Yep, we would have to have minimal pairs, but also (morpho-)phonological processes that somehow change one into the other.sangi39 wrote:Oo, now that looks interesting, but then you've got to think about what conditions any alternation might appear in, otherwise, would it be meaningful to talk about "correspondences" between short and long vowels at all?Creyeditor wrote:What if length contrast is phonemic and the short vowels are more marked, so there are fewer of them?
Combining both ideas:
/i: ɪ ɯ u:/
/e: ɘ o:/
/ɛ: ɔ:/
/a a:/
With the following correspondece:Code: Select all
Long short /i:/,/e:/ /ɪ/ /u:/,/o:/ /ɯ/ /ɛ:/,/ɔ:/ /ɘ/ /a:/ /a/
We might talk later about this, but this could also be related to stress, with e.g. both quantities occuring in stressed syllables but only short ones in unstressed, which of course implies that sometimes stress changes somehow.