The thing is, the smelly chamberlain isn't really an example of authority, but an example of various cases of assent (or non-assent).
<snip>
what becomes true in our shared imagined events doesn't become true simply because a rule says it is true. It becomes true because we all assent to it being true. Rules are just another form of authority.
That's the negotiation I think Vincent is talking about.
Sure.
But I think what is also being said, in the discussion of rules that I originally mentioned, is that they are meant to facilitate this process of assent, by establishing expectations about who will propose what, and who is obliged to assent to it. The comments on the Smelly Chamberlian blog bear this out - I can't remember his precise wording, but Vincent uses the idea that in certain contexts (eg when a participant is doing something that the rules say they should do) assent will be almost automatic - and
not assenting is likely to raise serious social contract isssues ("Hey, we all agreed to play the game like this - what do you think you're doing?").
So if the game
did have a NPC body odour chart, such that the players' secret plan in the Smelly Chamberlain case
was a case of drifting, then I think this would be likely to change the social dynamic in certain ways. And, as I responded to the Jester, I think if we move from something that is typically peripheral in most games - body odour - to something that is often more central - moral/religious commitments and allegiances - then I think the interplay between authority, drifting, expectations of assent and actual assent will show up pretty obviously.
To tie this back to my OP: if the rules are gameworld-modelling style rules (of the RQ or RM sort, or the sort desired by those who say that Come and Get It should be an attack vs Will, so that the mechanical resolution of the power itself explains why it is that the assailants close on the fighter) then this generates an expectation of assent with the model in question, and imposes a cost on attempted departures from it (especially by non-GM players - at least in traditional groups, the GM is generally awarded a lot more leeway here, and some rulesets expressly confer it - for example, Rolemaster criticals specify injuries to particular body parts, but the rules say the GM may adjust these "if appropriate" eg if a leg crit is delivered to someone taking cover behind a shoulder-high wall).
Whereas HeroQuest style mechanics, which separate the method of mechanical resolution from the ingame causal process,
oblige someone at the table to actively put forward a suggestion as to what actually happened in the fiction (assuming that the fiction is going to be discussed at all - let's assume that the mechanics in other respects make fictional positioning important, and therefore give the participants a reason to settle the content of the fiction). The process of resolution does not itself establish a default candidate for assent.
At least in my experience, this is not merely an abstract distinction. For examle, if we're playing D&D (any edition, I think) then - assuming no drifting - a roll of 1 on an attack die mandates assent to a description of the attack as failing to bring the opponent closer to defeat. But does it
also mandate assent to a description of the attack as incompetent, or poorly aimed, or otherwise something the attacker should be ashamed rather than proud of?
The rules of every edition are silent on this point, I think - D&D has never had fumble rules. But there is a clear tendency on the part of D&D players to treat the die roll as modelling the "oomph" or skill of the attack, with a 20 being "maximum skill" and a 1 being "maximum ineptitude". Under this approach to the rules, then, 1 in 20 attacks is going to have the attacker doing something which is an instance of the maximally inept. Whereas if we treat the roll of a 1 as simply stipulating an outcome, with the fiction to then be inserted independently of the process of mechanical resolution, it becomes open to narrate the roll of 1 as the most tremendous demonstration of swordplay ever!, but matched by an equally dramatic series of parries and ripostes by the opponent.
In the "D20 roll as model" case, it is clear to me in what way the rules are providing the basis for the shared fiction: they establish a default proposition towards which assent must be given if the game is not to be drifted. In the "d20 roll as mere outcome" case, then the rules provide boundaries or parameters on the shared fiction - without drifting, there is a default expection of assent to the proposition that the attacker came no closer to defeating the defender - but the bulk of the fictional content, and in particular the ingame causal process whereby that default outcome was achieved - remains to be established more-or-less independently of the mechanics. In this latter case, only in a rather thin sense to I feel that the rules are providing the basis for the shared fiction. Most of the work is being done by assent to suggested fictional content where the main consideration is not the mechanics, but genre expectations, general table conventions, and the like.
EDIT: All of the above is probably tangential to what I suddenly realise (or feel as if I realise) is your real point - that rules can't establish the fiction, only assent can. I agree with that. But I'm also assuming (and as I indicate above I think that here I'm just following Vincent Baker) that the rules can play a significant role in establishing expectations of assent, default propositions to assent to, and so on. And I'm trying to distinguish two ways of introducing those default propositions - one which treats the game's mechanical processes as a model for the gameworld's own causal processes, and one which doesn't. And I took Monte Cook's comment to presuppose that mechanics ought to be designed to work in the first of these ways.
