Abstract
It is said that new vocabulary items in contemporary Japanese are more likely to be loanwords (especially from English) rather than Sino-Japanese words. While this is true if attention is limited to Sino-Japanese words written with two characters, Sino-Japanese words written with three or four characters are still created quite productively. Such longer words are commonly used but have not attracted as much attention as two-character words, in part because many of these longer words are not listed in dictionaries. This study investigates the formation patterns of four-character Sino-Japanese words and the diachronic changes in those patterns during the Edo, Meiji, and contemporary periods.