You seem to have gotten lost in some bizarre misreading of my arguments, but I’m not sure I can help you get clear of it, since I don’t for the life of me see how the above could plausibly be a reasonable reading of what I wrote, and you steadfastly ignore any attempt of mine to clarify, persisting in this bizarre interpretation despite me repeatedly pointing out that I don’t believe anything of the sort.
But I’ll try to be as explicit as possible for one last time, and maybe then at least you could tell me where our understanding differs. I’ll even paint you a picture.
________ (1)
/
----------x
\________ (2)
We are at the point x, with some ancestral mental state that has no particular preference regarding salads. There are two distinct successor mental states, (1) and (2), with:
(1): ‘I love salad, so I should eat some’
(2): ‘I love burgers, so I should only eat those’
Those confer a different selective benefit, (B1) and (B2), with (B1) > (B2). That’s just due to the fact that those who eat salad typically live longer and have more offspring/care for their offspring better/what have you. In this case, the argument from selection works: eventually, the population will be dominated by those who eat salad, and they will eat salad because they love salad, and hence, we can validly say that they love salad because of the adaptive benefit the resulting behavior confers.
Nothing about this has anything to do with natural selection somehow ‘making’ a trait beneficial. I’m not even sure what that would mean; it’s just that some traits turning up by chance confer a selection advantage, which means that those with the trait will on average have more offspring.
Now consider the following scenario:
________ (1)
/
----------x--------- (3)
\________ (2)
Now, we have the following mental states:
(1): ‘I love salad, so I should eat some’
(2): ‘I love burgers, so I should only eat those’
(3): ‘I hate salad, but I should eat healthily, so I should eat some’
In turn, these confer reproductive benefit (B1), (B2), and (B3). Individuals with mental states (1) and (3) end up eating salad in just the same quantities. As in this simple model, there is no other determining factor, that entails (B1) = (B3) > (B2). Now suppose we are again at point x. Natural selection can’t differentiate between (1) and (3), because it’s sensitive only to fitness differentials, and there’s no fitness differential between the two. Individuals with either will eat just as much salad, and outcompete the burger-munchers in just the same way.
Again, I stress that this is not because of any way by which natural selection might ‘make’ a trait beneficial; it’s just by it acting on differences in benefit generated by these states, and there not being such a difference between (1) and (3).
Suppose we then find a population in which only (1) obtains. In this case, the argument from selection—clearly!—does not work: there is no selective benefit of (1) over (3), so it can’t be preferentially selected. Rather, it might have been genetic drift, or pure chance (the mutation responsible for (3) just never occurred), or the whim of some aliens tinkering with the genome, or whatever.
Again, this is what I tried to make clear by the coin toss example. Both (1) and (3) lead to ‘winning the million dollars’, so to speak, hence, the fact that one wins can’t be explained by (1) coming to pass, because whether it does fails to change the outcome. So, in this scenario, we can’t argue that individuals in this model population have a love of salad because of the fitness benefit this provides. It’s just logically wrong (once more, affirming the consequent) to do so!
What confers the fitness benefit is the amount of salad eaten. We can, thus, validly claim that the individuals eat that amount of salad because of the fitness benefit this provides. But (3) would’ve led to the same amount of salad-eating, so selection can’t account for the prevalence of (1) over (3).
Now, it’s been said that, say, (3) might not be available for a given ancestral population. Say, there’s for some reason no feasible mutation that gets it to the basin of attraction associated with ‘wanting to eat healthy’. Like for a creature with scales, it probably wouldn’t be feasible to evolve fur as a means of conserving body heat, rather than feathers. So, we’d be in a situation like this one:
________ (1)
/
----y-----x
\ \________ (2)
\_____________ (3)
Here, at point x, we don’t have the option of evolving (3). So then, the argument from selection should work, right? Well, no: all that this tells you is just that you chose your point of origin such as to obtain your desired conclusion. After all, why choose x, and not y? As long as there exists a point from which the evolution of a creature that differs from the actual population only in showing (3) rather than (1) is possible, the argument from selection fails. If you could somehow show that such a development isn’t possible, thus restoring a situation like the one in the first drawing (which is the one where there is that necessary connection between the behavior and the mental state I pointed out was needed earlier), then you could patch up the argument. But you can’t simply stipulate it isn’t possible: then, you’re just assuming your conclusion. And since creatures like us, that behave exactly like us, but for different reason are certainly consistently possible, I see no feasible way to go this route.
Again, as I hope the above makes finally clear, I’m very much not saying that. All I’m saying is that as long as a difference in mental state makes no difference in phenotype (e.g. in the amount of salad being eaten), there is no difference relevant to natural selection. That’s an utter triviality. Your ‘parody’ would be apt if, by some means, being blind would yield to exactly the same behaviors as having eyes. But it quite clearly doesn’t. However, being compelled to show a certain set of behaviors, and feeling as if one choses that very same set freely, quite clearly lead to the same behaviors.