I think that if you say: dependency treebank, most computational linguists will expect syntactic dependencies. I agree that the notion is blurred sometimes, e.g

I think that if you say: dependency treebank, most computational linguists will expect syntactic dependencies. I agree that the notion is blurred sometimes, e.g. by including word senses or preferring semantic attachments in some cases over semantic attachment. There may also be treebanks of the semantic-y kind, but I haven't seen or used them often.