When my players elect to explore eastward, for example, they find the foothills for the mountains the can already see. In those foothills they will find new challenges.
<snip>
If I did that, for instance, then I would foreshadow that fact prior to the question coming up, as in "the gate is guarded by Bob the Guard, and you, Bob the Rogue, know from your underground contacts (established by Bob the Rogue at character creation or during play) that Bob the Guard is generally held to be unbribable. He is, however, known to be rather dim."
So, for me, a DM who uses heavy prep, the goal of play here would never be for the players to declare actions to find out my notes on bribing the guards, they'd either already be told in framing what to expect or they could try and I'll let the mechanics determine things.
<snip>
Let's assume the case where the DM has notes that say the map is in the kitchen. The players don't know this, and so they have hope that they can find the map. This is negated, yes, but that doesn't negate the map as an objective of play, it just says that the fictional positioning is not yet right and they need to try again elsewhere. This isn't all that dissimilar to having to get past the guards before having a chance to search for the map. Again, broader play where agency is spread out vs narrower play where every scene focuses on the maximum agency moves.
To me, this all speaks of very heavily GM-driven play.
The GM has established that there are foothills. The GM estabilshes the material for the challenges in said foothills. The GM estabilshes the personality of Bob the Guard, and determines what it is that Bob the Rogue's contacts have told him about Bob the Guard. The GM establishes where the map is, and the players have to work ther way through the GM's pre-established fiction until they make the right move to find the map.
As you are describing these things, I am not seeing a great deal of player agency over the shared fiction.
The choice to go to this room to search instead of the other room to search is agency if it comes at a cost, regardless of outcome. If there's a time limit, or chance of guards appearing, then the choice to move to this room and search has a consequence, and therefor agency as the players spend a limited resource (in this case time or danger) to achieve a goal.
<snip>
The guards arriving are a check made by me based on the events ongoing -- make a lot of noise, check gets a plus. Quiet, check gets a minus. Near the guard post? Plus. Crawling through airducts? Minus.
From the OP:
In classic D&D, the dungeon was a type of puzzle. The players had to map it, by declaring moves (literally) for their PCs. The players, using their PCs as vehicles, had to learn what was in there: this was about inventory - having enough torches, 10' poles, etc - and about game moves too - searching for secret doors, checking ceilings and floors, and so on. And finally, the players had to try and loot it while either avoiding or defeating the monsters guarding the treasures and wandering around the place - this is what the combat mechanics were for.
The game is something of a cross between a wargame and a complex refereed maze. And *worldbuilding* is all about making the maze. I get that.
Perhaps your game is something like that? So player agency is really in beating the maze/puzzle?
It' s not clear, though. Eg in classic D&D, the "time limit" comes from a known-to-the-players wandering monster mechanic, plus known-to-the-players rules about torches burning down, light spells running out, etc. So the "time limit" isn't anything like a story element - it's a parameter of the puzzle that has to be taken into account in arriving at a solution.
Here's another way a "time limit" can work: (i) there are 5 rooms; (ii) you can search 3 of them for the map; (iii) then you're done. That doesn't seem to allow much player agency - it's just a gamble. Suppose there is an extra rule: use your Passwall resource and you can search an extra room. (In the fiction, the spell creates a shortcut.) That's a very modest bit of agency, but not as much as in classic D&D.
It's not clear to me what you haVE in mind when you talk about a time limit. Likewise with the guards, which perhaps are like wandering monsters but perhaps not.
There's absolutely nothing wrong with defining agency as the ability to introduce new fiction through action declarations
<snip>
The player expressing a desire to introduce new fiction that is actually introduced when agreed to by the GM or when the dice indicate success of that introduction isn't a different thing that the player being able to introduce new fiction through action declaration. Also, your phrasing fails to account for the fiat introduction of Jabal of the Cabal as part of an action declaration that is accepted regardless of the outcome of the mechanics
As far as Jabal is concerned, a long way upthread I already distinguished between two different Circles declarations:
* Jabal is a leader of my cabal. I reach out to him.
* As we travel, I look for signs of any members of my knightly order.
The first involves establishing a new element of the fiction; the second does not. Establsihing a new element of the fiction is not a canonical element of a Circles check.
And I do not define player agency as "the ability to introduce new fiction through action declaration". I have repeatedly talked about "player agency over the content of the shared fiction", and have pointed to multiple ways that can be exercised. Action declaration is one. Providing material for GM narration of framing and consequences is another.
The exanmples you gave, that I quoted above, illustrate this. How does the GM decide what challenges will be found in the foothils? Who decided that foothilss would figure in the game at all? Why has the GM framed a scene with Bob the Guard, and established that Bob the Guard can't be bribed? Who made acquisition of the map a goal of play?
At least as you describe these things, I get an impression that you regard all this as very GM-driven. Here's a furtehr cause of that impression:
Where do you get your framing? I don't know, it's largely unimportant. You can make it up on the spot. You can use things that tie it to your characters for future links and challenges, you can read your notes that you prepped and are still useful. What does it matter?
In the "standard narrativistic model", the source and content of the framing is hugely important. The players build characters; those characters have dramatic needs; the GM's job is to frame those characters into situations that will speak to those dramatic needs and provoke choices, which lead to consequences, which feed into further framing, and so on until the game is done.
This is an important mode of player agency in respect of the content of the shared fiction, which has nothing to do with action declaration.
the players lose agency by having the GM force them into crisis of the GM's choosing. That the GM references the notes on their characters before doing so doesn't mean the players suddenly have agency at being framed into a crisis of the GM's choosing.
Are you envisaging the players framing their own situations? That really does sound like "collaborative storytelling".
But in any event I simply can't agree. A player who signals "I want to play a game in which my PC tries to free his brother from possession by a balrog, starting with the attempt to find items that will help that task" is not having his/her agency negated by the GM saying "OK, you're in a market, and a peddler is offering to sell you an angel feather." That player's agency is being affirmed!