Whose? Dro's? That's real, not part of the fiction. Harguld's? That's part of the fiction, but at the moment of action declaration is fully subsumed within the action - indeed, it's this relative thinness of the established fiction pertaining to Harguld's intention, ie we know nothing more than that he is shooting a bolt at the Gnoll, that permits the subsequent retcon created by use of the trait.
This isn't really aimed at you (AbdulAlhazred) as on this I think we're agreed. But I don't think that either
@clearstream or
@Fenris-77 has said whose intention they are having regard to, when they say that the intention is a component of the fictional position.
Something Baker wrote that I find interesting is -
Fictional positioning is only and always retroactive. You can guess what your position is, and you can plan for your future position, but it's only when you test your position by making a move that you learn whether the move is legitimate
It's hard to see if this means that the fictional position is ever known. One way to read it might be to suppose that at time T I don't know my fictional position, and at T+1 I make a test to establish something about it, so that at time T+2 presumably the outcome of that test is known and thus I know something about my fictional position. The nature of the test is to make a move and learn whether it is legitimate.
Alternatively, it might rule out establishing with certainty any of the contents of fictional position at any time, so that a fact that legitimated a move and thus might seem to have been established at time T+2
cannot be reliably known at time T+3, and I can only
guess at the result of testing something also (seemingly) connected with it at T+4.
Perhaps this is more a comment on the nature of declarations: avoiding assuming that future (and thus not yet known) declarations can be reliably connected with the contents - established or otherwise - of a fictional position. Suppose that a player's subsequent declaration is identical to their first? At T1 "I pick a pine needle from the Christmas tree" which we test and say that it's okay. At T4 "I pick a pine needle from the Christmas tree"... is it no more than a guess that this will be legitimate in the absence of intervening change?
We regularly speak of established or prior fiction, so for now my view is the first one. Through speech acts and tests we gradually establish some known contents of our fictional position. Even if we never establish all of its the contents.
Dro declares Harguld's action, which is to shoot at the Gnoll with a crossbow bolt. Although Dro's literal words are "I put a bolt in his face!", we all know that (i) Torchbearer has no hit location, so "in his face" is mere flavour, and (ii) a player saying it doesn't make it so, so the description of the action as "putting a bolt" in the Gnoll really means "shooting at the Gnoll, hoping thereby to stop and perhaps kill it".
We don't need any deeper analysis of Harguld's mental states to know what the declared action is. Nor do we need any notion of potentiality. We have all the relevant components of the fiction established: a cave, with a mouth and beyond that a shadowy tunnel; Harguld at the mouth, having just launched a bolt from his crossbow at a Gnoll in the tunnel.
Suppose we had a description of fictional positioning that assumed it was a set of facts. Earlier you suggested that Harguld's fictional position is that he is standing, in a cave mouth, crossbow cocked and loaded, waiting for Gnolls. These facts seem to include both imaginary physical facts (imagined tension in the spring arm of a crossbow) and imaginary mental facts (waiting for Gnolls).
What I believe Baker might have been dealing with is that Dro can say something like "H picks a pebble up off the cave floor" - and everyone may well agree that yes, cave floors no doubt have pebbles and picking up a pebble is something H can do. In that light, it seems hard to pin fictional position down to a finite set of facts, rather it has to be thought of as a scene with some known contents and some unknown.
What contents become known? Only those we intend to know.
The question is, what happens next. Whatever answer is given, the established fiction constrains the subsequent narration of that new fiction; no one can change the fact that Harguld is there, that the Gnoll is there, that Harguld has shot at the Gnoll.
But none of that subsequent narration is part of Dro's fictional position when he declared that Harguld shot the Gnoll. Not even if that subsequent narration establishes something that was true, in the fiction, when Harguld shot - such as the Gnoll got in close because Harguld waited too long to take the shot, trying to cunningly lure the Gnoll in.
We must avoid assuming here that Dro's fictional position is completely known at time T when Dro declares that H shot. According to Baker, and my own reasoning, it is not. Whatever is known about its contents at time T omits something that comes to be known at time T+1.
The only known contents of the fictional position are those we intend to know, and at time T+1 it has gained additional or modified contents that will legitimate (or not, e.g. rule out as reaching) Dro's further declarations.
What I've just posted can be explained without needing any account of Harguld's mental states beyond the description of his action: he shoots at the Gnoll hoping to stop or kill it ("put a bolt in his face!"). And adding some richer analysis of Harguld's mental state won't change the way the use of a trait works, the retcon works, and the GM's narration of consequences works.
That's why I think its unnecessary machinery.
Of all the things we can argue about, arguing about what we are allowed to include in our process of understanding seems to me the least appealing. Especially given the nascent state of game studies and ludology.