When I first played Loved, I couldn't understand the art style - all black, with red pixels as the enemies, and a soft blocky green cloud surrounding my avatar. It took a couple more playthroughs for me to understand my control over the world - each decision to obey of disobey was a vote for the aesthetic style of the game. Despite the initial confusion, the mechanic has made me think a lot more about 'uncovering' design choices in linear games.
By acting 'constantly' in Loved, your world is shaped in a way that is more beautiful (or at least more unified) than it would be if your choices sometimes agreed and sometimes disagreed with the text. Whether a squirming pixalated platformer or a detailed yet morbid limbo-esqe is preferable to the player, the world itself provides no consistent appearance to even be evaluated unless the player submits completely or revolts entirely against the narrator.
So why is this innovative? Platformers and other narrative-based games operate under similar pretenses, whether purposeful or not - Mario will see only the first world if he disobeys his premise, and Chell will never unfold the broken beauty of Aperture Labratories if she does not perform her tests. Obtaining a game's objective is synonymous with unraveling a cohesive world, and to some degree, unraveling the world rewards a player for aligning their goals with the characters, and therefore participating in gameplay. This concept is fundamental to most narrative games, and could even be said to be a reward mechanic that is effective enough to replace a cohesive narrative altogether (think the impossible game or battletoads).
The innovate switch appears here. If the game takes survival of the main character to be axiomatic (a fair assumption for a game using the pretenses of platformers), a unified world is never made whole. Sometimes, the voice tells you to collect checkpoints, and other times it tells you to die, even after exceptionally difficult obstacles. If the player aligns with the goals of the character, we actually never get to see the world fully abstracted or realized. It's not only that playing the game 'correctly' doesn't lead to a aesthetically pleasing world, but that either of the these two worlds arent even put fully in front of the player to view or choose between - and unless you're replaying, once you make an in-game choice that doesn't align with your previous decisions regarding the text, you'll never get to see the full art style, even if you try to be consistent with the rest of your actions.
Loved made me reconsider exploration, design and visuals as a reward mechanic because it demonstrates the problems that occur when it's not implicitly a reward. Because we're so accustomed to an evolving, slowly revealed world to be indicative of progress and success, our own control over it leaaves us confused, without a strict sense of success or progression. Loved lets consistency take the role of revelation - and in this way, progress in the game is left more abstracted.
Good point! Usually doing the 'wrong' thing in games doesn't let you progress, but in this case it can completely the aesthetics of the game. Never thought about it that way!