抄録
Ramsey & Unger (1972) proposed pJ voiced obstruents *b, *d, *g, *z that lenited to w, y, Ø, s ~ Ø at some pre-OJ stage, distinct from both voiceless *p, *t, *k, *s and later prenasalized voiced obstruents. In her review of Unger (2008), Robbeets (2009) calls this proposal “the classical lenition theory”, which I had posited as a sufficient basis for reconstructing a pJ velar nasal phoneme *ŋ. Here I propose an amended lenition theory: proto-Japanese had just one series of obstruents but also had a rule of automatic intervocalic voicing. This modification accommodates objections to the “classical theory” without significantly diminishing its explanatory power, including the reconstruction of a pJ *ŋ, and has important implications for further research into the Korean-Japanese relationship and other aspects of Japanese language history.
https://ror.org/00rs6vg23 
