from the angle of "explaining observations" a fundamentally incorrect theory can work well. but it breaks down as soon as you interact with the object of the theory in new ways. in epicycles' case that means traveling to a satellite that doesn't orbit the earth. usually the theory is disproven before it kills people, but you know, possibility's there

I don't believe in fundamentally correct theories. All theories have limitations. They work until they do not. Is general relativity fundamentally correct ? I wouldn't call it that (others may disagree), because it doesn't explain quantum effects or very early universe. So to me theories are just tools that explain what we see and predict the future.

I'm with you - "All models are wrong, but some models are useful".

I do think dark matter is tantalizing because it hints strongly that our model needs a fundamental change in understanding at some other level. It just feels like too much wall paper to be adding to the model in an attempt to cover the holes.

Doesn't mean I find any of the current alternatives more viable then our current models.

exactly, so when we attempt to put general relativity into practice at the limit of its predictive capability, we may see something is very wrong, then extrapolate a better model, and a few decades later everyone will look back on general relativity as an example of science's ability to carry us away from grossly false beliefs. as with epicycles. and maybe dark matter.

Theories, sure, but knowledge is not confined to theories, i.e. isn’t there positive knowledge, a.k.a. “facts”?

Yes there are facts. I.e. there is a particle called electron, there is a Sun, Cosmic Microwave background, Milky Way galaxy. I'd say those are facts. But I do think the boundary what's a 'fact' vs model/theory dependent statement can be different among different people.

There are theories with super high levels of supporting evidence, so we treat them as true.

This isn't to say that is a bad thing, it is a great thing.

I treat them as true because they are true. I guess there is more than one definition of truth.

But I think what you mean that you treat some theories as axioms (temporarily, I hope). It's a great aid in thinking but a great source of confusion for newbies as well.

There are absolutely multiple definitions of truth. "1 + 1 = 2" is not true in the sense that "Paris is the capital of France" or "Water boils at 100C" are true.

Outside of mathematics, not really. Special relativity and QM proved that 300 year old "facts" were actually subtly wrong, for example.

I think the OP's point is that we should understand our current models of natural phenomenon to be just that, models. Outside of their predictive capabilities we should remain humble about how much "truth" they contain and be skeptical of their ability to tell us what's "really going on".

Kepler and Newton's models produce better predictions over a long period than Ptolemaic models, which is great. But it's a bit of delusion to think that necessarily means they are analogous to the cosmological mechanisms that produce orbital mechanics.

> I think the OP's point is that we should understand our current models of natural phenomenon to be just that, models

this was my point with the initial epicycle comparison