@danderson Unicode addresses that in terms of "it contains all the characters". Yes.

But it doesn't address it in terms of "which character to use" as it's locale specific and unicode is just a character set (which is locale-agnostic)

I could imagine that ICU though has some features for that.

And yes: It's always based on the general lamguage. Not on the quoted language

@danderson Is that based on the language of the main text or of the quoted text?

/me runs for cover