A means for producing spoken messages in response to signals from a data processing or control system. The selection of messages is produced by assembling speech sounds from a set of fundamentals that may be artificial in origin or may have been extracted by processing human speech.