It's archaeology's version of a theory. You see a huge city that appears to have vanished pretty fast, so you come up with some ideas about how/why that happen

It's archaeology's version of a theory. You see a huge city that appears to have vanished pretty fast, so you come up with some ideas about how/why that happened, like, say, ecological collapse (hypothesis). Once you find some plausible evidence that for that, you flesh out what that would look like more broadly. Now you have a theory (narrative) which should have some explanatory power. Then, as you encounter other evidence, you expect it to fit with and support that theory (narrative). If it doesn't, it may turn out your theory (narrative) was wrong, or at least incomplete, just as in the rest of science. Just like theories in other fields, these narratives can be kind of sticky and survive longer than the evidence suggests they should. It's not a perfect mapping, but that's basically what's going on.