The present invention relates to a word sequence output device in which emotional synthetic speech can be output. The device outputs emotional synthetic speech. A text generating unit 31 generates spoken text for synthetic speech by using text as a word sequence included in action command information in accordance with the action command information. An emotion checking unit 39 checks an emotion model value and determines whether or not the emotion of a robot is aroused based on the emotion model value. Further, when the emotion of the robot is aroused, the emotion checking unit 39 instructs the text generating unit 31 to change the word order. The text generating unit 31 changes the word order of the spoken text in accordance with the instructions from the emotion checking unit 39. Accordingly, when the spoken text is "Kimi wa kirei da." (You are beautiful.), the word order is changed to make a sentence "Kirei da, kimi wa." (You are beautiful, you are.) The present invention can be applied to a robot outputting synthetic speech.