AS of writing, only 116 sentences with dependency annotations
Using UD 2.0
Bambara is predominatly isolating
The Daba analyzer tool was used to create the original Bambara Reference corpus
Morphological features generated by looking at both the glosses and the morphological breakdown in CBR (the reference corpus).
compounding and derivation not treated productively so lemmas are not split into compound components
Original reference corpus did this thing where it had multiple POS tags in cases where the POS was ambiguous. These were resolved using largely manual methods.
All copulas were annotated as verbs? Weird choice not to have them as aux.
Topicalization involves resumptive pronouns in Bambara
No Comments