When I enter a lowercase three-letter ISO code into the general
LANGUAGE tag field for an MP4 file, it is stored verbatim for the file as a whole, but it is not propagated into the proper field for any audio or video stream. That means, there is no way in Mp3tag to see or set any metadata per stream.
I would expect that at least if only a single audio stream and perhaps a single video stream is present in the container, Mp3tag would (also) store the language therein.
I think multiple streams should be displayed like chapters, optionally.