tresoldi
commented
6 years ago As a general rule, I'd only add them if they are provably distinctive. It boils down to our common problem of researchers marking phonetic properties when what matters for historical linguistics is almost always phonology. If the feature cannot be explained in terms of complementary distribution, context, etc., I'd leave them at the side or just carry them on while having no weight in the alignment.
2018-01-28 8:36 GMT-02:00 Johann-Mattis List notifications@github.com:
When testing on the phonetic alignment benchmark http://alignments.lingpy.org I found many ultra-short consonants as well as mid-long consonants. I guess this is dialect data which is often very fine-grained. The question is, whether we want to add these to the diacritics, or leave them at the side for the time being.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/cldf/clts/issues/97, or mute the thread https://github.com/notifications/unsubscribe-auth/AAar9zlfx7gDPSrner8BUabknE5GYbwmks5tPE2mgaJpZM4Rvpgt .
LinguList
When testing on the phonetic alignment benchmark I found many ultra-short consonants as well as mid-long consonants. I guess this is dialect data which is often very fine-grained. The question is, whether we want to add these to the diacritics, or leave them at the side for the time being.