The system then checks whether that word or phrase is already known, and if it is not then in step S3, it prompts the user to enter the same word or phrase via the microphone 7, and associates the utterance with the corresponding text input in step S1.