General formulation of Problem of Evidence/Epistemology

Two general problems in trying to ascertain knowledge:

--> Look too closely, and you deal with problems of interpretation and meaning
:: Translation, Miscommunication, in general Solipistic problems

--> Look too far out, you over-generalize
:: Stereotyping, over-extrapolation, in general Platonic problems

So, where is the solution?

Preliminary question to be re-visited:
What is the purpose of evidence and of knowledge? Why are these things sought by you, by humans?


-----
Assuming a unified, objective reality, clearly the only possible solution resides in “looking closely”,
but presumably in a specific, systematic way. 

:: Perhaps resolve problems of meaning via Godelian or Hofstadterien methods?

Once you ground sufficiently verifiable and practicable human knowledge in
your objective reality, via a process of abstracting a finite object from (presumed) infinitely divisible reality…

:: note that “verifiable” and “practicable” are attempts at getting closer to WHY anyone cares about knowledge, the question outlined earlier; my contention is that humans in fact should NOT care about an absolute knowledge, but only the kind of knowledge that makes sense to talk about at the level of concepts humans deal with.
      --> :: For instance, we don’t care about the energy contained in each of the trillion particles in
                the room, but only about the aggregate, average “temperature”.
      --> :: Humans deal with big things, because on the scale of the universe, we are rather big. We
                should not be ashamed to admit our nature.

…, you can build an arbitrarily  “far out” or arbitrarily large (as large as you decide, but not infinitely large) perspective out of your finite objects. And you can change scope of perspective as is appropriate.

All knowledge/Evidence arrived at should work for any degree of scope in perspective, whether long-term or short-term, I think?