> - Maybe I don't fully understand why LIGO needs two arms. If you had a clock that could accurately measure light wave crests, could you do it only with one arm?
Yes, that's absolutely right. The two arms cancel out the frequency fluctuations of the laser itself. If we had a perfectly stable laser, we could make do with just one arm.
Regarding the question about what stretches and what doesn't, I think the general rule is that rigidity prevents "stretching". For example, a hydrogen atom in expanding space would lose momentum over time, because it redshifts, but the atom itself wouldn't get any bigger. There's no need to invoke a higher dimension here, just some things are rigid (like laser cavities and the Earth) and some things aren't (like electromagnetic waves). In fact, in general invoking higher dimensions without a strong reason to is discouraged when discussing general relativity, simply because the math is already very complicated in 4D.