PURPOSE: To improve efficiency for correcting a result of recognition by combining and utilizing certainly and voices.

CONSTITUTION: The result of recognition is not only read out as the result of speech synthesis but the output of the speech synthesis is changed corresponding to the exactness (certainly) of the result of recognition. Namely, corresponding to the degree of certainly, speed or the volume of voices to read the speech synthesization is changed and informed of a user. Concretely, corresponding to the certainly of the recognized result, a display method is changed such as inverting characters, the shape of a cursor is changed or in the case of uncertain characters, compressed character images are displayed at suitable positions.