Closed laiyunfan closed 4 years ago
WE have the original data anyway in the field VALUE. So deleting the brackets is much easier than keeping them, as this would mean: you have to list ALL things that occur in bracketsin the orthoprofile and transform them accordingly. This can of course also be done, but I assume this will take some time to do it. Alterantive is of course: given that it's just 120 entries, one could make a manual representation that is giving the data in correct BIPA, but I think it is better to use orthoprofiles for all data here (otherwise, you'd have to manually add IPA for all datapoints).
Concerning the non-IPA/not-in-Leipzig-rule Old Chinese annotations by B-S (such as in qʰo[n], ŋ(r)aj, etc.), I am reluctant of deleting them, since they are informative, rather than meaningless.
One solution could be something similar to this:
[a] --> a/.
(a) --> a/∅ (with an empty set symbol, not the vowel [ø])
As for the Leipzig-rule symbols, such as "-", "<>", and "." I think we can simply delete them since they are not informing about uncertainty.
What do you guys suggest?