Anda di halaman 1dari 21

The Cardinality of the Real Numbers

David Corne

852038

1
Contents

1 Introduction 3

2 A Short Biography 5
2.1 Georg Cantor (1845-1918) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Ernst Zermelo (1871-1953) and Abraham Fraenkel (1891-1965) . . . . . . . . 5
2.3 Kurt Gödel (1906-1978) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4 Paul Cohen (1934-2007) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

3 An Infinity of Infinities 7
3.1 Theorem: R is Uncountable . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2 Cantor’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

4 The Continuum Hypothesis 9


4.1 The Continuum Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

5 Zermelo Fraenkel Set Theory and the Axiom of Choice 11


5.1 Axioms of ZF Set Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
5.2 The Axiom of Choice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

6 Gödel’s Consistency 14

7 Cohen’s Independence and Forcing 16


7.1 Forcing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

8 Conclusion 18
8.1 Martin’s Axiom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
8.2 Martin’s Maximum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
8.3 The Answer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

9 References 20

2
1 Introduction

I will begin with a discussion of Cardinality as this will be a key concept throughout this
project.

Definition: Cardinality
The cardinal number associated with a given class

Definition: Cardinal Numbers


A measure of size of a set not accounting for the size of its members

[1, p.65]

For example the set {1, 2, 3, 4} and the set {N, Z, Q, R} both have cardinality 4 as they both
have 4 elements. Although this is an easy case these two sets have the same cardinality as
they can be put in a one to one correspondence with each other. [2, p.1-2] This is the same
as saying that there exists a bijective function between them. The example given above has
the following bijection:

f : {1, 2, 3, 4} 
→ {N, Z, Q, R} s.t
N x=1








Z x = 2

f (x) =




 Q x=3



R x = 4

This notion of a one to one correspondence is crucial when one starts to deal with infinite
sets, as for many cases intuition is wrong. One such example of this are the integers (Z) and
the even integers (2Z). General intuition is that Z should have twice as many elements as
2Z and hence double the cardinality. However the function

3
f : Z −→ 2Z

x 7−→ 2x

obviously maps from Z to 2Z and is a bijective function. Therefore the cardinality of Z is


equal to the cardinality of 2Z meaning that there are as many even integers as there are
integers.
The absolute value symbols are used to represent the cardinality of a set, for example,
if A = {1, 2, 3} then |A| = 3
Many mathematicians also use the notation c to represent the cardinality of the real numbers.
c stands for the continuum, which is another word for the real line.

4
2 A Short Biography

2.1 Georg Cantor (1845-1918)

Georg Cantor is often called the founder of set theory, due to his development of naive set the-
ory. Cantor worked on one to one correspondences, countability and uncountability and also
transfinite numbers. Cantor’s work often conflicted with the views of other mathematicians
at the time, particularly Kronecker. Cantor’s principals did, however, have their supporters,
such as David Hilbert who famously wrote, “No one shall expel us from the Paradise that
Cantor has created”. The most famous result of Cantor’s is probably the formulation of the
continuum hypothesis, which is a main subject of this project.

[3, p.778],[4]

2.2 Ernst Zermelo (1871-1953) and Abraham Fraenkel (1891-1965)

In 1908 Ernst Zermelo published the first axiomatisation of set theory. However, he did not
consider the axioms of foundation or replacement. These were formally added by Abraham
Fraenkel and Thoralf Skolem in 1922. This set of axioms then described what is known as
Zermelo-Fraenkel set theory, which can also be combined with the axiom of choice. Zermelo-
Fraenkel set theory is still seen as the standard model of set theory that is used today.

[3, p.619],[5, p.213]

2.3 Kurt Gödel (1906-1978)

Kurt Gödel is widely agreed to be the greatest logician of the nineteen hundreds. Gödels
major contribution to mathematics were his first and second incompleteness theorems. In
the context of this project I will be studying his proof of the consistency of the continuum
hypothesis with respect to the axioms of Zermelo-Fraenkel set theory.

5
[3, p.619],[6]

2.4 Paul Cohen (1934-2007)

Paul Cohen published two papers in 1963 proving the independence of the continuum hy-
pothesis to the axioms of Zermelo-Fraenkel set theory. In these papers he used and developed
the technique of forcing. Forcing remains an important mathematical tool in set theory to
this day.

[3, p.624-625]

6
3 An Infinity of Infinities

Georg Cantor published his paper Über eine eigenschaft des inbegriffes aller reellen alge-
braischen zahlen (translation: On a characteristic property of all real algebraic numbers) in
1874. This was published in ”Journal fur die reine und angewandte mathematik”, which is
commonly known as Crelle’s journal. In this paper, Cantor proved the uncountability of the
real numbers and the countability of the real algebraic numbers. This important paper is
seen by many as the birth of set theory. To prove the uncountability of the reals he initially
used a proof relying on an infinite sequence of nested intervals. He then took their limit
point using the completeness of the Reals.[7, p.55]
In 1878 Cantor’s next paper, published again in Crelle’s Journal, then defined precisely the
concept of a one to one mapping between sets to determine their size.
It was in 1891 that Cantor published his famous diagonalisation argument for the uncount-
ability of the real numbers.

[4]

3.1 Theorem: R is Uncountable

Proof: Diagonalisation Argument


It is sufficient to show that (0,1) is uncountable in order to demonstrate that R is uncountable.
This is because any subset of a countable set is countable, so showing that the Reals have
an uncountable subset shows that they are uncountable.
The decimal expansion of real numbers needs to be used in order to prove this. Decimal
expansions are, however, non-unique.
e.g. 0.29999 . . . = 0.30000 . . .
Therefore, any decimal ending in a non-terminating string of nines shall be taken as the
equivalent value ending in zeros.
Suppose (0,1) is countable
let the elements of (0,1) be listed as a1 , a2 , a3 , . . . and let A = {a1 , a2 , a3 , . . .}

7
It can now be demonstrated that there is an element b ∈ (0,1) missing from this list. To do
so a1 , a2 , a3 , . . . each need to be written as a decimal expansion. Construct b as follows; for
the nth digit of b take an and if its nth digit is 0 let the nth digit of b be 1 and if its nth digit
is not 0 let the nth digit of b be 0. Constructing b as such makes it differ from each of the
numbers in A by at least one digit.
We have reached a contradiction as b 6∈ {a1 , a2 , a3 , . . .}
Therefore (0,1) is uncountable and as such R is uncountable. 

[5, p.24-25]

This proof implies that there exist sets of cardinality greater than that of the natural
numbers i.e. that there are uncountable sets and R is such a set. Following on from this
idea it will be seen that for all cardinal numbers there exists a greater one.

3.2 Cantor’s Theorem

For X a set and P(X) the power set of X (the set of all subsets of X) then
|X| < |P(X)|
The proof for Cantor’s theorem involves using the function f : X → P(X) such that f (x) =
{x} This is injective as |X| ≤ |P(X)| It then remains to show that |X| =
6 |P(X)| which can
be done by taking an arbitrary function f : X → P(X) and showing it is not onto by taking
the set Y = {x ∈ X : x 6∈ f (x)} which is an element of P(X) but is not in the range of f.

[5, p.65]

Cantor’s Theorem asserts that for all sets their power set will be of greater cardinality. The

cardinality of the power set is defined as |P(X)| = 2X for every set X. It is easy to show

by induction that |P(X)| = 2X for finite X and it is defined as such for non-finite sets X.
This notion implies that there is no greatest cardinal number and that there are infinitely
many of them.

8
4 The Continuum Hypothesis

The continuum hypothesis, or the continuum problem, as it is sometimes known, was first
formulated in 1878.[3, p.623] It was conjectured by Cantor and was the first of the Hilbert
problems. David Hilbert famously introduced these problems in Paris, at the Second Inter-
national congress of Mathematicians in 1900. They comprised a list of twenty-three unsolved
problems he deemed to be the most important of the time.

[3, p.789]

4.1 The Continuum Hypothesis

The Continuum Hypothesis states that every infinite subset of the real numbers is either
countable or has the same cardinality as R. If Zermelo Fraenkel set theory with the axiom
of choice is used, this is equivalent to saying that the cardinality of R is ℵ1 . This is also
equivalent to saying that ℵ1 = 2ℵ0 . I will discuss Zermelo Fraenkel set theory with the axiom
of choice later, as opposed to naive set theory.

[3, p.623],[8, p.102]

It is evident that there are many ways of stating the Continuum Hypothesis, but I believe
that the simplest interpretation is that there is no set whose cardinality is strictly between
that of the integers and that of the real numbers.
This hypothesis is obviously of great importance to the problem of finding the cardinality
of the real numbers, as one of the statements of the continuum hypothesis is that c = ℵ1 .
Therefore, I will now look at the problem of proving or disproving the continuum hypothesis.
There is also the Generalized Continuum Hypothesis which is stated
2ℵα = ℵα+1 for any α

[9, p.65]

Cantor published his “Grundlagen einer allgemeinen Mannigfaltigkeitslehre” (Foundations


of a general theory of manifolds) which is often referred to simply as the Grundlagen. In

9
the Grundlagen he had still not managed to prove the continuum hypothesis but still hoped
that a proof would be arrived at soon. However, as this project will demonstrate, such a
proof is not possible.[4, p.101-102]

10
5 Zermelo Fraenkel Set Theory and the Axiom of

Choice

Thusfar, only naive set theory as developed by Cantor has been utilised. For the necessary
mathematical tools and rigor to investigate the continuum hypothesis, discussion of axioma-
tised set theory is required. In 1893, Frege published an axiom schema of comprehension
which gives an intuitive idea of a set it states:
If ϕ is a property then there exists a set Y = {X : ϕ(X)} which contains all elements with
property ϕ.
This was shown to be false due to the well-known Russell’s paradox. This paradox states
that there is no set S = {X : X ∈
/ X}. Therefore the schema of comprehension fails.

[5, p.7]

For this reason mathematicians looked beyond naive set theory and developed axiomatic set
theory. The standard form is Zermelo Fraenkel Set Theory which is often abbreviated to ZF
Set Theory. The axioms of ZF Set Theory are as follows

5.1 Axioms of ZF Set Theory

Axiom 0: Empty Set* There exists the empty set ∅.

Axiom 1: Extensionality If x and y have the same elements then x is equal to y.

Axiom 2: Schema of Separation For every formula ϕ(s, t) with free variables s and t, for every
x and for every parameter p there exists a set y = {u ∈ x : ϕ(u, p)} that contains all
u ∈ x that have the property ϕ .

Axiom 3: Pairing* For any a and b there exist a set x containing precisely a and b.

Axiom 4: Union*For every family F there exists a set U = ∪F that is the union of all subsets
of F.

11
Axiom 5: Power Set* For every set x there exists a set P(x) the power set of x that is the set
of all subsets of x.

Axiom 6: Infinity There exists an infinite set.

Axiom 7: Replacement Scheme For every formula ϕ(s, t, U, w) with free variables s, t, U, and
w every set A, and every parameter p if ϕ(s, t, A, p) describes a function f on A
by f (x) = y ⇔ ϕ(s, t, A, p), then there exists a set Y which contains the range,
f [A] = {f (x) : x ∈ A}, of the function f .

Axiom 8: Foundation Every non-empty set x contains a set y that is disjoint from x.

[5, p.211-213]

These are the Zermelo-Fraenkel set theory axioms. The axioms marked with a * are a
stronger version of the basic axioms obtained using the Schema of Separation. Not included
in these axioms is the controversial axiom of choice.

5.2 The Axiom of Choice

The axiom of choice states:

For every family F of disjoint empty sets there exists a selector, the set S. This
means that S intersects every x ∈ F in precisely one point.

[5, p.212]

Zermelo Fraenkel Set Theory including the axiom of choice is often abbreviated as ZFC. The
axiom of choice was controversial, as it is very non-constructive. Many mathematicians at
the time rejected it, for example Lebesgue and Baire. It is mathematically equivalent to
many other results, such as:

(i) The Well Ordering Principal

(ii) Cardinal Comparability

12
(iii) Zorns Lemma

(iv) Every vectorspace has a basis

(v) Tychonoff’s Theorem

“The Axiom of Choice is obviously true, the Well Ordering Principal is obviously
false, and who can tell about Zorn’s Lemma.” (Jerry Bona)

[10, p.6]

The axiom of choice also leads to counterintuitive results such as the Banach-Tarski Para-
dox. This ‘Paradox’ has been rigorously proved and shows that the unit sphere in a three
dimensional space (U = {(x, y, z) ∈ R3 : x2 + y 2 + z 2 ≤ 1}) can be split into finitely many
pieces and reformed rigidly into two separate copies of U .

13
6 Gödel’s Consistency

Cantor and many other mathematicians tried to verify the continuum hypothesis, but no-
body could find a logical proof. In 1938 Gödel proved that both the generalized continuum
hypothesis and the axiom of choice are consistent with the ZF axioms. For this proof he
assumed that the axioms of ZF are consistent. With this assumption Gödel constructed a
model of ZF calling it T . He subsequently found a way to construct another model of ZF
within T which he called L, where the continuum hypothesis holds. This model L is often
known as the constructible universe. An immediate corollary of this proof was that the ax-
iom of choice is also consistent. This is because Gödel did not use the axiom of choice in this
proof and so it is not required to hold in T . However once L is constructed it can be shown
that the axiom of choice holds in L. This shows that both the continuum hypothesis and
the axiom of choice are consistent with the axioms of ZF and thus neither can be disproved
within them.

[3, p.623-624,819] [6]

By 1940, when this proof was published in “The consistency of the continuum hypothesis”
by Gödel, the notion of proving the consistency of a result was quite new. Only nine years
earlier his second incompleteness theorem stated that:
Any system of axioms T with a language L if:

1. T is consistent

2. It is possible to and there is an effective algorithm to decide for any sentence in L


whether it is in T or not

3. T is complex enough to contain addition within the integers

Then it is impossible to prove within T that T is consistent.

[5, p.4-5]

14
An immediate corollary of this is that the ZF set theory axioms could not be proved consis-
tent. The proof that the generalized continuum hypothesis is consistent, as already estab-
lished, relies on this assumption.
Gödel wrote an article discussing the continuum hypothesis and his views on it. This
article was published in The American Mathematical Monthly in November 1947. In this
he gives two arguments as to why the continuum hypothesis should be false. He then
provides four more examples of paradoxical results which can be deduced from the continuum
hypothesis. Gödel’s final remark in this paper is:

“it is very suspicious that, as against numerous plausible propositions which


imply the negation of the continuum hypothesis, not one plausible proposition
is known which would imply the continuum hypothesis. Therefore one may on
good reason suspect that the role of the continuum problem in set theory will be
this, that it will finally lead to the discovery of new axioms which will make it
possible to disprove Cantor’s conjecture.”

[11, p.523-524]

This emphasises Gödel’s belief that not only is the continuum hypothesis hypothesis false,
but that its lasting effect on mathematics will prompt the discovery of new set theory axioms
and systems. It illustrates that he thought his proof of its consistency would not lead to its
proof, but would mean that it is undecidable and so can neither be proved nor disproved in
ZFC.

15
7 Cohen’s Independence and Forcing

It turns out that Gödel was proved correct, as in 1963 Paul Cohen published two papers in
the National Academy of Sciences called “The independence of the continuum hypothesis”.
The first result he gives is as follows:

“Theorem 1
There are models of ZF in which the following occur: ... (3) The Axiom of Choice
holds, but ℵ1 6= 2ℵ0 ”

[12, p.1143]

In these two papers, he proved the independence of the continuum hypothesis by inventing
a revolutionary new mathematical technique. Along with the proof that the continuum
hypothesis is consistent provided by Gödel, this proves that the continuum hypothesis is
undecidable in ZFC set theory.

7.1 Forcing

Forcing is a powerful tool for constructing models of ZFC. It is used primarily to construct
models of ZFC, in which a particular result can be proved consistent. It can also be used to
prove independence of results. The overall method to do this is as follows:

Define a sentence φ as a formula without free variables. Then the technique of forcing
allows you to prove that φ is independent. In order to do this ZFC+“φ” and ZFC+“¬φ”
must both be proved to be consistent. Proving their consistency is done using the forcing
principle.

[5, p.167-169]

The development of forcing has precipitated much new research. Results proved by this
technique do not just concern set theory, but have had important implications in fields such
as topology, combinatorics and analysis.

16
[3, p.627]

For the result of the independence of the continuum hypothesis, Cohen received the Fields
Medal in 1966. To this day it remains the only Fields Medal to be awarded for logic.

17
8 Conclusion

As demonstrated in the previous chapters, the continuum hypothesis is both consistent with,
and independent of, the axioms of Zermelo Fraenkel Set Theory, even including the axiom
of choice. This shows that Cantor’s conjecture about the cardinality of the real numbers, i.e
that c = ℵ1 , can neither be proved nor disproved. This does not, however, mean that the
continuum hypothesis is no longer of interest, or even that it has stopped being researched.
The fact that it is undecidable in ZFC set theory simply means that mathematicians should
be looking for new models of set theory and new sets of axioms. Results are still being found
and proved in these advanced levels of set theory, such as forcing axioms and Martin’s axiom

8.1 Martin’s Axiom

“In every compact Hausdorff topological space satisfying the countable chain
condition, the intersection of ℵ1 dense open sets is non-empty”

[3, p.633]

If this axiom is accepted it shows that the continuum hypothesis is false, as if there are ℵ1
T
real numbers then {R \ {r} : r ∈ R} = ∅. This does not, however, resolve the question of
the cardinality of the real numbers. Although the system ZFC + Martin’s Axiom + 2ℵ0 = ℵ2
is consistent. [3, p.633],[13, p.611]

Some mathematicians, such as W. Hugh Woodin, now believe that there will be a solu-
tion to the continuum problem and that it will not remain undecidable. Woodin wrote a
paper entitled “The continuum hypothesis” in two parts, publishing it in the notices of the
American Mathematical Society in 2001. In these papers he discusses building new models
of ZFC set theory and the axioms of projective determinancy. His main result in this paper,
however, is a theorem stating that if certain technical conditions hold, then the continuum
hypothesis is false. In the conclusion of his paper he writes:

18
“So, is the Continuum Hypothesis solvable? Perhaps I am not completely con-
fident the solution I have sketched is the solution but it is for me convincing
evidence that there is a solution. Thus, I now believe the Continuum Hypothesis
is solvable,”

[14],[15]

8.2 Martin’s Maximum

Another axiom called Martin’s Maximum is the provably strongest form of Martin’s axiom.
This axiom has some very far-reaching and surprising implications. Martin’s maximum
was published in 1988 by Foreman, Magidor and Shelah in a joint paper titled ‘Martin’s
Maximum, Saturated Ideals, and Non-Regular Ultrafilters’. Included in these are |R| = c =
2ℵ0 = ℵ2 .

[16, p.1,17]

8.3 The Answer

In conclusion, therefore, the answer to this project’s question, “what is the cardinality of
the real numbers?”, is as follows. With the current Zermelo Fraenkel set theory axioms
including the axiom of choice, the answer is undecidable, a fact that most mathematicians
accept. However, if Martin’s axiom and Martin’s maximum are accepted, it can be proved
that the cardinality of the real numbers is ℵ2 .

19
9 References

1. E.J. Borowski and J.M. Borwein, Collins Dictionary of Mathematics, 2nd Ed. (Glas-
gow: HarperCollins Publishers, 2002)

2. M. Hallett, Cantorian Set Theory and Limitation of Size (Oxford: Oxford University
Press, 1984)

3. T. Gowers (ed.), The Princeton Companion To Mathematics (Princeton: Princeton


University Press, 2008)

4. J.W. Dauben, Georg Cantor: His Mathematics and Philosophy of the Infinite (Cam-
bridge: Harvard University Press, 1979)

5. K. Ciesielski, Set Theory for the Working Mathematician (Cambridge: Cambridge


University Press, 1997)

6. K, Gödel, “The consistency of the axiom of choice and of the generalized continuum
hypothesis”, Proceedings of the National Academy of Sciences of the United States of
America, Vol. 24, No. 12, p.556 (1938)

7. P.E. Johnson, “The genesis and development of set theory”, The Two-Year College
Mathematics Journal, Vol. 3, No. 1, p.55 (1972)

8. P.R Halmos, Naive Set Theory (Princeton: D. Van Nostrand Company, 1960)

9. F.R. Drake, Set Theory An Introduction To Large Cardinals (Amsterdam: North-


Holland Publishing Company, 1974)

10. C. Good, General topology, University of Birmingham, Birmingham (2009)

11. K, Gödel, “What is Cantor’s continuum problem”, The American Mathematical Monthly,
Vol. 54, No. 9, p.515 (1947)

20
12. P.J. Cohen, “The independence of the continuum hypothesis”, Proceedings of the Na-
tional Academy of Sciences of the United States of America, Vol. 50, No. 6, p.1143
(1963)

13. J. R. Shoenfield,“Martin’s axiom”, The American Mathematical Monthly, Vol. 82, No.
6, p.610 (1975)

14. H. Woodin,“The continuum hypothesis. part I.”, Notices of the American Mathematics
Society, Vol. 48, No. 6, p.567 (2001)

15. H. Woodin,“The continuum hypothesis. part II.”, Notices of the American Mathemat-
ics Society, Vol. 48, No. 7, p.681 (2001)

16. M. Foreman, M. Magidor and S. Shelah, “Martin’s Maximum, Saturated Ideals, and
Non-Regular Ultrafilters. Part I”, The Annals of Mathematics, Vol. 127, No. 1, p.1
(1988)

21

Anda mungkin juga menyukai