Translating introductions and wishes in audiovisual dialogue. Evidence from a corpus