As it raised its head in my ongoing project to review Dembski & Licona's Evidence for God I thought I would mention I was introduced several months ago to Alvin Plantinga's Evolutionary Argument Against Naturalism (EAAN). I understood at the time that the EAAN was expounded in Warrant and Proper Function, an expensive book that I wasn't inclined to buy. However I also understood that a shortened version of the EAAN was contained in Plantinga's paper "Content and Natural Selection". Having looked at the paper I was glad not to have shelled out for the longer work, as it seemed likely to be impenetrable to me.
As for "Content and Natural Selection", I couldn't get much beyond the first page. It starts off with a proposition that in English appears to say that the probability of natural evolutionary processes selecting for reliable belief-forming mechanisms is low. I wondered if this was intended to represent what atheists maintain in order to explain what they see as the preponderance of false god-belief. In such general form it would indeed seem to be self-refuting — if evolution gives us lots of false beliefs as a survival mechanism, then any belief, be it belief in God, in metaphysical naturalism, or in the proposition itself, is more likely to be false than true.
Regarding the initial probability argument, unless I've misunderstood the initial thrust (which I may have done), Plantinga actually starts off with the proposition that the probability of evolution selecting (indirectly) for true belief is low. I don't see any support for this proposition (to begin with I thought he was proposing this as the naturalistic position, but apparently he's not). He quotes Quine and Popper who suggest that evolution would tend to select for true belief, but doesn't refute them other than by quoting other authorities. Without such refutation maybe we should go with Quine and Popper. So in his own terms Plantinga appears not to have a foundation for his very first premise. Naturalism therefore stands as a warranted worldview.
Notwithstanding the above, how does the EAAN deal with the idea that god-belief isn't an evolutionary advantage in itself, but merely a side-effect of our bias towards belief in agency? In its general form the proposition applies across the board, ignoring the possibility that natural evolutionary processes could select for behaviour resulting both from true beliefs about some things and from false beliefs about others. They could, for example, select for behaviour resulting from true beliefs about natural evolutionary processes and for behaviour resulting from false beliefs about gods — or vice versa.
Plantinga seems to be suggesting that beliefs as a result of evolution are present fully formed, with no account taken of experience. People's beliefs are not wholly formed by their genes, they are also formed by what they perceive in their lives. Their perception may be influenced by their belief-forming mechanism (whether or not that mechanism is a result of evolutionary processes) but mostly they will believe something because they perceive it to be true. (You can be sure, however, that Plantinga's supporters will point out that perception is part of our belief-forming mechanism.)
But is any of this valid? Are "beliefs" — true or false — the kind of things that are so intrinsically bound up with behaviour that they can be naturally selected for? Only if different beliefs have behaviours in common, which themselves can be selected for. Our beliefs are not inherited genetically, and our belief-forming mechanisms are only partly inherited. Beliefs are, however, often passed on to children through indoctrination, so the selection mechanism may well be similar.
Plantinga is proposing that the truth or falsity of a belief is only an indirect selecting factor, because it's likely that the truth or falsity of the belief may be irrelevant to its survival value. What matters about the belief is that it encourages or discourages particular actions. It's those actions that are acted on by natural selection, regardless of whether they are instigated by true belief or by false belief. Some actions may have good survival value despite resulting from false belief, and vice versa. Despite the convoluted hypothetical examples Plantinga has given elsewhere, it's clear to me that there are likely to be many more true beliefs that lead to survival-promoting behaviours that there will be false beliefs leading to survival-promoting behaviours.
The proposition's generality makes it unsound. God-belief could be a false belief of a special kind, a kind that has fewer or weaker consequences than false belief in agency in general. So our tendency to believe in agents where there are none may have stronger consequences than our belief in a non-existent god. In that case natural selection would have a greater effect on our belief in overall agency than it would on our belief in a particular god.
In "Content and Natural Selection" and in Warrant and Proper Function (a version of which I have since found online) Plantinga employs a fairly dense style and contracts much of his argument (at least in the final chapter of WaPF) into hard-to-parse mathematical notation in order to show that belief in naturalism is unwarranted. Yet despite a whole book leading to this conclusion, his contention that this doesn't apply to theism is tossed off in a vague paragraph about man being created in the image of God. This is, at the very least, disingenuous. Also I note he's using his own particular definition of naturalism ("the belief that there is no such person as God") that appears to be calculated to favour his thesis, so that when he claims to show that naturalism is unwarrranted, it automatically follows that God exists.
Rather like presuppositionalism, Plantinga's thesis seems to be a negative argument — casting doubt on the reliability of our cognitive mechanisms. We think something is true (or false) but our basis for determining truth is apparently undermined. This is a bit like saying, "You can't disprove the existence of God, therefore he exists." I don't buy it.
MATS AI Safety Strategy Curriculum v2
1 hour ago