9
7. What will Really Cook Your Noodle ............................................................................................10
8. Russell’s Paradox...........................................................................................................................10
9. You Must Remember This.............................................................................................................12
10. Appendix; Various Proofs..............................................................................................................13
1. A Proof that √2 is Irrational ...............................................................................................13
2. A Proof that the Subsets of Á are Uncountable.................................................................13
3. A Proof that the Irrational Numbers are Uncountable .......................................................14

1. Review
We have seen that, in English, number-words are used both as adjectives and as nouns, as
illustrated in the following examples.

I have two brothers

two is the only even prime number

Our semantic account of number-words furthermore maintains a straightforward relation between

numerical-nouns and numerical-adjectives, given as follows.

a numerical-noun names what a numerical-adjective expresses

Adjectives express properties, and numerical-adjectives in particular express properties of sets,

specifically their sizes. To put it succinctly:

numbers are sizes

The size of a set is how many members it has. The following list summarizes the various cases.

0[A] i.e., A is memberless

1[A] i.e., A is single-membered
2[A] i.e., A is double-membered
etc.

These concepts in turn have rigorous logical formulations as follows.

Hardegree, Infinite Sets and Infinite Sizes page 2 of 16

0[A] ü ;∃x { x ∈ A }
1[A] ü ∃x { A = {x} }
2[A] ü ∃x∃y { x≠y & A = {x,y} }
3[A] ü ∃x∃y∃z { x≠y & x≠z & y≠z & A = {x,y,z} }
etc.

The size of set A – denoted #(A) – is then defined as follows.

#(A) = 0 ↔ 0[A]
#(A) = 1 ↔ 1[A]
#(A) = 2 ↔ 2[A]
etc.

2. Infinite Sets
How many natural numbers are there? How big is the set Á of natural numbers, which is
defined as follows.

Á ü { x : x is a natural number }
ü {0, 1, 2, …}

Upon reflection, the following should be apparent.

#(Á) ≠ 0 Á is not empty

#(Á) ≠ 1 Á does not have exactly one member
#(Á) ≠ 2 Á does not have exactly two members
etc. etc.

This is associated with the intuition that, no matter how far you count, you can always count farther.
The natural numbers are endless! Another way to describe this is to say that:

there are infinitely-many natural numbers

This poses a serious conceptual problem. Natural numbers are sizes, but no natural number is
the size of the set of all natural numbers. Given this result, there seem logically to be two choices.

(1) the natural numbers do not form a set, contrary to what we presumed above, and
accordingly the natural numbers do not have a size (no set, no size);
(2) there is at least one size that is not a natural number.

A minority of mathematicians and philosophers deny set-hood to the natural numbers, so for them the
problem of how many natural numbers there are does not arise. For these mathematicians and
philosophers, there are no actual infinities, but only potential infinities.1

1
By similar reasoning, although there may be a boundless future, which is a potential infinity, there cannot be a boundless
past, which would be an actual infinity. On the other hand, it seems that whether the past is infinite or not is an empirical
matter, not a logical matter. To be sure, our minds are boggled when we consider the possibility of an infinite past, which
might even contain infinitely-many actual humans. But boggling the human mind is not a reliable sign of impossibility, as
history has repeatedly attested. I accordingly reject this sort of reasoning.
Hardegree, Infinite Sets and Infinite Sizes page 3 of 16

Most mathematicians and philosophers, however, are perfectly happy to grant set-hood to the
natural numbers, and even more vast collections, and accordingly must come to terms with the question.

what is the size of the set of natural numbers?

#(Á) = ?

The glib answer is “infinity!”. Many modern school children can even recite this, and may even be able
to reproduce a symbol for it – T. More interesting perhaps, the glib answer seems to have satisfied
humanity for thousands of years, and even the greatest minds were unable to get past it.

This all changed in the 19th Century when Georg Cantor2 began to enquire a little more deeply
into the issue of infinity, and in the process invented both set theory and the theory of infinite numbers.
We have already discussed the basic insight – that two sets are equally large precisely when one can set
up a one-to-one correspondence between them. What we haven’t discussed, however, is how this
insight can be applied to infinite sets.

First, a set is finite if and only if it can be counted by a natural number.

A is finite ü #(A) = n for some natural number n

The definiens (right side of the definition) can in turn be thought of as an infinite disjunction with the
following disjuncts.

#(A) = 0 or
#(A) = 1 or
#(A) = 2 or
etc.

On the other hand, a set is infinite if and only if it is not finite, which means that it cannot be counted by
a natural number.

A is infinite ü A is not finite

In this case, the definiens can be thought of as an infinite conjunction with the following conjuncts.

#(A) ≠ 0 and
#(A) ≠ 1 and
#(A) ≠ 2 and
etc.

The set Á of natural numbers is a prime example of an infinite set:

#(Á) ≠ 0 and
#(Á) ≠ 1 and
#(Á) ≠ 2 and
etc.

2
Cantor was born in St Petersburg, Russia on March 3, 1845, and died in Halle, Germany on Jan 6, 1918. Cantor’s father
was German; his mother was Russian. By way of a little historical context, St. Petersburg was the capital of Russia. Czar
Nicholas was Russian, Czarina Alexandra was German, and also a granddaughter of Queen Victoria of England. The Royal
couple, along with their five children and several servants, were executed by firing squad by the Bolshevi ks on July 17, 1918.
Hardegree, Infinite Sets and Infinite Sizes page 4 of 16

Notice, however, that Á is merely one of many examples of infinite sets; the following are a few more
examples.

the set of even numbers {0, 2, 4, 6, …}

the set of odd numbers {1, 3, 5, …}
the set of prime numbers3 {2, 3, 5, 7, 11, 13, …}
the set of numbers Õ 10 {10, 11, 12, …}
the set of multiples of 10 {10, 20, 30, …}
the set of powers of 10 {1, 10, 100, 1000, …}

3. Comparing Infinite Sets

We have discussed size-comparison at great length, beginning with the Generic Theory of Size.
We know how to compare the sizes of finite sets; how do we compare the sizes of infinite sets? Under
what circumstances do we say that one infinite set is bigger than another infinite set? Does this even
make sense? For example, which of the following sets is bigger?

the set of even numbers

the set of all numbers

Here is an argument that the latter is bigger than the former.

A1. half the numbers are even, and half the numbers are odd;
therefore, there are twice as many numbers as there are even numbers;
therefore, there are more numbers than even numbers.

Here is a more general argument.

A2. every even number is a number, but not every number is an even number;
therefore, the set of even numbers is a proper subset4 of the set of all numbers;
therefore there are more numbers than even numbers.

Now, it is pretty clear that the above reasoning applies to finite sets and finite sizes. What is not
so clear is whether it applies to infinite sets and infinite sizes. What Cantor discovered was that many of
our intuitions about size do not apply to infinite sets. But what intuitions do apply?

Recall that in order to compare the sizes of two sets, we do not have to count them. Even before
humans could count, they could compare sets directly; counting was an invention for indirect
comparison of sets. If we can compare the sets directly, then counting is unnecessary. For example, I
don’t have to count the fingers on my right hand, and the fingers on my left hand, to know that my two
hands have equally-many fingers. I don’t even have to know how to count! All I have to do is pair up
the fingers of the two hands.

Now, here is Cantor’s insight.

3
A number is composite if and only if it can be factored; for example, 6 can be factored into 2 and 3; i.e., 6 = 2ö3. A
number is prime if and only if it cannot be factored; for example, 7 cannot be factored.
4
We say that A is a subset of B precisely if every member of A is also a member of B. We say that A is a proper subset of
B precisely if A is a subset of B but B is not a subset of A [in which case A ≠ B].
Hardegree, Infinite Sets and Infinite Sizes page 5 of 16

sets A and B are equally-big

if and only if
one can set up a one-to-one correspondence between
elements of A and elements of B,
even if A and B are infinitely-big.

So back to the question whether the following sets are equally-big.

the set of even numbers

the set of numbers

To see that they are equally big, we need merely set up a one-to-one correspondence between them. The
following is probably the simplest.

0 1 0
1 1 2
2 1 4
3 1 6
etc.

We can do the same thing with all the sets mentioned above.

0 1 1
1 1 3
2 1 5
3 1 7
etc.

0 1 2
1 1 3
2 1 5
3 1 7
etc.

0 1 10
1 1 11
2 1 12
3 1 13
etc.

0 1 10
1 1 20
2 1 30
3 1 40
etc.

0 1 1
1 1 10
2 1 100
3 1 1000
etc.
Hardegree, Infinite Sets and Infinite Sizes page 6 of 16

4. Denumerable Sets
It is customary to call a set denumerable if it has the same size as the set Á of natural numbers.
In other words:

A is denumerable ü #(A) = #(Á)

Accordingly, the following sets are all denumerable.

the set of even numbers {0, 2, 4, 6, …}

the set of odd numbers {1, 3, 5, …}
the set of prime numbers {2, 3, 5, 7, 11, 13, …}
the set of numbers bigger than 10 {11, 12, 13, …}
the set of multiples of 10 {10, 20, 30, …}
the set of powers of 10 {1, 10, 100, 1000, …}

Notice, however, that we do not yet have a special symbol for the size of Á. We have names of
all the finite sizes – 0, 1, 2, … But we have no names for infinite sizes yet. For this purpose, Cantor
introduced the following symbol.

ℵ0

which is read “aleph naught” or “aleph zero”.5

Let us next consider some proper super-sets of Á.6 As the reader is doubtless aware, the natural
numbers have negative counterparts – the negative integers – which are defined as follows.

the set of negative integers ü {−1, −2, −3, …}

These are not "natural" numbers, because they are not set sizes; for example, “minus four” is not an
admissible answer to a how-many question. Nevertheless, the negative integers are valuable
mathematical devices for other measurement purposes, like measuring the temperature7 on a very cold
day, or measuring your net worth if you are in serious debt!

Next, we obtain the set of integers by combining the negative integers with the non-negative
integers (i.e., natural numbers).

the set of integers ü {−1, −2, −3, …} ∪ {0, 1, 2, 3, …}

How big is the set of integers? It is also denumerable, as seen by the following one-to-one
correspondence.

0 1 0
1 1 1
2 1 −1
3 1 2
4 1 −2
etc.

5
Note that aleph is the first letter of the Hebrew alphabet, or what is sometimes called the ‘alephbet’, since its first two letters
are not alpha and beta, but aleph and bet. Perhaps, we should call our alphabet the ‘ay-bee’.
6
A is a super-set of B if and only if B is a subset of A.
7
In degrees Fahrenheit, or in degrees Celsius, but not in degrees Kelvin, which does not have negative temperatures.
Hardegree, Infinite Sets and Infinite Sizes page 7 of 16

There are other numbers besides integers, the simplest being the fractions. In this context, by a
fraction I mean a quantity that may be obtained by dividing one positive integer by a larger positive
integer.8 English provides common words for the simplest of these quantities:

one-half one-third one-fourth …

two-thirds two-fourths two-fifths …
three-fourths three-fifths three-sixths …
… … … …

As with the negative integers, the fractions are not natural numbers; for example, “one-third” is not an
admissible answer to a how-many question. On the other hand, it can serve as an answer to an how-
many-of question, as in “how many of the students got an A?” Similarly it can serve as an answer to a
how-much-of question, as in “how much of the pizza did you leave for me?”

How many fractions are there? If a table has m rows, and n columns, then it has a total of mön
cells. Now the chart depicted above has infinitely-many rows and infinitely-many columns; what is ¢0
ö ¢0? Surely, a "zillion-zillion" is bigger than a "zillion"! If that is not convincing, consider the
following reasoning: between any two fractions, there is another one; for example, between one-third
and one-half is five-twelfths. So, when you go to count all the fractions, since they are so densely
packed, you can’t even get from one to the next! No matter what you want to count next, you have to
count something else first!

Notwithstanding these initial intuitions, as it turns out, there are no more fractions than there are
natural numbers. The proof of this astonishing fact employs the following counting technique invented
by Cantor. Here, for the sake of brevity, we write the fractions in the customary fashion.

1/2 " 1/3 1/4 " 1/5 1/6 " 1/7 …

' & ' & ' &
2/3 2/4 2/5 2/6 2/7 2/8 …
$ & ' & ' & '
3/4 3/5 3/6 3/7 3/8 3/9 …
' & ' & ' &
4/5 4/6 4/7 4/8 4/9 4/10 …
$ & ' & ' & '
5/6 5/7 5/8 5/9 5/10 5/11 …
' & ' & ' &
… … … … … … …

In order to carry out the counting (pairing), one follows the arrows, starting at 1/2, then going to 1/3,
then to 2/3, then to 3/4, and so on. 9

8
These are sometimes called ‘proper fractions’, which are then distinguished from "improper" fractions, which include (e.g.)
zero-thirds, three-thirds, and seven-thirds.
9
Notice that we have made no provision for the minor detail that one-half and two-fourths are presumably the same
quantity, but this does not affect the result. If we wanted, we could skip each fraction that has already been counted under a
separate guise; for example, when we get to 2/4 we skip it since we have already counted it when we counted 1/2.
Hardegree, Infinite Sets and Infinite Sizes page 8 of 16

5. Uncountable Sets
Next, we introduce the concept countable, and its complement uncountable.

A is countable ü A is finite, or A is denumerable

A is uncountable ü A is not countable

Recall that a set is denumerable precisely if it has the same size as Á.10

Are there any uncountable sets. From what has transpired so far, one might naturally conclude
that all infinite sets are denumerable, which implies that all infinite sets have the same size, which
implies that there is just one infinite size – namely, ℵ0. In other words, infinite is infinite…period!
Accordingly, there are no uncountable sets.

Once again, however, our intuitions are trampled; in particular, as first proved by Cantor, we
have the following.

(1) the set of irrational numbers is uncountable.

(2) the set of all subsets of Á is uncountable.

First, a rational number is not a number that behaves rationally, nor is it a number that believes that all
knowledge proceeds from reason. Rather, a rational number is one that can be expressed as a ratio of
two integers. In other words.

x is rational ü ∃m∃n { m and n are integers and x = m/n }

On the other hand, an irrational number is one that cannot be expressed as a ratio. The most famous
irrational number is π (pi), which is the fundamental mathematical constant that relates a circle’s
diameter to its circumference.11

The discovery of the irrational numbers was revolutionary. Legend has it that a disciple of
Pythagoras12 took the master’s famous theorem, pictured as follows,

10
Notice that Denumerable = Infinite + Countable. Sometimes, however, uncountable sets are called ‘non-denumerable’.
Notice the mildly illogical fact that ‘non-denumerable’ does not mean ‘ not denumerable’. A set that is not denumerable may
be either uncountable or finite, whereas a non-denumerable set cannot be finite.
11
The letter ‘π’ probably comes from the Greek word for ‘perimeter’ which is ‘perimetron’ (περιµετρον).
12
Pythagoras (ca. 569 – ca. 475 BC) was born in Ionia on the west coast of Asia Minor (present day Turkey), and traveled
extensively – including to Egypt and Babylonia, where he evidently consulted extensively with the high-priests, who were
the guardians of mathematical knowledge. Following the Persian conquest of Asia Minor, Babylonia and Egypt, Pythagoras
moved west and founded a philosophical and religious society in Croton (modern day Crotone in southern Italy). Its
practices included communal living and vegetarianism, at least for the inner circle, who were known as mathematikoi. On
the other hand, the outer circle, known as akousmatics, were permitted to live in their own houses, to have their own
possessions, and to eat meat. Consult: [http://www-groups.dcs.st-and.ac.uk/~history/Mathematicians/Pythagoras.html].
Hardegree, Infinite Sets and Infinite Sizes page 9 of 16

and applied it to the case in which a = b = 1, and concluded that the hypotenuse c must have a length of
√2. He/she 13 went on to demonstrate that this number cannot be expressed as a ratio of integers, and is
accordingly irrational. As the story continues, since this discovery seriously undermined the religious
teachings of Pythagoras, the disciple was subsequently tossed over the side of a boat in the Adriatic Sea,
thereby perpetrating one of the earliest (attempted) cover-ups in history.

The proof that √2 is irrational, and the proofs of Cantor’s two results, are presented in an
appendix (Section 10).

6. There are Infinitely-Many Infinite Sizes

Let A and B be sets. Then A is said to be included in B precisely if every member of A is also a
member of B. This definition may be symbolically rendered as follow.

AÚB ü ∀x { x ∈ A ² x ∈ B }

Alternatively, A is said to be a subset of B. Next, the power set of A is, by definition, the set of all
subsets of A. In symbols:

Ó(A) ü {X:XÚA}

The following are examples.

Ó(∅) = {∅} 1 member

Ó{1} = { ∅, {1} } 2 members
Ó{1,2} = { ∅, {1}, {2}, {1,2} } 4 members
Ó{1,2,3} = { ∅, {1}, {2}, {3}, {1,2}, {1,3}, {2,3}, {1,2,3} } 8 members

Notice that this is the beginning of a process known as a "combinatorial explosion"; in particular, each
set is twice as big as the set before; accordingly, this list does not proceed too long before the set in
question cannot be physically encoded by humans or even their machines. For example, to physically
write all the subsets of the set of the first 1000 numbers would require more molecules than have ever
existed in the universe.

Nevertheless, we can write a simple theorem about the sizes of these sets, which is given as
follows.

Ó(A) is bigger than A, and in particular:

if #(A) = n , then #(Ó(A)) = 2n

This pertains to finite sets. Is there a corresponding result for infinite sets? According to the more or
less standard view, we have the following.

Ó(A) is bigger than A, and in particular:

if #(A) = ¢n , then #(Ó(A)) = ¢n+1

In particular, there are infinitely-many infinite-sizes, as follows.

13
Note that women were permitted to be, and were, members of the Pythagorean Society.
Hardegree, Infinite Sets and Infinite Sizes page 10 of 16

¢0 the first infinite size (countably-infinite)

¢1 the second infinite size (uncountably-infinite)
¢2 the third infinite size (seriously-uncountably-infinite)
etc.

7. What will Really Cook Your Noodle 14

Clearly, there are infinitely-many infinite sizes:

¢0, ¢1 ¢2, …

Unfortunately, this infinite series is just the beginning of a staggering array of infinite sizes.15 How
many infinite-sizes are there? For this, even the word ‘infinitely-many’ is totally inadequate. We
simply have no further concept for how many infinite-sizes there are.16 As a pathetic substitute for a
concept, we propose a word – ‘hyper-infinite’. One would think that surely we can figure out what
‘hyper-infinite’ means just as Cantor figured out what ‘infinite’ means. This proves more difficult than
anyone might have suspected, as we see in the next section, which will further cook your noodle.

8. Russell’s Paradox
In 1900, the ornate but well-mannered (dare I say "Victorian") world of sets that Cantor had
established came crashing to the ground, when Bertrand Russell noticed that Cantor’s theory of sets
contained a fatal flaw. Cantor’s theory of sets, further developed by Gottlob Frege,17 which is now
called ‘naïve set theory’, is based on the following two very simple postulates.

(s1) the axiom of extensionality;

no two sets have the very same members;
∀x { x ∈ S1 ↔ x ∈ S2 } ² S1 = S2
(s2) the axiom of comprehension;
for any property, there is a set that comprehends18 precisely the objects with that
property;
∀à ∃S ∀x { x ∈ S ↔ x has à }

14
I unashamedly borrow the expression ‘what will really cook your noodle’ from the movie “The Matrix”. In one of the
scenes, the Oracle tells Neo not to worry about the vase near him, as a result of which Neo turns quickly around only to
knock the vase off the table, thereby shattering it. He asks how she knew this would happen (well, duh Neo, she’s the
Oracle!) She tells him not to worry, and goes on to say that what will really "cook his noodle" is when he wonders whether it
would have happened even if she hadn’t warned him. This is a movie that presents numerous philosophical perplexities (in
Greek aporia).
15
By ‘staggering’ I mean staggering even to mathematicians who are completely comfortable dealing with infinities of
infinities of infinities of …
16
The difficulty is that any attempt to wrangle all the infinite sizes into a set to contain them all involves a complete logical
collapse of set theory. See Section 8.
17
Friedrich Ludwig Gottlob Frege (1848-1925) invented modern symbolic logic, and also the entire field of semantics.
18
The word ‘comprehend’ basically combines ‘com’ [together] and ‘prehend’ [grasp]. In this connection, note that some
monkeys have prehensile (i.e., grasping) tails. The first meaning of ‘comprehend’ is understand; we even use the word
‘grasp’ as an occasional synonym. Its second meaning, which leads to the derivative word ‘comprehensive’, which means
to take in as a part. Note that the cognate word ‘comprise’ produces serious usage problems, insofar as it is often used in a
manner exactly opposite its literal meaning. In particular, the word ‘comprise’ comes from ‘comprisé’, which in French is
the past participle of the verb ‘comprendre’, which of course is a cognate of ‘comprehend’. If we use the word according to
its strict meaning, the U.S. comprises its 50 states, whereas the 50 states do not comprise the U.S.
Hardegree, Infinite Sets and Infinite Sizes page 11 of 16

In this connection, it is useful to introduce a common philosophical notion – extension. The extension of
a property is, by definition, the set of all objects that have that property. In other words:

the extension of à ü { x : x has à }

We can then interpret the Axiom of Comprehension as saying the following.

(s2*) every property has an extension

Thus far, we have taken this principle for granted. In particular, whenever we have written down
a set-abstract of the form

{x:…x…}

we have naturally presumed that such a set exists.

This has never seemed to be an issue. The following seem to be some fairly natural examples,
both simple and complex.

property corresponding set

being a natural number natural numbers
being an even number even numbers
being a prime number prime numbers
being a memberless set memberless sets
being a single-membered set single-membered sets
being a double-membered set double-membered sets

The sets all look pretty clear-cut, so what could be the problem? Well, basically Russell did to Cantor
and Frege’s theory of sets what Eubulides19 did a couple of millennia earlier to Aristotle’s theory of
truth.20 In particular, he presented a paradox that showed that the theory is logically inconsistent. Now
logical inconsistency is to theories what 50 megaton bombs are to cities – fatal!21

Russell asked us to consider a property that we will express by the word ‘normal’; in particular,
an object is said to be normal precisely if it is not a member of itself. In other words:

x is normal ü x∉x

Notice that any object that is not a set is automatically normal; only sets have members. Furthermore,
most sets are normal; indeed, it is hard to imagine a set that is not normal. There is nevertheless a
prominent example of an abnormal set – the universal set È – which is defined as follows.

È ü { x : x=x }

19
Eubulides of Megara (not far from Athens) was a contemporary, and continual philosophical opponent, of Aristotle.
20
This is the famous "Liar Paradox". Suppose that I say that I am lying, and nothing else; then, am I telling the truth, or am I
lying?
21
Only one such "mega-bomb" has ever been detonated, in 1961 by the Soviet Union. Witnesses report that the blast
knocked down people fifty miles away, and that the fireball was visible 600 miles away! For the sake of comparison, the
Soviet mega-bomb was 2500 times more powerful than the bombs unleashed on the unfortunate cities of Hiroshima and
Nagasaki.
Hardegree, Infinite Sets and Infinite Sizes page 12 of 16

It is easy to show that everything is a member of the universal set, since everything is self-identical. So,
È is a member of È, so È is abnormal.

Now, according to Axiom of Comprehension, every property à has an extension, which is just
the set of objects that have that property. The property of being normal is a property, so it has an
extension, which is simply the set of normal objects. Let us call that set q. In other words:

q = { x : x is normal }

What follows from this and the Axiom of Comprehension is that:

∀x (x ∈ q ↔ x is normal)

But this formula applies to everything, so it applies to q, so substituting q for x yields the following
seemingly innocuous claim.

q ∈ q ↔ q is normal

In other words q is an element of the set of normal things if and only if q is normal. But what does it
mean to be normal? Well, we know that

q is normal ↔ q ∉ q

so, putting these together, we get

q∈q ↔ q∉q

which has the following logical form:

P ↔ ;P

The latter, however, is a self-contradiction, since it says that the truth-value of P is the same as the truth-
value of ;P.

Ouch!

As a result of this crisis, a number of mathematicians and logicians undertook to reformulate Set
Theory in a way that avoids logical inconsistency. What was produced was Axiomatic Set Theory, as it
is called. We will not venture into this subject. Suffice it to say that Axiomatic Set Theory is generally
regarded as a reliable foundation for mathematics, and no one has proved that it is inconsistent… yet.

9. You Must Remember This

We have gone far enough for the moment. The main thing to keep in mind are the following
facts, which are sufficiently mind-boggling in their own right.

(1) there are just as many even numbers as natural numbers!

(2) there are just as many natural numbers as rational numbers!
(3) there are more irrational numbers than rational numbers!
(4) there are more subsets of natural numbers than natural numbers!
Hardegree, Infinite Sets and Infinite Sizes page 13 of 16

10. Appendix; Various Proofs

1. A Proof that √ 2 is Irrational

Below, we offer a proof (in more or less standard mathematical form) that √2 is irrational.

In order to prove that √2 is irrational, we proceed by reductio ad absurdum (alias indirect

derivation). In particular, we assume that √2 is rational, and show that this leads to
contradiction. So, suppose to the contrary that √2 is rational. Then by definition, it is equal to a
ratio of integers. Let us suppose those integers are m and n; i.e., √2 = m/n. By a well-known
theorem of arithmetic, every fraction can be "reduced" to a fraction in which the two integers
have no common factors [for example, 3/6 reduces to 1/2]. Let us accordingly reduce m/n to its
simplest form, which we suppose is p/q. Accordingly, we have:
(1) √2 = p/q where p and q have no common factors
By a few simple algebraic manipulations, we can deduce the following equation.
(2) p2 = 2q2
From this it follows that p2 is even, from which it follows that p is even [since odd ö odd =
odd].
It now follows that q is odd. For if q is even, then both p and q are even, in which case they
share 2 as a common factor, which is contrary to (1).
As stated earlier, p is even, so we have the following – for some integer, say, r.
(3) p = 2r
Squaring both sides yields:
(4) p2 = 4r2
Combining (4) with (2), we obtain:
(5) 2q2 = 4r2
from which we immediately obtain:
(6) q2 = 2r2
from which it immediately follows that q2 is even, and so q is even. But this contradicts our
earlier claim that q is odd.

2. A Proof that the Subsets of Á are Uncountable

In what follows, we offer a proof in more or less standard mathematical form:

By way of constructing a reductio ad absurdum argument, we suppose to the contrary that Ó(Á)
is countable, in which case Ó(Á) is either finite or denumerable. It is fairly easy to dismiss the
possibility that Ó(Á) is finite (exercise!), so we concentrate on the second possibility, that Ó(Á)
is denumerable. It follows that there is a one-to-one correspondence between Ó(Á) and Á.
Define f so that:
f(n) = the subset of Á uniquely paired with n according to the alleged correspondence
Call a number n "proper" if it is not an element of its counterpart set f(n). Consider the subset s
of all proper numbers, so understood. In other words,
(1) s = { n : n ∉ f(n) }
Hardegree, Infinite Sets and Infinite Sizes page 14 of 16

By hypothesis, every subset of Á corresponds to some number; so s corresponds to some

number – call it p. Given the definition of f, we have:
(2) f(p) = s
Now, of course, either p is proper, or p is improper. In either case, we have a contradiction, as
we show below.
Suppose (case 1) that p is proper, in which case:
(3) p∈s
Combining (3) with (1) yields:
(4) p ∉ f(p)
But combining (3) with (2) yields:
(5) p ∈ f(p)
which contradicts (4).
On the other hand, suppose (case 2) that p is improper, in which case:
(6) p∉s
Combining this with (1) yields:
(7) p ∈ f(p)
And combining this with (2) yields:
(8) p∈s
which contradicts (6).

3. A Proof that the Irrational Numbers are Uncountable

The set Å of real numbers is defined as follows.

Å ü the set of rational numbers ∪ the set of irrational numbers

Earlier we defined a fraction to be a ratio of integers m/n where m<n. Let us henceforth call these
ordinary fractions. More generally, by a "fraction" we mean a non-negative real number that is strictly
less than 1.

x is a fraction ü x is a real number, and 0 ≤ x < 1

Hardegree, Infinite Sets and Infinite Sizes page 15 of 16

Some fractions are ordinary (i.e., rational) fractions; others are extraordinary (i.e., irrational) fractions.
In what follows, we argue that the set of fractions, both ordinary and extraordinary, is uncountable. We
employ a technique, originally devised by Cantor, called a ‘diagonal argument’.

Suppose, to the contrary, that the set of fractions is countable. Then it is either finite or denumerable.
The first case can be quickly dismissed, and is left as an exercise. In the second case, there is a one-to-
one correspondence between the fractions and natural numbers. Define f as follows.

f(n) = the fraction uniquely paired with n in the alleged one-to-one correspondence

Next, we appeal to the theorem that every fraction can be expressed as ("encoded by") an infinite
decimal sequence, the following being familiar examples.22

1/2 0.500000000000000…
1/3 0.333333333333333…
1/4 0.250000000000000…
2/3 0.666666666666666…
3/4 0.750000000000000…
7/22 0.318181818181818…
1/π 0.318309886183791…?
1/√2 0.707106781186547…?

Some of the sequences "terminate", which means that at some point the remaining digits are all 0’s.
When a sequence terminates with an infinite sequence of zero’s, we of course omit them in the
pronunciation – for example we say "point five", not "point five zero zero zero …", unless we are trying
to make a point about precision (about which we will later speak). Other sequences do not terminate,
but are nevertheless completely regular, in the sense that the digits eventually group into repeating units,
as in the encodings of 1/3, 2/3, and 7/22. Still others are "helter-skelter", like 1/π 1/√2, which have no
obviously discernable pattern. These, basically, are the irrational fractions.

Now suppose that we have a one-to-one correspondence between the natural numbers and the set
of real fractions, thought of as their decimal encodings. Define f so that:

f(n) = the fraction uniquely paired with n according to the alleged correspondence

Construct a decimal-encoding § as follows.

take the nth digit of f(n); call it δ;

if δ is 0, then §n = 1;
otherwise, §n = δ −1.

Notice that §n will never be 9; this constraint is included in order to avoid producing an inadmissible
encoding (one with a terminal infinite sequence of 9’s). The following is an example, where the
relevant digit in each sequence is highlighted.

22
Note that the algorithm does not automatically produce a unique encoding for each fraction. The problem is that "point
five" is identical to "point four nine nine nine …"! This can be seen by subtracting the latter from the former and noticing
that the result is zero. In our encoding, we mean to stringently exclude the goofy ones; in particular, we do not count as
admissible any encoding that has a sub-sequence of endlessly repeating 9’s.
Hardegree, Infinite Sets and Infinite Sizes page 16 of 16

number fraction resulting §

1 .412535286438453485... 3
2 .582938752320309532... 7
3 .340859403890349589... 1
4 .928347584349857348... 2
5 .839453002439549350... 4
6 .483974583945834953... 3
7 .897405723293409855... 6
8 .284634530498503934... 2
.... .... …

We next check to see whether the resulting encoding § [in this case .37124362…] is one of the items in
the list of fractions. Well, it is not the first one, since § disagrees in the first spot, and it is not the
second one, since § disagrees in the second spot, and so forth. Thus, § does not appear on the list. It
follows that there is a fraction that we have failed to count. 23

Thus, no matter how we attempt to set up a one-to-one correspondence between fractions and
natural numbers, we can always produce a fraction that has been omitted. It follows that there is no one-
to-one correspondence between Á and the fractions, and therefore the fractions are uncountable.

This proves that the fractions are uncountable. We next use this to show that the irrational
fractions are uncountable. The fractions are given as follows.

fractions = {0} ∪ ordinary fractions ∪ irrational fractions

Although we will not prove it here, we appeal to the following theorem.

if A and B are both countable, then A ∪ B is also countable.

We have already shown that the ordinary fractions are countable; it is obvious that {0} is countable; so,
if the irrational fractions are countable, then the fractions are countable. But we have just proved that
the fractions are not countable. Accordingly, the irrational fractions are not countable either.

Next, since the irrational fractions are a subset of the set of all irrational numbers, the set of all
irrational numbers is uncountable.

Finally, since the set of irrational numbers is a subset of the set of real numbers, the set of real
numbers is uncountable.

23
There is a potential problem; we might end up producing an inadmissible sequence, that is, a sequence with an infinite
repeating sequence of 9’s. But this is ruled out, since our technique generates a decimal encoding in which there are no 9’s
whatsoever.