The basics of how I view the universe in terms of epistemology and stuff
To start with, I divide things into two categories -- the physical and the metaphysical. The physical universe includes everything that can be observed, detected, or measured. It includes phenomena which can be analyzed, for which models can be formulated to predict future events, and from which past events can be deduced. The way I use the terms, "science" can be defined as "all inquiry into the physical universe." Not all sciences are advanced enough to make accurate predictions (economics is an example of a science in its infancy, which, at its current level of development is somewhat useful in some circumstances, but which can not, yet, be generally trusted. However, it IS a science by my definition, because the phenomena which it studies ARE observable and modelable, even if the models aren't there yet), and, indeed, it may be that some phenomena simply can't be actually modeled, because they are complex enough that analyzing them would take more computing power than can theoretically exist within the universe, but, nonetheless, you can tell how they COULD be modeled and predicted.
Truths about the physical universe, I call "facts."
The metaphysical universe, on the other hand, is made up of things which CAN'T be observed and analyzed, by their natures. Not stuff that we can't CURRENTLY observe, but stuff which simply couldn't be observed at ALL. Concepts of ethics, morality, justice, honor -- these are metaphysical things. But they genuinely exist and are real. The question of "is this specific action right or wrong" has an actual, true answer, just as much as "what wavelengths of light does chlorophyll metabolize most efficiently?" The difference is that the answer to the latter is a "fact", while the answer to the former -- and this is one of the confusing bits -- is an "opinion". By one definition of "opinion."
I use the word "opinion" to refer to several very different things. The first meaning of "opinion" is "a thing that a model of the physical universe predicts, when the model is not yet reliable." If the model works, then the opinion will actually be a fact. If the model is flawed, then the prediction may fail, and the opinion will turn out NOT to be a fact.
This definition of "opinion" falls within the physical universe, and is part of scientific inquiry. It means "best guess". This form of opinion is useful, in that the models being used, while perhaps not completely reliable, are still accurate enough to give a better-than-random-chance of being facts. A doctor's medical opinion is an example of this definition. If my doctor's opinion is that my symptoms will most likely be ameliorated by a particular treatment, that opinion is more likely than not to be a fact -- and whether the opinion was correct will only be known later.
Note, however, that this kind of "opinion" IS either right or wrong.
The second form of "opinion" deals with questions which fall neither within physical OR metaphysical inquiry, and therefore DON'T have actual right or wrong answers. This is probably the most common meaning -- and the least useful. If I say that "Kurosawa's THRONE OF BLOOD is the best filmed Shakespeare in existence," that statement would only be true or false if Plato's idea of Ideal Forms was accurate, and there exists a platonic ideal filmed Shakespeare, and a true way of determining closeness to the ideal, and if we define "best" as "closest to the platonic ideal."
In my understanding of metaphysics, none of those things are true, and therefore not only is it impossible to determine the truth of "Kurosawa's THRONE OF BLOOD is the best filmed Shakespeare in existence", but the statement actually lacks a truth value at all. I'll make the statement, and mean it, but I understand that this is a matter of taste only, and not one that is actually true or false.
Also in this category is the creation of categories. Whether a particular thing falls within a particular category is a scientific question, about the physical universe, the categories themselves are human-defined and have no external existence. One clear example of this is the demotion of Pluto from planet status. Nothing about Pluto changed, but our arbitrary human definitions of categories did, and an object that fit an old definition didn't fit the new one. Whether Pluto fits one definition of "planet" or another is a scientific question. What set of attributes the category called "planet" defines is an opinion with no truth value. There is no inherent "planet-ness".
I believe that categories have no physical OR metaphysical existence. Nonetheless, they're darned USEFUL -- it really helps communication. If an astronomer declares that they have discovered an extrasolar planet in a different solar system, it's nice to have actual understandable and agreed-upon definitions of "extrasolar", "planet", and "solar system".
Which brings us to the final category in my understanding of reality, and my third mutually-exclusive definition of "opinion". This consists of things like "right" and "wrong", "justice", "mercy", "honor", things like that.
In my understanding of reality, these are things that actually metaphysically exist. Other people believe that these are human-created categories, like "planet" or "bug" or "nation", but I believe that they are actual things that exist external to culture and belief.
They're not detectible by scientific method, though, and therefore, there is no way to determine it.
My third definition of "opinion" is related to this, and, sadly, is therefore difficult to tell from that second meaning. If I say that an action is ethical, my "opinion" is a statement with a truth value -- I might be genuinely correct or incorrect about that. Because these statements are not approachable with scientific inquiry, there is no generally-accepted way to test them.
All I can do is state my postulates, and use logic to build up my argument to explain how these postulates demonstrate the ethical, honorable, or whatever nature of the action under question. It's possible to work out through inquiry that IF these postulates are true, then the conclusion holds -- but it is NOT possible to use inquiry to determine the postulates in the first place.