W dniu 2012-05-01 15:41, Anders pisze:
> Depending on whose analysis we follow. As I recall, Griffith (Die
> Sprache 45, 2005) assumes that the Late Common Celtic form was *-uns <
> *-u:ns < *-o:ns (with analogical *o: from nom.pl. *-o:s and gen.pl. *-o:m).
Yep. Kim basically follows Griffith except that he takes *-o:ns to have
been the "last common ancestor" of the IE acc.pl. endings (on the
combined evidence of all the branches), so there's no need to invoke
analogy. There was once Ronald's paper will be published in the
proceedings of the 2nd "Sound of IE" conference in Opava, but you can
download it from his Adademia.edu page. Jens Rasmussen also argued for
PIE *-o:ns in one of our discussion here... oh, God! almost ten years ago:
http://tech.dir.groups.yahoo.com/group/cybalist/message/15770
(and the rest of the thread)
Piotr