Press the captions and listen to the pairs of words. Classify them to three columns.