It feels like this can be read in two ways, one of which I heartily agree with! Two recent essays on simulationism that have influenced me are Tuovinen and Macris.Various threads - eg the "why rules?" thread, and the agency threads - have prompted some conversations with RPGing friends.
At least from my perspective, a recurring theme or element in discussion has been a contrast, between:
*Play where the players establish dramatic needs for their PCs, and these drive play (in part because the GM presents situations that speak to those dramatic needs);*Play where the GM establishes the drivers of play.
There are different ways of doing the second. One popular one involves the GM preparing, presenting and adjudicating a setting. One of my friends offered the following description of this way of using setting, which I though was pretty good. Here it is (as paraphrased by me):
In this sort of setting-focused play, what is central is the GM's orientation towards framing, and towards content introduction more generally. The GM is akin to a guide. (Which contrasts with the first approach above, where the GM is provoking the players to action based on the player-established dramatic need, and using the relevant system tools.) And the setting is the default protagonist: each NPC, each faction, each "side-quest" that the GM presents to the players, has its own dramatic need that propels play; and it's the players' job to get in on the action, and hopefully to find some overlapping interest with one or more of these setting elements.
This approach also brings some related expectations and techniques (and here's some more paraphrase):
*Players must be willing to accept the role of the setting as default protagonists. In this sort of play, it is disruptive for a player to subvert the focus on the GM's presented elements, and to try and make play about their PC's own dramatic needs. (A label I've seen for such disruptive play is main character syndrome.)*There is a very "breezy", relaxed approach to free play - ie the players declaring low-stakes exploratory actions for their PCs (basically wandering around the world), which the GM responds to by revealing appropriate elements of the setting. The ethos is one of "We'll get there when we get there." In this sort of play, for the GM to be pushy or obviously proactive about one of the setting elements they have established is seen as tantamount to railroading, or at least to be the thin edge of the railroading wedge.
It is sometimes assumed that the only alternative to this sort of play is "linear", railroad-y, adventure path-style RPGing. But that's obviously not true - see the first entry of my two contrasted approaches at the top of this post. Once we put aside that assumption, and try and describe the setting-sim approach on its own terms rather than in terms of some single conjectured contrast, I think we can get a clearer picture of how it works.
Pursuing the reading that I agree with - I feel you are right that once we set aside the dichotomy we possibly assumed (as contrasted by you at the top of your post) it opens our eyes to a much broader range of approaches. Two recent game texts that speak to this are Fantasy Flight's Legend of the Five Rings, and Chaosium's RuneQuest: Roleplaying in Glorantha. You've focused on setting, but let me shift that a bit and say that the essence of simulationism is focus on subject.
In L5R players are "samurai, who are torn between their personal desires and their sworn duties" while RQ "lets you experience that [mythological] state of mind and explore that [sacred] way of life through the mythic realm of Glorantha." Defining features of the L5R setting (of Rokugan) include the celestial order and the code of bushido, in RQ, Genertela is part of a mythological world in which gods, spirits and sacred acts are visible, active, and necessary. In both game texts, considerable investment is made up-front to situate your character into setting... but is it setting that is the focus, or the affordances for roleplay that separates characters from our ordinary world and make them suitable for exploration? That is, of the tension between their desires and duties in a strictly organised society, or of their sacred heroism in a vividly mythological reality. I don't think we have to answer this question in just one way: the game designs offer utility for multiple modalities.
Anyway, long-story short, to me your closing thought in your OP is superb. It might be you specifically want to focus on setting-sim, but I am not sure that it is quite that simple. I suspect that setting itself is serving purposes connected with what it is possible to explore via roleplay. It seems easy to think "well, I want to explore Glorantha" but that is not at heart what RQ is about. Not that simply. Exploring Glorantha is about exploring the sacred and mythological. You can easily see the similar argument with regard to Rokugan.