The value question would be open after the evolutionary explanation is completed.
Yes, but any values-based sub-optimal strategy (in Darwinian terms) must survive the argument that it is sub-optimal in Darwinian terms. The two-earner middle class family is a case in point. Surely, it gives members of both genders more autonomy, but when the pandemic comes, and Mom isn't already at home to watch the kids who can't go to school, there is hell to pay. All of a sudden, "freedom" becomes self-indulgence. In other words, when competing norms are weighed, the Darwinian consequences go in one pan or the other, and, all else equal, that weight makes all the difference.
“Entrenched” means that the roles have been established, but roles can be established in unjust ways.
But that does not mean that they were so established. I'm not saying that all norms are good because they are entrenched; I am saying that some norms are entrenched because they are good. The Darwinian superiority of a norm as a social strategy is one factor in deciding whether any given norm is entrenched because it is a good norm.
When you say that maleness and femaleness make for better gender roles than egalitarianism, are you saying the social priority should be to reinforce that traditional dichotomy...
Yes.
... and that that value judgment somehow follows from the evolutionary reasoning about who ends up on top of certain struggles for dominance?
No. (But nice try at suggesting the first thing implies the second.) Gender roles needn't include dominance and submission. We have evolved past that. At common law, the wife's identity merged into the man's and any child born to a married woman was by law the child of her husband. Neither of these rules was "fair," but they probably conferred some Darwinian advantage in the way those long antlers that ended up being the undoing of that species of elk once had an advantage. Things change, and norms change. Natural selection is not an originalist enterprise.
Why not go back to ...
Because strategies (including mores and norms) obsolesce in the sense that they no longer confer a Darwinian advantage. Technology creates new plus-sum opportunities, and the winning strategies change. Natural selection may favor conservatism, but it does not favor rigidity. "Hard to move" and "impossible to move" are dichotomous opposites. One means movable, and one means not movable.
I think the idea of a "dominance hierarchy" is too vague. The issue, I believe, as does Mr. Will as you quoted him, is what sort of dominance hierarchy one favors. Does "meritocracy" not create a dominance hierarchy? Every time I call a Trump voter "stupid" - I try very hard to say "foolish," but we're all human - am I not asserting that people who are not so stupid should "dominate" those who are so stupid?
I think your definition of "laissez-faire" is an ex post notion, if I am using term correctly. IF "government" stays out of the bargaining long enough for the plutocrats to capture the government, then the government was badly constituted, and "laissez-faire" is an outcome and not a cause. The problem is not "capitalism" but governance. There's no end to the bad things that can occur under bad government. But under good government, capitalism is not "laissez-faire," so it would be best not to include the word "capitalism" in the description of the problem.