And what about that could a character in that world not do? The use of authoring here in our discussion implies the player does something the character could not do.
Let's unpack this. The character never does anything at all. The character is imaginary. The player is doing everything. You can put constraints on the player's options, like acting in character, but this doesn't change this fact. And, because of this, everything a player does that results in a change in the fiction is authoring fiction. Granted, in many games, this is gated or shared authoring, but this isn't talking about if authoring fiction occurs, but rather who has what authority to author a thing. In effect, we're discussing the different authorities to author.
Your approach is that the player only has the authority to author action declarations and statements about the internal feelings and thinking of the PC. However, even this is gated, because the GM exercises veto authority over even these things, having the authority to declare an action declaration or statement about thinking to be incompatible with the GM's conception of the fiction. This means that there are many constraints on a player's ability to author fiction, but it doesn't change that what they do is author fiction.
In other games/systems/approaches, the how fiction gets authored is moved around, but it's rarely without it's own constraints. These constraints look different from your approach though.
Let's look at a classic example: Searching for a secret door. The scene is that the PC (or PCs) are fleeing from some guards, and, for whatever reason, have ended up in a dead end. Desperate, they search the dead end for a secret exit.
A) in your approach (and I'm only saying your here as a differentiator, I use this as well when I run 5e and similar games), the PCs do the search -- the player has authored this into the fiction. The resolution of the search, however, is entirely based on the GM's conception of the fiction. Normally, the GM would have determined pre-play if a secret door was present in this dead end (which was also likely determined pre-play), or would use their prep to inform their decision as to whether or not a secret door was present here. Regardless, the presence or absence of a secret door would be entire up to the GM -- that the players have searched for a secret door will have almost no bearing on it's existence in the fiction. The only thing that will happen here is that the GM will consult their conception and, if a door is present, will ask for a check to see if it is located. I'm ignoring that a GM might ask for a check even if it's not because that's not important to the point -- it's just some theatrics to further hide information from players.
B) in some other approaches, the question of a secret door is looked at differently. Since there is no pre-play determination of fiction, the question becomes 'is it possible a secret door may be here?' The answer to this is yes, unless something specific in the fiction or genre would prevent it (like it was determined in previous play that no secret door exists in this dead end). So, yes, it's possible, but we don't know yet. This would be resolved via a check. A success would reveal that a door does exist, a failure would add complication to the scene or pay off a consequence. Here, the door's a question for everyone at the table.
So, to get back to authoring, in A, the players are still authoring into the fiction the actions their PCs are taking. This is the hard limit, though, as the result of those actions is entirely up to the GM to author. In B, however, the very act of declaring the action also defines part of the solution space. So, the player is still authoring the action of their PC searching for a secret door, but also adding the potential that a secret door exists. On a success, the fact that a secret door exists can be said to be because the player authored that bit of fiction, even though it was mediated through the mechanics of the game.