Becomes accepted -> by the peer review system and here is important to put a note that it may become accepted as viable theory, not yet as factual data.
The theory of evolution goes around the development of the species until what we know today, it's not about the "origin of life" itself but "life as we know it" if you prefer.
You can have strong evidences without empirical evidence, but you won't get out of the theory realm unless you can prove it empirically. And still, science isn't meant to assume something to be true beyond contest, even the most proved thing can be being analyzed by the wrong prism and therefore providing false results. "Truth beyond contest" that's what the religious Dogma is for.
1) To me, the major problem with the peer review system it that it prevents grant funding for anything that isn't backed by academia, e.g. the hobbyist genius who fucks with explosives in his methlab and sent his cat on a round trip to outer space. Consequently, this leads to 'business' science in which the vast majority of grant money is funneled to corporations and product research. Basically, keeping the peer-review system as the staple for scientific progress is a good way to ensure that progress happens slowly.
2) Not even "life as we know it." All it's about is adaptation and the mechanisms that cause it. One of the reasons for this is that a theory on the evolution of specific species requires a good definition of species. The problem is that there's currently no known (at least, I've never seen it) definition of species that includes every single living thing, without exception. My personal opinion is that the best definition of species is "the offspring of the same species." For example, I am human because I have two human parents. Science has yet to figure out a way to model a theory of species on top of evolution.
3) No, you always have theories in science. The only proof in science is 'proof within a margin of error'. A sound math proof or a logical tautology contains no margin of error.
4) Science makes assumptions it can't even study via its own methods. And, interestingly, every scientific theory ever produced contains assumptions; this is self-evident because of the statistical margin of error I mentioned in point #3.
5) Science itself is simply a theory about knowledge acquisition. Any discipline is such a theory, but they range in scope. Philosophy is the most broad and it provides the tools necessary to analyze both abstract and empirical information. Mathematics and physics represent the abstract and empirical descendants of philosophy. Math then branches off into other abstract disciplines (e.g. geometry, calculus, trigonometry, etc.), and physics branches into other empirical disciplines (e.g. biology, chemistry, anthropology, etc.).