Gödel on Cantor's Definition of Infinite Number

Gabe Barhydt
May 9, 2024
Introduction
In his paper What is Cantor’s Continuum Problem (1947), Kurt Gödel argues that
Cantor’s definition of infinite number was inevitable. In this paper, I reconstruct and evaluate
Gödel’s argument, and discuss the philosophical and mathematical context against which Gödel
is making this claim. I first discuss Cantor’s definition of cardinality and infinite number in order
to contextualize Gödel’s argument. I then discuss Cantor’s continuum problem, and explain how
Gödel’s argument fits into the grander scheme of this problem. I then move into a reconstruction
of Gödel’s argument itself, and evaluate the argument. Lastly, I remark on certain philosophical
points regarding intuition and infinity.
Cantor’s Definition of Cardinality
Cantor’s definition of infinite number is an extension of his theory of cardinal numbers.
His definition of equality, or inequality between sets states that the cardinality of a set A is less
than or equal to the cardinality of a set B if and only if there exists an injective function from the
elements of A to the elements of B. The sets A and B then have the same cardinal number if and
only if A is less than or equal to B, and B is less than or equal to A. This definition amounts to
the “one to one correspondence” intuition of equality between numbers, or the idea that two sets
are equal if and only if their elements can be put into one to one correspondence (in simple
terms, every element of the one set can be put into a pair with an element of the other, with no
elements leftover). Cantor refers to his theory of infinite sets as the theory of “transfinite
cardinals,” the first value of which (denoted “aleph-null”) is defined as the cardinality of the set
of natural numbers. Cantor is largely building off of Dedekind, who defined a set as infinite if
and only if it can be put into one-to-one correspondence with a proper subset of itself. This type
of definition contrasts with the “part-whole intuition,” or the theory of numerosity, which I will
expand upon later on. I will now turn to the topic of Cantor’s continuum problem, the discussion
of which motivates Gödel’s argument in favor of Cantor’s definition.
Cantor’s Continuum Problem
Cantor’s continuum problem came to light after Cantor’s discovery that the set of real
numbers is uncountably infinite. A countably infinite set is one which can be put into one-to-one
correspondence with the natural numbers, while an uncountably infinite set is one which contains
too many elements to correspond to the natural numbers, and instead requires a more rich set of
numbers to be put into one-to-one correspondence with, such as the real numbers. Gödel states
Cantor’s continuum problem simply as “How many points are there on a straight line in
Euclidean space?” More generally, Gödel is trying to show that this question is the same as the
question “what is the relationship between the cardinality of the natural numbers and that of the
real numbers, beyond simply the cardinality of the real numbers being greater than that of the
natural numbers?” Before properly answering this question however, Gödel must first address
the concern of whether the extension of the concept of “number” to infinite sets can be effected
in a “uniquely determined manner,” or if differing methods might be equally correct. He claims
that Cantor’s definition does in fact have this uniqueness, and that in a “very striking manner.” I
will now turn to Gödel’s argument that Cantor’s definition of equality is not only correct, but
“inevitable” when applied to infinite sets.
Reconstruction of Gödel’s Argument
Gödel begins by remarking that it will be an essential characteristic of whatever
definition of “infinity” we ascribe to “number” that the number of elements belonging to a class
remains the same regardless of any changes we make to the objects in the class, be it their
properties or relations to one another. He then claims that it immediately follows that Cantor’s
definition of equality between numbers is correct, namely that all the elements of the sets can be
brought into one to one correspondence. Given any pair of sets which has this one to one
correspondence, he claims it will be possible to consider a transformation of all of the properties
and relations of one of the sets into those identical to the other. Hence the two sets will be
completely indistinguishable, and have the same cardinal number. Gödel then considers an
example involving a square and a line segment filled with mass points. It follows, assuming that
there exists a one to one correspondence between the points of the line and those of the square,
that the points of the one can be rearranged so as to fill out exactly the points of the other. He
remarks that such an example could only be applied to physical objects, though this is no issue as
a definition of number which relies on numbered objects could not be satisfactory. Gödel thus
concludes that Cantor’s definition of equality between numbers must be accepted, further
remarking that it can easily be extended to infinite numbers.
Evaluation of Gödel’s Argument
Gödel’s argument rests on a single premise, that our definition of equality between sets
must be such that whatever changes we might make to the nature of the elements themselves or
to their relations, the number of elements will remain the same. This premise paints a clear and
intuitive picture when considering finite sets. When extended to infinite sets however, the
premise may appear problematic, or indeed even question-begging. To understand why, I will
now delve in the “part-whole intuition,” and the theory of numerosities. The theory of
numerosities relies on what are known as “approximating sequences.” We can consider these as
functions which pull certain elements from a given series, and can tell us the relative size of this
subset as the series approaches the limit. For example, given the set of natural numbers 0, 1, 2,
3,..., an approximating sequence which picks out all the even numbers will read 1, 1, 2, 2, 3, 3, 4,
4, etc.; 0 is the first even number, which is why we start out with one, while 1 is not an even
number, so the value remains the same, increasing by one every other number. Were we to do the
same for the multiples of three, the approximating sequence would read 0, 0, 0, 1, 1, 1, 2, etc.
When comparing the approximating sequences for the even numbers and the multiples of three,
we can immediately see that the sequence corresponding to the even numbers is increasing at a
much quicker rate. For a proponent of the theory of numerosity, this will yield the conclusion
that the set of even numbers is larger than the set of multiples of three, while a proponent of the
theory of cardinality will still claim that they are the same, as the numbers can still after all be
placed into one to one correspondence. We can now see why Gödel’s premise is perhaps question
begging; it already assumes that the part-whole intuition is false. This is because, when regarding
infinite sets, the grouping of the objects within the sets will affect the approximating sequence,
which is indeed a relational property. If the approximating sequence is translated into a graph,
the size of the infinite set will correspond to the slope as the sequence approaches infinity;
depending on the grouping, the line could converge faster or slower, yielding different answers
as to the size of the set for those who share the part-whole intuition.
Philosophical Remarks on Intuition and Infinity
It is important to note that the conflict between the part-whole intuition and the
one-to-one intuition arises only when they are used to describe the same notions of “size.”
Cantor himself even remarks on different sizes of infinity that they can be conceived in different
ways; he claims that the set of natural numbers is richer, and contains “more reality” than the set
of even numbers, though the cardinality of the even numbers is nevertheless the same as that of
the natural numbers, and hence they are of equal sizes. It does seem as if he may be too heavily
reliant upon the abstractness of the subject matter, especially when he refers to the “reality” of
the two sets, the one possessing more than the other. What could “more reality” refer to if not
simply there being more within one set than the other? Moreover, it seems as if Gödel’s claim
that Cantor’s definition is “inevitable” is too strong, as it simply will not suffice to those who do
not share the one-to-one intuition. He is trying to show Cantor’s definition is necessary and
sufficient, but has succeeded only in showing that it is sufficient. While his points are often very
intuitive, and perhaps very attractive, when dealing with a concept as abstract (and perhaps
fundamentally unknowable) as infinity, we are somewhat stuck between whether we should
follow our intuitions, or deviate from them. For the seeming unknowability of the infinite might
yield the belief that our intuitions are crucial in order to evaluate the situation, as we haven’t
much else to springboard off of. On the other hand, one may argue that, given we are far more
accustomed to the finite than the infinite, that our intuitions regarding the infinite are
fundamentally mistaken, as we are merely trying to fit the infinite into a framework which is
compatible with our understanding of the finite. I contend that, while it will be useful in
mathematical practice to take a firm stance on the situation (so as not to fall into a contradiction),
one ought to heed this sort of caution regarding their philosophical stance on the situation.
Indeed, both the part-whole intuition and the one-to-one intuition seem equally plausible in this
situation, and this is something which Gödel has plainly overlooked.
Conclusion
In this paper I have discussed Gödel’s argument that Cantor’s definition of infinite
number was inevitable. I have provided the philosophical and mathematical context against
which Gödel is making this argument, that being Cantor’s definition of cardinality and its
application to his continuum problem, as well as the tension between the opposing “part-whole”
and “one-to-one” intuitions regarding how to define infinite number. I have argued that Gödel’s
argument rests on a premise which already assumes his opposition to be false. Lastly, I have
remarked that one ought to be cautious of their reliance on intuition when dealing with a concept
as abstract and unknown as infinity.
 References
Godel, K. (1947). What is Cantor’s Continuum Problem? The American Mathematical Monthly,
176-177