You've raised objections that I believe are interesting, to which the answer lies in the nature of play itself.
Upthread you raised a concern with GM scaling down player declarations. Can player notwithstanding what GM says, insist upon their scaled up declaration? If not, then GM has said "No". One may believe that norms in force at the table will lead players to give, in such confrontations, but it's still "No - you can't declare what you wanted to declare"... or it's "Yes, let's proceed with your scaled-up declaration even though it would work better scaled down!" A connected observation is that players as often say "No" to themselves.
It's an obvious glitch in CR - commented on by multiple folk - that I could declare "punch the Earth in half" and if we aren't resolving my performance - how strong my punch is - then we could end up with a result that doesn't feel legitimate. I've already discussed some practices commonly used to fix that. And it is right to point out that these resolution methods do not exist in isolation: they're part of a web of practices that work together. One can to an extent analyse a single component of an Internal Combustion Engine, but one cannot complain too much of fault if the action of said component is regulated by another component leading to collective functionality.
Which leads to the solution that I think you find unsatisfactory. The playful negotiation rarely if ever leads to "I discover that X is Y". It's overwhelmingly more likely to land on "What's the deal with X, is it Y?" Logically, yes - you are right. There are all kinds of legitimation failures possible in CR. Your example of asserting a fact contrary to those already settled is just one of them. Picture "I know we said the sky is blue, but I discover it is pink." Roll... is the sky still blue? "I know we said Brother Jo is wearing a linen shirt and heavy oilcloth coat, but I discover he's standing at the pulpit stark naked." Roll... is Brother Jo naked?
Play occurs because players adopt the appropriate lusory-attitudes for that play. The cases you are speaking of are far more extreme than you might be picturing. According to Huizinga, they are those of the "spoilsport", who shatters the magic circle by abandoning the attitudes that sustain play.
There are many mechanisms which exist in Narrativist systems which obviate all of the above concerns, at least in the better systems. All of those issues were addressed through discussion and iteration of design in the discussions of Sorcerer, The Forge posts, and various blogs, mostly in the early '00s. At this point it is well-trodden ground!
So, we discussed the role of fiction, for example in DitV. It isn't there to serve as a menu of things that players must discover in order to navigate some existing plot. There IS no plot, and the fiction is simply revealed in a narratively appropriate way. So, sure, if you don't go to the Mayor's Office you may not see the safe, and maybe you thus don't have the opportunity to get the papers, but (demonic forces aside) its not the focus of play to test your ability to crack open safes. That fact isn't revealed simply because there's no narrative logic leading to its revelation. Presumably the character went somewhere else, and revealed some other facet of the highly pregnant with conflict potential situation! Very soon they will cross paths with NPCs who will oppose the PCs goal, purging the town of its evil influence. Conflicts will involve convincing people, maybe intimidating them, maybe spiritual and physical struggle, etc. Those things are not part of any secret backstory, and whatever information is required to bring them forward and resolve them will be presented as required. Thus there is no potentiality for 'facts to get in the way' of intent.
In terms of things like 'punch the Earth in half' we have basic garden variety solutions to this problem that have existed since the first days of Dave Arneson's game! The rules don't allow for such actions! In DitV if I stated that my character's desire is to march down to the Governor's Residence and seize power the GM will simply present me with obstacles which my character is entirely incapable of coping! This is just basic narrative integrity and gamist game integrity stuff, it is the least possible concern! I'd note that the 'scale down' advice in DitV, from the sound of it, isn't really so much concerned with THAT as it is with the more nebulous kind of "well, we just cut to the heart of the conflict and toss a few dice." sort of issue. Since intent type resolution systems don't really INHERENTLY present a 'scale' this can be a concern. I mean, you could imagine the same sort of question in a 4e campaign "Hey, why don't we just have a skill challenge to decide if the RQ takes over the whole Lattice of Heaven or not?" I mean, you COULD, the rules don't really put a limit on the stakes of an SC... That's where one of the GM's main tasks in Narrativist play comes in. Note how it is approached in AW/DW, moves only have limited scope and everything that has stakes will trigger SOME move pretty quickly, so there's no way to trigger "I win the game" in one shot! Beyond that AW's strongly articulated principles/agenda means the GM shouldn't want that play, and has the power to make moves which will effectively 'scale down'. It is just not an issue in any actual Narrativist play I am aware of.
As for the idea that any of the above amounts to 'GM Control' or 'saying no', nonsense (and you mention lusory attitude, which I think is sufficient to cover this, so I think we agree here). The rules of the game are plain and agreed prior to play, as is the premise, agenda, and principles/techniques. While it may be too much to ask that every player is fully cognizant of every implication of what they're agreeing to play, it is pretty clear that a basic agreement on the satisfactory nature of the 'rules of the game' is a precursor to success. Nor is it any more 'GM control' in AW to say you can't punch the Earth in half, than it is to say in D&D you can't just walk through the walls of the dungeon and see every secret door. This is all just part of how the game works.
So, frankly, I don't see any of
@Crimson Longinus objections as being really substantive. I will grant that, if you have only understood RPGs from the Trad D&D point of view, then some of these may be questions you will ask, but the answers exist and are well-practiced.
