(12-03-2025, 09:19 AM)oshfdk wrote: You are not allowed to view links. Register or Login to view.Roman numerals are not a natural language. I assume Frigorifico is primarily talking about natural languages in this thread, since the original post starts with:
Ah, I see. So the question is rather: does it become more, or less language-like if we assume positional allography?
I know that when transcribing a medieval text for literary purposes (so we can read the story), a lot of allographs are either not represented in the transcription, or glossed over by the reader. Elaborate forms (initials, top line ascenders...) are merged with their regular counterparts. The allography of capital letters is maintained, but we know they have the same phonetic value as their counterpart since we understand the writing system.
This is different when doing a paleographic transcription, which does not flatten these details. The EVA transliterations are most definitely of the more detailed kind. But we kind of treat them as if they are literary transcriptions, religiously maintaining distinctions even though LAAFU/PAAFU effects only make sense as allographs.
Unfortunately, it's all very tricky and hard to test, since it isn't very clear which glyphs may be variations of one another, and what we'd do with the system once hypothetical allographs have been flattened out.