Sunday, August 01, 2010



"""Once we allow some principle of recursive reasoning, we are landed with the infinite. There are an infinite number of natural numbers: for 0 is a natural number, and if anything is a natural number, its successor is a natural number. So there are natural numbers, and for any finite number, n, it is demonstrably false that there are exactly n natural numbers. We feel impelled to allow the question `How many natural numbers are there?', and the only possible answer seems to be `an infinite number'. But we have qualms.

Infinity seems out of this earth. It smacks of Platonism, mysticism and theology. The word `infinite' is a negative concept, contrasting not only with `
finite', in a strict mathematical sense, but with
`definite' and with `comprehensible'. Often, especially in theology and ancient philosophy, the Infinite is the Whole....the Universe, the Absolute, whose logic is difficult and fraught with inconsistencies. We are wise to be wary.

Although Parmenides, Plato, Augustine, William of Alnwick, Leibniz, Cantor, Dedekind and most modern mathematicians are fairly happy with infinity, other philosophers have had doubts. Aristotle allowed the existence of the potential infinite but denied
the actual infinite, and was followed by Aquinas and most of the Schoolmen. It is fair to place Weyl and the modern Intuitionists in that tradition. Locke had considerable difficulty in articulating a coherent and satisfactory account of infinity. Berkeley was deeply
critical of infinitesimals and mathematical infinity generally, and modern Finitists likewise reject every sort of infinity, and try to confine themsleves to finite numbers alone. But infinitistic arguments
keep seeping in, and strict Finitists distinguish themselves from their laxer brethren, and still are not stringent enough in their scepticism to escape the strictures of the ultra strict."""
(in other words, Cantor resides in Hades. Yay)

It is easy to stifle doubts, and accept the infinite as part of the
mathematical exercise, justified by the general success of mathematics.
Geometry runs more smoothly if we allow points at infinity:
then we can say that two straight lines always intersect, parallel
lines meeting at a point at ifinity. If we add the symbol 1 to our

other numerical symbols, we are able to discuss series, sums and

integrals much more incisively. Once we allow the question `How

many natural numbers are there?' it is illuminating to answer @0.

It seems silly to let philosophical scruples deter us from entering

Cantor's paradise.

But credulity conduces to inconsistency. Easy acceptance is not

on. It is important, though di cult, to articulate the objections

to in nity that we are inclined to feel. What tends to happen is

that the objections are only half-formulated, and never adequately

considered; instead, the mathematician gets used to operating with

in nity, and, as it were, represses the doubts he once felt without

ever thinking them through properly and either seeing that they are

valid and accepting them or seeing that they are not valid and why.

As in psychoanalysis, repression causes trouble later on. Many Intuitionists

show symptoms of not having come to terms with in nity

in their youth, and su ering in consequence from psychological

lesions in middle age. The only prophylactic is to give full rein to

doubts at the first encounter. Doubt now, doubt bravely: dubita

fortiter to parody Luther.


Anonymous said...

I took her out once.
She was an Absolute ZERO!

(I am staying as Anonymous
because once my secret is out
I might get way-laid by the likes of Jessica Smith
or worse: One of her Clubbies

One Brow said...

Actually, we recognize a positive definition of infinity: a set is infinite if it is the same size as a proper subset. This is true whether we are using a count or a measure for size.

In particular, this definition makes it clear why the William Lane Craig arguments are such failures.

J said...

Hey OB.

I was hoping you would say something on this topic. I am opposed to Craig's arguments...yet the strictly constructivist points (ie nothing to do with theology) against Cantor's hierarchy of sets don't seem completely off base, as Lucas points out. Is mere successorship sufficient for infinity? Furthermore the idea that sets could be both greater than (set off all integers, vs set of all primes, for instance) and yet...equivalent (both infinite) seems rather...preposterous (as Lucas points out). Is that what pro-mathematicians call a bijection? It seems absurd to say that...two (or ..countless) different sized infinities exist (tho' I am not as of yet completely decided on the issue).

So I am at least intrigued by the constructivist arguments (Wittgenstein rejected Cantor)--note also Lucas's point that the universal generalization (usually a V with line, or (x)) if used sloppily certain seems absurd--e.g. to posit a universal about an infinite set (ie the set of all primes, for instance) which cannot really be verified...

Quine was not so down with Cantor or really mathematical realism.

One Brow said...

I left this question in my blog, which we could also discuss here: are locations real? If I set up some arbitrary starting point and method of determining direction, the place that is one inch north from my origin is easily and uniquely identified. Is it real? If you can identify such a place for every member of the successor set, that mean there are an infinite number of real places?

I'm more formalist than realist, mathematically.

J said...

to take a constructivist view, the point, line, equations indeed entire coordinate system is itself...posited, made to fit "the world", specific problems, projects, etc. And what are points? obviously building a bridge, or skyscraper or trench, the points are "real" in space--or we take them to be....(if you mean absolute space as per Newton or something, a rather different discussion). So in an ordinary sense--pragmatic even--there's no need for infinity...(and the little sideways 8 tends to confuse students, even in calculus--it merely means something like "no end stop to the function" ....not the universe)

Given an infinite domain, you would not be able to say what actually holds, can you. That was Lucas' point on the universals (and actually doesn't that lead to some difficult math, Goldbach conjecture, etc). I mean, people do, but logically speaking it's not established... not suggesting nominalism...but its hardly "more absurd" than the fullblown platonism of Cantor (or most academic mathematicians, really))

Custom Search

Blog Archive