Hey, Moonplum! I really loved reading your manifesto, you really did tell things current language deduction games don't have that should.
Reading your manifesto reminded me of a game I'm making. It's called RGSA, and it's about, well, learning a new language. I currently have it on my itch page as I'm typing this, but I'll make it private because I need to add in a page and make new puzzles to replace things I think should be further into the game.
Furthermore, I should explain what RGSA is about.
You find a dusty book on a shelf inside an abandoned basement inside your house. You grab it, dust it off, and place it on the table next to the shelf. You open it, read it carefully, and you notice that whatever you're dealing with in this book is a language. It's even written on the first puzzle: "WORD".
RGSA is a paper puzzle rule discovery (it's more language discovery, but "rule discovery" works here too.) game about learning a new language. The way you do it is simple: Just draw a line passing through letters. Just that.
The goal of each puzzle is to match the image to the solution word ("WORD" puzzles) or make a sentence in which the solution matches what's happening on the image ("SENTENCE" puzzles). You may also learn a different aspect of the language ("SUFFIX" puzzles).
There's no feedback. There's no writing of the language (except for Puzzle 3/1-3 which is a SENTENCE puzzle and the SENTENCE puzzle tutorial). There's just you, the book, the images, the language, and the lines you draw.
I feel like even though you have multiple points, RGSA might've broken some of yours. Here's the points I think it breaks (from most definite to least definite):
RGSA breaks Point VII. I just wanted to include this because it's aligned to what I made this section of the message for and also because I just wanted to LOL
RGSA breaks Point VI. Every "WORD" puzzle is just a 1:1 mapping of an existing word in English. I mean, the first words in the game are literally nouns. How is that not a 1:1 mapping of English? I feel like giving more context about what the language is about would spoil the game a bit, because primarily all you were given was a language and no context.
RGSA breaks Point I. Even if my intention for RGSA was to give the game a bit of a mysterious feel, like "Why do I have a book with this random language? What is the purpose of it?", basically every single image in RGSA's puzzles can be described using English and perhaps the language too. I wanted to try to make something original (I even used some of RGSA's ideas in a headcanon for a different conlang!), but even then, I feel like there's just way too much English packed into a dusty book.
RGSA breaks Point IV. As said in my previous "RGSA breaks Point N", RGSA's images in its puzzles can be described using English, so can the language itself. It's a reason why I think that some parts of RGSA's language just feel like English if it had a tie dye-looking face (it's the best I can describe it as, LOL). Though these "tie dye face" parts only start appearing near the halfway point of RGSA, so I don't think I can talk much about it.
Either way, I feel like RGSA has a bit of what you want it to have. You are more than free to talk about what you think of RGSA breaking said points, all our words matter. If you're interested in RGSA, you can check the demo (which is what is on the itch page rn) after I update it! You can add it to a collection if you want, though I'm not so sure about that part (I saw RGSA was in a collection before).