You are missing the point of my argument.
To have a more realistic language system -- anything more complicated than what we have now -- requires tying those languages to probably a combination of region and people. You can't have a French language without the French people and a part of the world called France (not going to nitpick about demonyms).
Similarly, you cannot have a Pomarji language without a Pomarj people and a part of the world called the Pomarj.
This is why your approach doesn't work. It isn't generic, and almost by definition cannot be. Current D&D languages are generic. That they are bland and not-quite-sensible is not something I'd argue against.
Yes. The languages will not be generic. The SYSTEM is generic. The LANGUAGES are setting-specific.
Not sure how to make that more clear. And it's a thing I've been saying the entire time.
This isn't where the complexity expenditure bugs me. The excessive complexity comes from having to track all these relationships. If you're trying to model languages at all on par with real world complexity, or even say half that, it's going to be a table lookup every single time you have a social interaction.
Again, you're trying to say "Like real world complexity" and as you can see in the example: IT DOESN'T.
Not only does it not reflect real world complexity in the slightest, it invents the "Romance" language as an intermediate regional proto-language somehow separated from Latin. I did that -specifically- to avoid getting into the actual complexities of West versus Southern Germanic and the whole structure of Proto-Indo-European.
Is it more complex than "Each race has their own language?" Yes. Is it "Real World Complexity?" Not remotely. Not -remotely-. Like. I'm suggesting people speaking French and Italian, which are -very- different languages for their similar linguistic origins, would be able to mostly understand each other without taking the time to learn the other language. But things like subtle threats or unique turns of phrase require insight checks to figure out.
That's not remotely trying to model languages to real world understanding.
Essentially I'm looking at something like what Pathfinder has for their human languages. 2-3 ancient languages and writing systems, and then the languages that came from them use those writing systems and are kinda similar but not exactly the same.
Give me an example of a game being run where your take on language wouldn't come up.
Curse of Strahd. Everyone speaks Barovian.
Unless you decide to have some NPCs speak other languages, like the Vistani speaking Vistana. But they already do that in Curse of Strahd so it's not much different.
It's another thing to track. And I can imagine said person saying something like "What the [bleep] is Pomarji and why do these orcs speak it and not these other orcs?" It's a fiddly detail, and those add up.
And like people do with Language for the most part and Encumbrance almost universally you can just ignore it if you don't like it, too.
Shocker.