Is there any "Cosmological Principle"? The term "Cosmological Principle" is widely invoked in cosmology discussions. I
think the basic idea here is that one might guess that the Milky Way is not at a special location in the cosmos as a whole, and one might guess further that each place in the cosmos is much like any other place. This guess is then elevated to the status of
"Principle". However it is obvious that some places are unlike others (the centre of a neutron star is unlike a galactic void) so this "principle" is only being
invoked as a statement about averages at large scales. So really it is a "principle"
which says that after all the stars, galaxies and galaxy-clusters one will eventually
arrive at a scale where things are homogeneous on average, and this scale is a good
deal smaller than the size of observable universe.
Now it happens that the observable universe
is indeed like that (smooth to good approximation on the large scale). However, as
far as I am aware there is no principle which says it had to be like that. It is not a principle but an empirical observation. Before looking, we have no a priori justification for expecting homogeneity. The universe might have been structured like a tree, with structure at all scales above some minimum. Indeed, in view of all the other scales of structure, one might guess that it would be quite likely to come out like that.
If the tree-like structure were the case, then one would have the issue whether or not the Milky Way should be located at a special place in such a tree. It is obvious that the answer is yes, up to a point, because it has to be at a place where a galaxy can form.
So you see I begin to doubt that there is any "cosmological principle" in the sense of a principle of logic or of theoretical physics. It is a misnomer for an interesting and important empirical observation.
My question is, then: is there any argument coming either from logic or from physics more generally (not observations such as sky surveys and CMB measurements) which says the cosmos must be expected to be homogeneous? And if it were not homogeneous then is there any argument deserving the name "principle" whose conclusion is that the physical conditions allowing life might be equally well expected at one place as another?
 A: I think this is very close to a discussion on definitions rather than physics, but I will give it a shot.
First we could argue it was a principle for historical reasons. And the terminology has stuck, so people still call it 'Principle' today.
On the pragmatic point of view it is the simplest possible scenario computation-wise and speaks directly about the simplest approximation you can make if you want to model the universe. And was used in the scientific community way before actual large scale observations.
Now let me argue from the theory side. There is in my opinion two important points. One is Naturality and the other is Symmetry. If you accept to an extent that a universe starting without preferred direction, location or very specific tuning of some quantity is more appealing that one violating any of the aforementioned, you again conclude an isotropic and homogeneous fluid is the simplest thing you can use (which is exactly the Cosmological Principle). We actually observe the laws of physics do not prefer a direction here on earth and observations of different locations seem to indicate location is also not important for fundamental laws, so we extrapolate... Let us say that happens everywhere.  We call this zero-th order approximation a "Principle" again due to custom.
Can we check every possible location? No, thus we need to extrapolate. Is it an approximation, yes we know. Have we gone beyond it? Yes we have and perturbations around such are extremely relevant to the CMB, structure formation, and in general the history of the universe.
Will it we be still called a principle...? Probably yes I am afraid
A: Not only is the cosmological principle not a logical requirement, but it is actually violated in many cosmological models such as eternal inflation.
However, we only have access to one Universe, so there's a limit on the largest scales we can probe observationally; beyond that you start getting into shouting matches about metaphysical things like  how to put a probability measure on an infinitely large space or what counts as science. The cosmological principle seems to work for scales between 100 Mpc and our cosmological horizon.
There are some proposals that we actually do find ourselves in an underdensity of the cosmological fluid, and perhaps this explains the tension between measurements of the Hubble constant with "local" measures like supernova and "distant" measures like the CMB. However as far as I understand, these models would require us to lie very close to the center of the underdensity (if we do lie in one), and so are fine-tuned.
A: There's no argument from fundamental physics or logic that the universe must be homogeneous. My basis for making this statement is argumentum ex silentio: if there is such an argument, then it would have been mentioned in this paper, and it isn't.
The name "principle" is sort of a misnomer, because it's not grounded in fundamental physics, and more reflects what we think ought to be true. The principle of relativity (i.e. the idea that the laws of physics must be the same for all observers) is another similarly-named principle. It would be intuitively really weird if the principle of relativity is false, but it is still a falsifiable statement, and we believe it because of empirical justification.
Similarly, there's already another "principle" that was proven incorrect in cosmology - the perfect cosmological principle, so-named because it says that we are not located in a special position in the universe and we are not located in a special time in the universe. It's intuitively appealing, but it was proven incorrect by the observations that led to Big Bang theory.
A: 
it happens that the observable universe is indeed like that... there is no principle which says it had to be like that. It is not a principle but an empirical observation.

Right! Physics explains the world with the simplest performant models. Our place is typical until proven otherwise. We don't need to justify its being typical to an armchair philosopher; we need only have explanations that fit the observations so far. You can learn very little without them.

is there any argument coming either from logic or from physics more generally (not observations such as sky surveys and CMB measurements) which says the cosmos must be expected to be homogeneous?

It's interesting you define physics to exclude the observations that test the mettle of its ideas. Ultimately, physicists tried various full-Universe solutions to general relativity, and realized the only tractable ones data didn't quickly refute were homogeneous and isotropic but not static. Ever since then, additional data has largely gone along with these ideas. The tree idea has been less successful.
I'll take "logic" to include philosophical analyses that are plausible but potentially open to empirical refutation, if only because tautologies won't tell us how the Universe looks. There are those who will argue the Copernican principle is a reasonable prima facie starting point (what else would be?), and that it motivates the cosmological principle.

And if it were not homogeneous then is there any argument deserving the name "principle" whose conclusion is that the physical conditions allowing life might be equally well expected at one place as another?

I won't try to decide what should be called a principle, but I can speak to whether inhomogeneity would make life more feasible in some places, provided it's a narrow notion of life. Certain kinds of star, and certain kinds of elemental distributions on planets etc., occur more readily with specific matter distributions, and are crucial to life as we know it. We could be here all day trying to give examples, every single one of which would be very controversial. To sign off with one, it's been conjectured that the crucial role evolution gave phosphorus in Earth's life, despite it being of low abundance, suggests (i) life needs it and (ii) other parts of the galaxy where phosphorus is rarer will have less life if any.
A: As far as I can tell, the use of "principle" in physics is largely interchangeable with "postulate". E.g., it's often said that Einstein's two postulates of special relativity were the principle of relativity and the principle of the constancy of the speed of light.
The cosmological principle is a postulate of FLRW cosmology. Its effect is to force the large scale geometry of spacetime to be a FLRW geometry. You could just assume a FLRW geometry instead, but the more abstract idea of homogeneity and isotropy sounds, I suppose, more principled. Principled or not, it's an assumption, not something you could know a priori. It's ultimately justified, like any postulate, by the agreement of the resulting theory with experiment.
I think it's as well justified as any other so-called principle. The principle of relativity has held up well, but it certainly could be wrong. The principle of least action is just an approximation in quantum mechanics. And so on.
A: 
Before looking, we have no a priori justificatioon for expecting homogeneity.

We do have justification for taking it as our starting point. It's akin to Occam's Razor: the simplest hypothesis is that everything is the same everywhere. As we get more data refuting that assumption in some particular senses, we come to conclude that the universe is nonhomogeneous in those senses, while keeping homogeneity in other senses our default assumption.
A: 
So you see I begin to doubt that there is any "cosmological principle" in the sense of a principle of logic or of theoretical physics. It is a misnomer for an interesting and important empirical observation.

No physical regularities can be derived entirely from logic alone. All physical regularities are ultimately empirical observations. As far as I can tell, a "principle" is simply an empirical regularity that happen to hold over a particularly broad range of regimes and situations. We are blessed to live in a universe that happens to contain empirical regularities, like the Schrodinger equation or the rules of special relativity, that seem to hold in a lot of different regimes. But of course we can't rule out the possibility that we'll eventually come across some patch of space where the Schrodinger equation or special relativity simply don't apply.
But I think the exact cutoff for what counts as a "principle" will always be somewhat subjective, and so I'd question the premise quoted above. It also depends on what you're trying to describe: a deep-ocean marine biologist could reasonably take it as a "principle" that the temperature is always 4 degrees Celsius, but a stellar astrophysicist cannot.

My question is, then: is there any argument coming either from logic or from physics more generally (not observations such as sky surveys and CMB measurements) which says the cosmos must be expected to be homogeneous?

No. It's an empirical observation that cannot be derived from "fundamental" (i.e. microscopic-scale) laws of physics.

And if it were not homogeneous, then is there any argument deserving the name "principle" whose conclusion is that the physical conditions allowing life might be equally well expected at one place as another?

No, I can't think of one.
