This paper studies the development of quasi-nominal particles such as no, ga, gan, to and na in Japanese dialects, through an analysis of language maps and dialectal data from the past. A survey of some of the language maps in Grammar Atlas of Japanese Dialects shows that no is widely found in central Japan, ga in some outer areas like Hokuriku and the southern part of Shikoku, gan in Niigata, na on the Japan Sea side of Tohoku and to in Kyushu. This distribution indicates that no is new and that ga and gan are relatively old; to and na seem to be unique to their respective districts. In the history of Standard Japanese, it is said that (1) as case particles, ga hosts a noun which is familiar to the speaker and no hosts one not familiar, (2) grammatical differences between the two have gradually become dominant since the Muromachi era. Ga functions as nominative case, no as genitive case; and (3) quasi-nominal particles ga and no developed from these case functions at the beginning of the Edo era. Based on dialectal data in the Edo era we can say that the same kinds of quasinominal particles are already found in those areas mentioned above. This developed in parallel with those of Standard Japanese. Thus, ga in the outer areas is derived from its use as a case particle almost independently of Standard Japanese when (3) is realized under the condition of (1), no in the central area occurred when (3) is realized under the condition of (2). Gan in Niigata is a combined form of ga and no. To in Kyushu is possibly derived from its use of a case particle. Na is supposed to derive from the auxiliary verb nari to express assertion.
View full abstract