There is provided a device for performing interaction between a user and a machine. The device includes a plurality of domains corresponding to a plurality of stages in the interaction. Each of the domains has voice comprehension means which understands the content of the user's voice. The device includes: means for recognizing the user's voice; means for selecting a domain enabling the best voice comprehension results as ht domain; means for referencing task knowledge of the domain and extracting a task correlated to the voice comprehension result; means for obtaining a sub task sequence correlated to the extracted task; means for setting the first sub task of the sub task sequence as the sub task and updating the domain to which the sub task belongs as the domain; means for extracting a behavior or sub task end flag correlated to the voice comprehension result and the subtask; and means for causing the machine to execute the extracted behavior.