Originally posted by twhitehead
They sound like physicists / philosophers not mathematicians.
[b]Other mathematicians will happily accept that a number can be infinite.
But not a Real number.
Most have a position in between. The difference is one of mathematical philosophy.
Nevertheless, the question in the OP, is not about whether infinite numbers can exist, it is ...[text shortened]... but instead say 'the observable universe' as that caters for both finite and infinite universes.[/b]
Look up L.E.J. Brouwer who amongst other things invented the fixed point theorem.
You've missed what I was getting at. When we make claims about the world it is either a claim generated by the theory or an experimental result. To make
where I'm making the distinction I'm making more clear, consider the discovery of the Higgs boson at LHC. What they actually saw were behaviours in their detectors. Their claim is along the lines of: "Our detectors exhibited a behaviour consistent with two photons with a combined energy of 125GeV a statistically significant number of times. We therefore conclude that a particle with the right properties to be the Higgs boson exists and has a mass of 125GeV.". The empirical part of that statement is: "Our detectors exhibited a behaviour.",
the whole of the rest of the statement is theoretical. That what they saw in their detectors corresponds to two photons is a theoretical claim (a very well justified theoretical claim, but nevertheless theoretical). That the theory being
verified is being compared with a null claim means that other theories which have scalars about the right mass and decay path, but not part of the Higgs mechanism are not ruled out.
Any answer to the question in the OP cannot be straightforwardly empirical. According to what we think our best theory is says, it's impossible for us to see everything that is a finite distance away, never mind verify the existence of things infinitely far away. So we have to look at the theories to see if they allow it. The theories have empirical justification, but that doesn't stop them being theories. Most of the time physicists implicitly assume that the quantities they are using are correctly described by the real numbers. Consider distances, we assume that the correct way to denote distances is to use the reals. For almost all applications it doesn't matter that it might not correctly represent distances at a fundamental level. So the question of whether the reals are the right set of numbers to use in fundamental physics is an open one.
That General Relativity uses the real numbers doesn't mean that it could not use the surreals, it doesn't make much difference at the level GR makes its predictions. This might matter for Quantum Gravity theories, but the discussion isn't about them. So, if we can use some number system that allows infinite distances in General Relativity then the answer to the OP would be "Yes, the theory allows objects infinitely far apart.". I don't think empirical evidence helps particularly, all we can say is that this region of the universe (the observable to us part) appears to be expanding at an accelerating rate. That doesn't really rule out any of the options. If it's possible to give a definitive answer it hangs on what numbers correctly describe things like distance and we can't be sure of the answer to that, but that hinges on whether things like surreal numbers are allowed in mathematics. If mathematics does not allow such structures it's hard to see them as an allowable object in a physics theory.
I didn't read all the posts where you and humy were discussing set theory. What I did notice is that you seem to have missed out the notion of measure. This is an additional property you impose on a set. On Wikipedia the place to start is probably the page on Lebesgue Integrals. Roughly, measure is a non-negative number assigned to a set, the interval [0, 1] has measure 1. There are a collection of axioms (basically to make it additive in the right way). Let m(S1) be the measure of the set of points S1 and S2 also a set with measure m(S2) and, S1 U S2 is the union and S1*S2 is the intersection of S1 and S2 then:
m(S1 U S2) + m(S1 * S2) = m(S1) + m(S2).
So, you are looking for a subset of the reals with infinite measure. Let's make a cut on the real line and exclude the point at the origin. This divides the set of reals into two subsets of infinite measure. Now, because we are using the reals then there is no point on the real line where we can make a cut to generate three subsets where one of them does not have finite measure (the bit between the origin and where we made our second cut). But, that the reals do not have this property does not mean that some space could not be constructed so that subsets with infinite measure can be found which are adequately continuous (see Banach-Tarski paradox for an example of what I'm trying to rule out with the words "adequately continuous".), think of the distance along a fractal curve for a simple example. Questions like this tend to require one to worry about things like the Axiom of Choice and how you've gone about constructing your sets, and I think you may have made implicit assumptions about the sets you are looking at which mean that your conclusions aren't as general as you think they are.