To obtain a printer in which occurrence of an error can be recognized and dealt with quickly and readily while simplifying and making inexpensive the arrangement.

A voice output section prepared for each state of error, or the like, and converting the storing part and the digital voice pattern data of a character string group for voice data represented by a standard inscription method into voice is provided in a printer (10) whereas a monitor (5) and a voice synthesizing section (6) for converting a character string of aforementioned format into digital voice pattern data is provided in a host unit (1) and then these units are connected through a bi-directional interface. Upon occurrence of an error, a character string group for voice data corresponding to that content is selected and transmitted to the host unit (1) which analyzes the character string and sends back digital voice pattern data subjected to setting of accent, intonation, or the like.