I still don't think I follow.OTOH, I don't think you have identified a relevant difference between simulationist and narrativist playstyles here, either.
Upthread I've described a certain bundle of approaches:
(i) to how a GM sets up ingame situations that the players then engage via their PCs (namely, with heavy consideration being given to how they relate to the established thematic direction of the game);
(ii) to how a player engages with those situations via his/her PC (namely, with heavy consideration being given to the thematic material that s/he has built into his/her PC, both via background and over the course of play); and
(iii) to how a GM determines the consequences that flow from what the PCs do (namely, with heavy consideration being given to the thematic point that the players have made in deciding what their PC's do, and being careful to build on that without crushing or invalidating it - as per my concerns about whether it would be a mistake to render serious something which the player has treated in a humorous and light-hearted fashion).
I regard this bundle of approaches as characteristic of the sort of narrativist play that I engage in, that Ron Edwards describes in the passages I quoted upthread, and that LostSoul's blogger (Eero Tuovinen) is talking about in the passage I quoted upthread.
This - and, in particular (iii) - which draws its rationale from (i) and (ii) - is a fairly elaborate gloss on my earlier phrase "the GM is not going to 'gotcha' the player". It makes fairly clear, I think, why the notion of "natural consequences" (whether understood as meaning "natural given the causal logic of the gameworld" or "natural given genre") is not the principal concern in GMing a narrativist game. As is indicated (for example) in the HeroQuest rulebook, and as mentioned upthread by myself and others, natural consequences set an outer limit on where the GM and players can go. But within that limit, choices in narrativist play are made primarily as per my (i) to (iii) outlined above.
Are you saying that you don't think (i) to (iii) mark a significant difference between narrativist and simulationist play? If so, I'm puzzled as to what you think narrativist play is, given that (i) to (iii) are pretty much a summary of what the standard texts on narrativism (Forge essays, rulebooks for Maelstrom Storytelling, HeroQuest, Burning Wheel etc) say in characterising that particular way of RPGing.
I'm familiar with those posts (and noted the possibility of a die roll in my post to which you replied). That is almost the exact opposite of how narrativist play approaches the issue - given that the point of play is to bring out and address the salient thematic material, it would make no sense to leave the question to a random table.See The Shaman's many postings about how he uses random encounters to simulate the common trope of coincidence within a genre framework.
(I would only roll dice if I thought the reasons telling in favour of going one way and the reasons telling in favour of going the other way were equally balanced, such that decision by lot was the only rational decision procedure. But this is pretty rare, and certainly doesn't apply in this case - because the episode can certainly stand on its own as a humorous one, that is the clear default unless the reasons for pushing it in a serious direction instead are clearly countervailing ones.)