Assumptions are
aways involved, in any human endeavour. Science is no exception. Making reasonable inferential assumptions (in a Humean sense), whose absence would raise more questions or require special conditions, is the best policy - for example, assuming the sun will rise tomorrow as it has done since records began, or assuming that the world (or the universe) doesn't coincidentally end at the limit of your vision.
Space is expanding, so the greater the distance between two objects, e,g, galaxies or galaxy clusters, the faster they recede from each other. At a certain distance their separation velocity will exceed the speed of light, meaning they are causally isolated from each other. No signal can reach from one to the other, no matter how many observers you have between them.
Not quite sure what you mean here. Multiverses are certainly not a conclusion - as already mentioned, they are predictions of theories (some well tested and established), given plausible assumptions.
Sure - some multiverses are the result of exploring, using current fundamental physical theories, how some prior state could produce big bang conditions that would develop into a universe like ours. Of course, that doesn't explain where the multiverse came from, but they're generally compatible with not having a beginning, i.e. having always existed, though not necessarily with an arrow (direction) of time. Sean Carroll's book '
From Eternity To Here' has some readable explanations of what these various ideas are and how they came about.