Word frequency and key word statistics in historical corpus linguistics