George F. R. Ellis writes in section 2.4.2 of his outstanding survey paper, Issues in the Philosophy of Cosmology:
Clearly we cannot obtain any observational data on what is happening beyond the particle horizon; indeed we cannot even see that far because the universe was opaque before decoupling. Our view of the universe is limited by the visual horizon, comprised of the worldlines of furthest matter we can observe—namely, the matter that emitted the CBR at the time of last scattering.
The picture we obtain of the LSS by measuring the CBR from satellites such as COBE and WMAP is just a view of the matter comprising the visual horizon, viewed by us at the time in the far distant past when it decoupled from radiation.
Visual horizons do indeed exist, unless we live in a small universe, spatially closed with the closure scale so small that we can have seen right around the universe since decoupling.
The major consequence of the existence of visual horizons is that many present-day speculations about the super-horizon structure of the universe—e.g. the chaotic inflationary theory—are not observationally testable, because one can obtain no definite information whatever about what lies beyond the visual horizon. This is one of the major limits to be taken into account in our attempts to test the veracity of cosmological models.
Let’s start by defining a few of the terms that Ellis uses above.
particle horizon – the distance beyond which light has not yet had time to reach us in all the time since the Big Bang
decoupling – the time after the Big Bang when the Universe had expanded and cooled enough that it was no longer a completely ionized opaque plasma; atoms could form and photons began traveling great distances without being absorbed
worldlines – the path of a photon (or any particle or object) in 4-dimensional spacetime: its location at each and every moment in time
CBR – cosmic background radiation
LSS – last scattering surface
COBE – Cosmic Background Explorer
WMAP – Wilkinson Microwave Anisotropy Probe
(And, Planck should be added now, too)
Now the question. Do we live in a small, big, or really big universe? The best answer we can give now (or, perhaps, even in the future) is that we live in a really big universe, though it is unlikely to be infinite. Ellis himself provides a cogent argument in section 9.3.2 of the paper referenced here that infinity, while a very useful mathematical tool, does not ever exist in physical reality. We shall investigate this topic in a future posting.
Even though general relativity shows us how matter defines the geometry of our observable universe, it tells us nothing about the topology of our universe, in other words, its global properties. Do we live in a wrap-around universe where if we set off in one direction and traveled long enough, we’d eventually return to the same point in spacetime? Is the topology of our universe finite or infinite? At the moment it appears that we are not able to observe enough of the universe to discern its topology. If that is true, we may never be able to understand what type of universe we live in. But observational cosmologists will continue to search for the imprint of topology on our visible universe at the largest scales.
References
Ellis, G. F. R. 2006, Issues in the Philosophy of Cosmology, Philosophy of Physics (Handbook of the Philosophy of Science), Ed. J. Butterfield and J. Earman (Elsevier, 2006), 1183-1285.
[http://arxiv.org/abs/astro-ph/0602280]
Liddle, A.R. 2015, An Introduction to Modern Cosmology, 3rd ed., Wiley, ISBN: 978-1-118-50214-3.