Thank you!Salmoneus wrote: ↑03 Jun 2020 14:54Well, it's best to think of this sort of thing not as a phoneme changing to either X or Y depending on environment, but as the product of two discrete changes, one of which was context-dependent.
So, there's really two possible questions for you to look at here: what would cause /o:/ > /u:/ (remaining /o:/ then breaks); or what would cause /o:/ to break (remaining /o:/ then raises).
I think the latter is more likely, probably. And why would /o:/ break? Well, the obvious culprits would be a following low vowel [ko:ta > koata > kuot] or less likely some sort of following coda consonant (ko:ht > ko@ht > /ku@ht > /kuoht/). You could also have /uo/ come from /oe/ - ie from a fronting or derounding (ko:ti > koeti > kueti > kuoti). And you could have breaking be due to a preceding raising element (jo:ta > juota)
With the plain raising of /u:/, the most likely thing would be an adjacent raised element (palatals, palatoalveolars, approximants, etc), before or after the segment (ones before could alternatively trigger breaking). However, /u:/ is usually also more rounded than /o:/, so it could be an adjacent rounding element. Or, of course, some sort of vowel affection or harmony (ko:tu > ku:tu).
Finally, in both cases, issues around stress and syllable weight may be relevant. Particularly if /o:/ > /u:/ is a merger, which could easily be more common in unstressed syllables. Breaking could happen only in stressed syllables (perhaps via an overlong allophone). It's less likely but conceivable it might only happen in unstressed syllables for syllable shape reasons (if long vowels are only allowed in stressed syllables, unstressed /o:/ could break into bisyllabic /o.o/, which could then remerge as /uo/).
Those are very similar ideas that I thought.
At last, I was boring and took the Italian/French sound change and conditioned it just with openness of the syllable.