You may be missing something. The gene view / kin selection / inclusive fitness theory predicts normal one family/ monogamous queen ant colonies just fine because the ants are related and have the haplodiploid sex system.
However, the problem is the super colonies for aforementioned reasons. The ants are cooperating with ants to whom they are not (very) related, which seems like it would not evolve and also would be unstable.
Reiterating quickly a couple of things:
- Altruism - as in a gene sacrificing it's fitness for the sake of other genes - is a losing strategy. The gene by definition doesn't take care of itself in a competitive way. Natural selection weeds such genes out.
- But if you somehow get a group of altruists, then it is generally very good to be a selfish individual in that group (take from all, give nothing vack). Hence, selfish genes and strategies would be expected to invade a group of altruists
Wilson, as far as I can tell, says that
groups of altruists outcompete groups of selfish individuals. But AFAIK his group selection idea doesn't explain how pressure on groups also turn into pressure on the individual. For example, queens and their genes compete against many other queens with their own genes. It's easy to imagine some of the selection pressure that would come from that. Queens and their genes make and shape colonies. Their everyday conflicts would shape those genes. While altruistic groups may be competitive, I still can't quite see what pressure there is on these genes to evolve 'nicely' on behalf of a largely unrelated group. What's the mechanism?
I currently don't feel like group selection is the answer, but I'll admit it's a conundrum. I suspect it's a varied explanation that has a lot to do with ant ecology coupled with possible constraints / limits. Like I wrote, if ants CAN'T tell which ants are kin and which are not (just nestmates from non-nestmates), then something as simple as that should be expected to turn them altruistic. The reason is their ancestral trait is that ALL of an anthill used to be kin. There was no other way, hence they didn't have to recognize kin. If they were friendly to nestmates, then they were nice to kin. Hence, that ancestral ant would be nice to a nestmate whether it was kin or not.
So maybe they could have carried such naivety with them from olden days? You wouldn't expect that because in a colony with several queens/families, selection would so favor any new mutation in ants that lets them recognize kin. But possibly, there could be some limiting factor to ants ability to evolve such a trait.
That's just my thinking so far and I expect it to change in the near future
Edit:
I'll check out Wilson's lecture when I find the time