Interesting! I wish there was more, but obviously there's only so much you can do within the confines of a field methods class.
There isn't much info on Anyi online, though I did find this paper
, which has a fair amount on the pronouns, and includes some asides on ATR-harmony, tone, etc. It's still unclear to me how important tone is, though it seems to play a more important role than you guys were able to determine. Apparently the few instances of vowel length are connected with tonal stuff too; i.e., syncope of an earlier vowel in CVL sequences leads to a long following vowel with combined tones, so blàá
"woman" is bàlá
in some dialects, apparently.
You also missed that the 2sg distinguishes subject and object forms as well, not only the 3sg (subject ɛ-
, object (w)ɔ́
-- as shown e.g. in (45) mi kluɔ
, "I love you" [cf. 177, 178a, 276-277a, 390a, etc.] and (105) mi li ɔ flua ɡɔ̃
, "I have a letter for you"), the object forms being the same as those used in possessive constructions and so on (with certain aspects, however, the subject pronouns apparently all take low tone while the object pronouns retain high tone, so in those cases subject/object is distinguished for all person/numbers). Interestingly, Burmeister says reflexives are formed with the object pronouns and the noun ŋwṍ
, but your data very clearly demonstrate the construction with bɔbɔ̃
, with numerous examples.
There's also a paper from the 70s on Anyi serial verb constructions here
. One interesting thing I noticed on skimming it (reeeeally briefly) is what is probably a dialectal difference; while your informant has adua
for "dog", Van Leynseele records "cụ̀á" (= tʃʊ̀á