Where can I find a labeled dataset of sentences classified by type (interrogative and declarative)? I'm particularly lookings for Russian sentences, but other open and labeled datasets of other languages will be helpful.

Basically, I need a dataset of sentences, which not consist only of question words, and sometimes don't have a question mark. It will be good, if dataset will have some declarative sentences with question mark (e.g. 'The bus has already left?')

  • 'The bus has already left? still seems like a question to me. You are asking for confirmation that the bus has already departed.
    – user4293
    Commented Mar 6, 2017 at 10:39
  • @JanDoggen Yes, the point is that this sentence is a question, but in an unusual form, so perfect dataset should also contain such sentences.
    – Amir
    Commented Mar 6, 2017 at 13:19
  • Can't you then classify sentences yourself if they end in question mark?
    – philshem
    Commented Mar 6, 2017 at 14:35
  • My point is your question text: don't give that as an example of a declarative sentence. Please edit
    – user4293
    Commented Mar 6, 2017 at 15:00

1 Answer 1


Mark Weisser's SPAADIA corpora consist of 35 transactional dialogues annotated at the sentence-level with function/mood (declarative, interrogative, imperative, etc)

He's also got a nice annotation guideline for expanding on this schema. martinweisser.org/index.html#spaadia_v02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.