Maxperson's game (and he can correct me where I'm wrong) advocates for actor stance, not necessarily for first person dialogue but for the character behaviour/thought process. So yes at times, probably many, the player will know more than the character about the in-game fiction, as the player dives into the role of the character.
The only definition of
actor stance that I'm familiar with is
Ron Edwards's:
In Actor stance, a person determines a character's decisions and actions using only knowledge and perceptions that the character would have.
This is a particular
mode of or
orientation towards action declaration. It says nothing about
who gets to decice what knowledge and perception the character would have.
Here are the other "stances"
that Edwards identifies:
In Author stance, a person determines a character's decisions and actions based on the real person's priorities, then retroactively "motivates" the character to perform them. (Without that second, retroactive step, this is fairly called Pawn stance.)
In Director stance, a person determines aspects of the environment relative to the character in some fashion, entirely separately from the character's knowledge or ability to influence events. Therefore the player has not only determined the character's actions, but the context, timing, and spatial circumstances of those actions, or even features of the world separate from the characters.
When the RuneQuest books say that players should have their PCs cooperate because that is necessary for the game to be fun, that's an example of advocating
author stance - that is, making decisions for one's character based on real-world (or, if you prefer,
metagame) priorities (in this case, having fun playing the game). I would be absolutely gobsmacked if there is not quite a bit of this in [MENTION=23751]Maxperson[/MENTION]'s game. Even [MENTION=29398]Lanefan[/MENTION] on these boards, who is more purist than Maxperson about the issue of "artificial" cooperation between PCs, has told anecdotes of doing stuff with one's PC because it's fun/exciting in the real world. Which is to say that even [MENTION=29398]Lanefan[/MENTION] plays in author stance from time-to-time.
Every time a player has his/her PC pick up on an "adventure hook" because it seems like it would be more fun in play than hanging out at the thermal baths, we have author stance.
It seems likely to me that players in Maxperson's game adopt director stance only during character creation - for instance, they are not allowed to introduce new backstory elements during the course of play. But deciding that my PC knows about trolls isn't an example of Director Stance. Deciding that s/he knows about them because Old Uncle Elmo told him/her is Director Stance
only if the existence of Uncle Elmo and his tale telling hasn't yet been established as part of the fiction.
Deciding that my PC wants to use fire to attack these trolls
because I want to get the encounter over and done with would be an example of author stance. If I don't offer some story to explain
why my PC wants to use fire, then it's pawn wtance.
If I don't
know very much about what my PC does or would know - which is very common in D&D play (eg consider the starting set up for most classic modules) - then most action will be pawn stance for the simple reason that my PC
has no knowledge, perceptions and motivations outside mine as a player knowing what the game expects. Conversely, to adopt actor stance in playing my PC I need a reasonably rich sense of what my PC knows and wants. Normally that would be established before play (if it's established by the player
during play then we're back into author stance). No version of D&D I'm familiar with says that this is the sole prerogative of the GM: classic D&D (OD&D, B/X, Gygax's AD&D) are silent on this matter, while 4e clearly says that the player establishes background and can establish quests for his/her PC. As I already posted, perhaps the 2nd ed AD&D DMG says that this is the sole prerogative of the GM - that's not a book I've ever read.
From Maxperson's PoV, he is not gating anything or playing a degenerate form of MMI. He, as referee, is ensuring that everyone follows the roleplay in actor stance.
He, as referee, is deciding
what the PCs do or don't know. That's not holding players to actor stance (assuming you're using that phrase in the only way I'm aware of it ever having been defined). That's
deciding what it is that the PC knows, and treating that as the GM's exclusive prerogative.
Whether or not it's "a degenerate form of MMI" isn't something for me to judge - presumably it's flourishing at Maxperson's table. But it's clearly a very strong form of GM gating.
EDIT:
when the attempt involves uncertainty about whether or not the PC knows something, the DM has the authority to adjudicate it and call for a roll.
But, as per the quote that follows, you appear to take the view that there is such uncertainty in respect of
anything that has not actually been revealed in play.
You're right about the DM not having to adjudicate things recalled that involve what the PC has encountered during game play. It's the stuff that isn't from game play that's at issue here.
If a player is not entitled to impute any knowledge to his/her PC other than what has actually been encountered in play, moving beyond pawn stance will be very difficult.
(Which is certainly a very traditional way to play D&D: modules like ToH, Keep on the Borderlands, the Pharoah series, Isle of Dread, all the ones being republished in Yawning Portal and by Goodman Games, etc assume pawn stance as the default: that is, that players will make certain choices because
that's how the game works, and the question of why the PC would make that choice isn't expected to be raised, let alone answered.)