oldman said:
I don't think anyone can.
For historical reasons it was assumed, shortly after the isotropic red-shift of the Hubble law was discovered, that it is caused by the isotropic recession from us of remote galaxies. It is of course impossible to check whether this is true by direct measurement, say with rulers or radar.
Instead, cosmologists soon developed a model, based on general relativity, of how the universe can be described as "expanding". This model has proved consistent with many observations, especially with the striking discovery of the cosmic microwave background in the 1960's. The assumption of expansion on which the model is based is now accepted as a working hypothesis (or maybe a stronger truth) by nearly all professional cosmologists.
...
I think your picture of history, because of an inverted narrative order, does not do justice to modern cosmology.
It is really important, REALLY IMPORTANT, in Science that a theory makes predictions of a testable effect BEFORE it is observed. That's what the whole current discussion about "falsifiability" is about. Theories have to risk refutation by making predictions BEFORE the experiment or the observation.
Predictions after the fact ("post-dictions") contribute little credibility, by comparison. They could be just fudging around to fit known data.
That is why I am bothered by your account, which seems to INVERT THE HISTORICAL ORDER.
Gen Rel was 1915, at a time when Einstein thought the universe was observed to be static, and it bothered him that his theory seemed to favor dynamic large-scale change. But they went ahead and tested. GR predicted lightbending and in 1919 they checked that.
People like Slipher observed galaxy redshift before Hubble as early as 1917, but they did not formulate a "Hubble Law" distance-redshift relation.
http://en.wikipedia.org/wiki/Edwin_Hubble
Hubble formulated Hubble Law in 1929. He presented it as an EMPIRICAL LAW that he got by plotting observations---something without theoretical basis. But in fact it was a check of General Relativity. This is something that Gen Rel had, in effect, BEEN PREDICTING ALL ALONG----as one of several possibilities for expanding or contracting dynamic universe---although not everybody realized it.
that is not the whole story----Friedmann and LeMaitre had each worked out the expanding universe consequence of Gen Rel before it was clear from the observations that that was what actually happening.
In rough outline Gen Rel (which says dynamic distances) was 1915 and the empirical Hubble Law was 1929. that is a 14 year lead. One can mention Slipher and Keeler and redshifts observed in 1917, but not many people knew about that or saw a pattern, so basically there's 14 years.
==================
So oldman I would fault your brief historical sketch here because of a certain NUANCE it has by telling the story backwards. It makes it seem as if they FIRST observed Hubble-Law expansion and THEN cooked up a model to fit the observations.
One reason Gen Rel has prestige is that (not only many of its predictions have been exquisitely accurate but also) IT PREDICTED THINGS PEOPLE WEREN'T EVEN THINKING ABOUT OR EXPECTING.
Like the Microwave Background was predicted theoretically actually back in 1948! And not observed until 1965.
As you probably know the 1948 prediction of a CMB around 5 K was published by Gamow, Alpher, Herman. (they came darn close to the right temperature)
=========================
It has been prophetic to an uncanny degree.
And the theory is a coherent organic whole, so it is not so easy to construct an alternative that somehow makes all the OTHER predictions (which involve dynamic geometry) and yet is rigid in largescale distance.
I don't know of anyone doing it. (and it would be after-the-fact POSTdiction, in any case---contriving to fit past data---so less convincing)
==========================
Don't misunderstand me. I feel sure that Gen Rel is WRONG. I expect it to be replaced by a theory that has some quantum corrections. This will, in turn, make some remarkable predictions which people will test and test---until they think they've got the right quantum-corrected version.
but I have no reason to expect that the new improved models will have static distance. Why should they? I see no evidence of static largescale distance in nature, so it is not the sort of feature I would expect a new theory of geometry to have!
I expect OBJECTS, like pieces of metal, to have fairly stable dimensions.
I don't expect largescale DISTANCES in empty space to be stable. I see no convincing evidence of that in nature. My experience does not teach me to expect it. It would surprise me if a theory was concocted that had static geometry and my first reaction would be skeptical. I would anticipate that observational tests (say of light bending) would refute such a theory.
============================
more random thoughts on the subject:
Gen Rel is our prevailing theory of geometry and says, among other things, that distances between stationary objects should normally be constantly changing: Spatial geometry evolves under the influence of matter.
Gen Rel is one coherent theory that makes testable predictions in a lot of different arenas------in solar system (light bending, GPS clock adjustment, Mercury orbit...)-----outside solar system (binary pulsar decay, expansion or contraction of distances). And it is highly successful.
I don't know of a way to buy lightbending (with the observed degree of precision) and all the other precise predictions without buying a geometry with dynamic distances along with the rest. You would have to construct a theory in which spacetime geometry is dynamic is all the ways Gen Rel says EXCEPT the variability of distance. I don't think such a variant of Gen Rel has ever been constructed. Space would have to curve the amount required to match light-bending data, for example, but distances would NOT be able to change with time. It is not obvious to me how one could do that.
