Having worked with linguists on an ads eval platform (human eval), I can definitely appreciate what they've put into formulating questions/answers/documents/etc, such that these do not sound ambiguous, and not only that.