Haha! I meant to just be articulating the claims of the first two polls you posted. I'll do my best to summarize Chakravartty's article ( https://plato.stanford.edu/entries/scientific-realism/#WhatScieReal)

(THREAD): https://twitter.com/KordingLab/status/1334703783687942148

First off, to clarify, the metaphysical, semantic, and epistemic dimensions of scientific realism you identify from the article are NOT mutually exclusive theses.They are 3 different general claims that the author identifies as loosely characterizing "scientific realism."

Many believe all three, others only believe 1 or 2, and some folks deny all three. I'll try to characterize each in turn (n.b., my gloss necessarily ignores all sorts of subtle refinements one can make to each claim! See Chakravartty's article for more detail):

The metaphysical thesis: science investigates a reality that is mind independent. That reality would exist even if we didn't represent it as such. Scientists are not merely cataloging how the world *seems* to them, but discovering how it *is* (using how it seems as evidence).

E.g., oxygen and hemoglobin exist whether we think they do or not-- and indeed whether we even exist or not!

(The entities posited by neuroscience require much subtle refinement in the notion of mind-independence, since there seems to be some identity between at least some minds and some neuro-psychological entities. The nature of that relationship is a whole other kettle of fish )

Contrast the metaphysical thesis of Realism with neo-Kantianism: we can only investigate how the world seems to us-- not as it is independent of our own minds. Perhaps there is a mind-independent reality but science can't discover it.

The semantic thesis: The claims science makes are either true or false. They are "truth evaluable." Either it's true that a Hydrogen atom has one electron or it's not.

Contrast the semantic thesis (ST) with Instrumentalism. Instrumentalisim says the claim, "an H atom has one electron," just means that under certain conditions a white streak will be observed in a cloud chamber.

ST says, no, the white streak is just *evidence* of the existence of an electron. Whether the H atom has an electron is true or false independently of whether a white streak is observed. After all, lots of other factors could be interfering with cloud chamber data gathering.

Notice, this semantic thesis meshes nicely with the metaphysical thesis. It's true that an H atom has one electron just in case... an H atom has one electron (in the mind-independent world).

Compare disputes over claims one sees in neuroscience: "neuron population N implements algorithm A." Some think that this claim is true only if N's behavior has a relevant homomorphism to aspects of A. Others just take it as shorthand for A relating the same inputs/outputs as N.

Finally, the epistemic thesis: The truth evaluable claims science makes about a mind-independent reality *are* true (...or "approximately true"), and we are justified in believing them!

So the claim that oxygen has 8 protons is (1) a claim about how reality is independent of how we think about it, (2) true just in case oxygen has 8 protons, and (3) true!

So what's with this "approximately true" business? Well, most realists are fallibilists: they figure it's possible for us to be wrong about anything. So, no matter how advanced our science is, there's always a possibility that it's not true.

Indeed, the history of science is of well-justified claims that we later realize are false. So, some folks do a "pessimistic meta-induction": science has always been wrong before, so it'll probably keep being wrong!

The more optimistic amongst us like to think that science is at least getting us closer to the truth. Our best theories may have some details wrong, but we're probably not wrong about oxygen and hemoglobin.

Some folks couch the epistemic thesis in terms of "the end of inquiry"-- scientific theories, once fully developed at a hypothetical point in the future, will be true, even if some of their claims now are still false.