I'll hang my hat here for the moment. Let me stress I am not saying this to avoid the failure. I love the failure of my other approach. I consider it temporary! I just thought the glaring problem of this term supernatural needed to be pointed out.
I actually think this is the solution to your difficulty. When you said
For now, I just can't get my argument through. I won't stop thinking about it. But as it stands, I don't see any way to swat down this possibility for evolution to result in false beliefs.
the first thing I thought of what Plantinga's well-refined Evolutionary Argument Against Naturalism. But the key here is that it is an argument that assumes the premises that both evolution and naturalism are true. By bringing the prefix super into this issue, I think you are recognizing that the task you have set out depends on whether one holds to a naturalistic evolution, or a nonnaturalistic one. Evolution and naturalism are often conflated, but as ARM would be quick to remind us, they are different categories and must be treated separately.
Plantinga argues that if both evolution and naturalism are true, then we actually have no reason whatever to suppose are beliefs are reliable. We only have reason to suppose that our beliefs are survival-beneficial. We also have to understand that we cannot take our experience of reality as we know it and glean from that that our beliefs are reliable, because we can't assume that this reality is defined by naturalism and evolution both being true. Thus we can't use our experience of apparently reliable beliefs as evidence contra Plantinga's argument; the argument aims to show what would be the case if both evolution and naturalism were true. The short form of the argument, without any exposition, runs as follows:
1) The probability of having reliable beliefs given naturalism and evolution is low
2) Anyone who accepts (1) and believes that naturalism and evolution are true, has a defeater for the belief in the reliability of beliefs.
3) Anyone who has a defeater for the belief in reliable beliefs has a defeater for any other belief one holds, including the belief in naturalism and evolution.
4) If belief in naturalism and evolution necessarily leads to a defeater for the belief in naturalism and evolution, then belief in naturalism and evolution is self-defeating and cannot be rationally held.
Proving (1) is obviously the critical step. And again, it amounts to differentiating between "cognitive indicators", like true data gathered from an environment to enable adaptive behaviour, versus true beliefs about this data or the environment. It is not clear that the beliefs themselves have to be clear for the cognitive data to be nonetheless received and processed adaptively. And so if the belief-about-data, for any given datum, is just as likely true as false, and if we expect "reliability of beliefs" to be defined, let's say, by a 3:1 ratio of true to false beliefs, then given a mere 0.5 probability of a belief being true, the probability of achieving these 3:1 ratio is a belief set of 1,000 beliefs is small enough that we can no longer consider beliefs as reliable, given naturalism and evolution.
That is a very brief summary of the argument. Plantinga goes into great detail attempting to pre-emptively answer specific objections, or dealing with semantical arguments, but if it is not necessarily the case that true beliefs would accompany true cognitive data, then I think the entire argument follows, and, indeed, I would agree with you in saying that one cannot "swat down the idea" of evolution resulting in false beliefs - in fact, if we are taking the conjunction of evolution and naturalism as true, it would seem it is just as probable for evolution (and naturalism) to produce false beliefs as true.
Which brings me back to how I began this comment. By bringing the prefix super into this argument, you are subtly gesturing to a theistic evolution, which is something entirely different from the "naturalism and evolution" conjunction. A conjunction, instead, of theism and evolution would, I think necessarily, lead to the reliability of beliefs, for the simple fact that, in this case, evolution would not be random - or rather, it would not be selecting merely for survival, but instead evolution would have been designed for the sake of selecting for gnosis - the ability of the evolved creature to understand the universe, and by extension, God. Which brings us to one of those critical three principles of natural law that mankind has, whether he knows it or not, come to reject as a result of the Protestant Reformation on the side of the religious right, and as a result of the Enlightenment on the side of the secular left - namely, the principle of the intelligiblity of the universe.
A principle, I might add, that anthropic realism - held by those "moronic primitives" - itself assumes.
So basically, I did argue myself to the truth, given the conditions of (a) naturalism being true and (b) evolution being true. Plantiga makes the case that if these are true, we have no justification for belief in any belief.
I'm reading a chapter in a Feser book at the moment, and I came across this quote. I thought you might like it for its relevance to your comment:
...elsewhere I have argued that our capacity to grasp abstract concepts and to reason in accordance with formally valid patterns of inference is something incompatible with naturalism, and that the naturalist cannot evade the problem by attempting to deny that we truly possess such concepts or reason in such ways. That alone is reason to reject any naturalist epistemology.
Scholastic Metaphysics (2014)
Here he is making reference to a much more in-depth argument to this effect that he made in Aristotle on Method and Metaphysics (2013).
Feser is on my reading list for sure.
(post is archived)