I'm a very cowardly person who hasn't looked again at the other thread that I posted, since I posted it. As a result, I don't yet know what reply I got to it. I hope this can still be a good forum contribution.

I still don't know how to make mods, so I'm proposing a mod idea in the open. The intention of this proposal is that anyone can claim it without necessarily consulting me on any detail of the idea's implementation (though I may still conceivably do my own later). I may do more of these posts if they're tolerated; no promises.

In the hopes of reducing the win more cult of personality outcomes that this game tends to produce (with frequent 95%+ popularity positions), I think a new feedback restraint mechanism could be added to the game by adding a stat called Values Conflict. Values Conflict would represent one of the problems with trying to produce an omni-ideology: people ultimately are not in fact identical, and the unipolar dominance of any particular single point is unlikely to be ideal for most people.

Powered by any faction with extreme approval values, my concept for the implementation is this:
Values Conflict should push down the Everyone popularity and feed into some of the negative situations in the game. Undecided: do any factions appreciate Values Conflict and grow more peaceful in societies that have become intent on considering the problem of meeting the needs of many diverse people?
The popularity of all factions should feed into this value as an increaser, scaled to worsen with increasing popularity. This stat is driven by the implicit conflict in ideological alignment.
The membership rates of all factions should feed into this value as a reducer, though scaled such that only unusually large factions should reduce it noticably, because the reason by which membership rates should reduce Values Conflict is that unusually unipolar societies would experience less of it.
Very few policies should affect Values Conflict directly. Notably, censorship policies should reduce it, since stifling embers becomes possible, but it might be appropriate to create a long-term growth in Values Conflict on such things due to their intrinsic reliance on preventing people's peaceful activities. (Few people include being personally censored in their view of an ideally governed society.) I think a 2-turn delay on the decrease and an 18-turn delay on a nullifying increase would be appropriate.

Closing words:
When any one particular attempt to solve the collective preference problem rises to dominance, the importance of its divergences from its constitutive ideals becomes increasingly important.