UD_Portuguese-Bosque
UD_Portuguese-Bosque copied to clipboard
sentences with missing nsubjpass or nsubj
we have 9128 sentences, do all of them have subjects? if we add 1 dep name="nsubjpass" to the 11833 nsubj we have all subjects, but this does not guarantee that every sentence has one subject. (and does nsubjpass correspond to subject of passive verbs? if so, 1 seems far too little)
Can we find a way to check that every sentence has at least one subject and if not, which ones do not?
I know that sentences such as "Chovia forte naquele dia" can be considered without subject by some, but maybe this is a way to check for bad trees.