I think when you get into the MMOVN thing, the 'VN' part is just the fact that the information given to the player is provided in text form, in a controlled manner. However, once the player is aware that there are players on the other side, it becomes more like a game, practically speaking. But you can play a VN like a game (giving yourself objectives) if you want to.
The difference is whether the author pre-created what's going to play out, or if the players themselves are going to create the emergent story. The latter would be a text-etc-interface multiplayer roleplaying. It could certainly feel like a VN because of the narrative and how each player sees the information, but I wouldn't call the whole thing a VN. For prose generated from an underlying game system, what would matter is how much of what happens determined by the author. If it's just the order switching around a bit, then it'd still be author conceived. Otherwise it would be emergent.
(If you call that type of presentation VN-style, then yes, the game would be a VN-style game, you might even think to call it a VN. But that would overlap with the current notion which has that VN's are something which authors conceive of).
I suppose an examples of that would be a game where all you do is explore this big manuscript that you view within the game, and when you go to different places you hear different sounds and stuff.
I have no idea what to call this. It's not quite the same as a set of real letters because you can incorporate sound and visual effects into it. Does controlling the order the player reads parts of the story make the work a VN or not? If you had a series of episodes viewable in any order, that would still seem like a VN. What about if you interacted in a point and click fashion to access the episodes? What about if you walked around a 3-d environment to access those episodes?
It's just a categorization problem. You wouldn't be able to make a classification for every conceivable thing without being very minute and detailed. You can pick a relatively easy way to categorize works, or you could end up aiming for a categorization that's actually impossible, because our notion of what makes a VN would turn out to be contradictory once you explore the array of things which are possible using text and interaction.