Re: THEORY: final features, moras, and roots [was: it's what I do]

On Thu, 5 Oct 2000, Marcus Smith wrote:

> Nik Taylor wrote:
>
> >Also, what about the voicing diacritic?  Based on my (admittedly scanty)
> >knowledge of Japanese, it appears that voicing was perhaps once
> >allophonic?  If so, what was the original conditioning factor?
>
> Part of voicing still is allophonic, but not all of it.  Originally there
> was a restriction that there could only be one voiced consonant per
> word.  That condition doesn't hold anymore due to influences from Chinese
> and English among others.  The allophany you're referring to (I think) is
> what the Japanese call rendaku.  When native Japanese compounds (or words
> that are felt to be Japanese rather than borrowings), you voice the first
> consonant of the second element.  Thus kana > hira-gana.  There are lots of
> conditions on when this can apply, but the short of it is that the voicing
> in these contexts is allophonic to this day.
A friend of mine at the University of Arizona wrote her dissertation
on the Tohoku dialect of Japanese, which has some very interesting
voicing phenomena. Essentially, everywhere Standard Japanese voices an
obstruent (whether the result of a rule or underlying) Tohoku has a
prenasalized stop. Elsewhere, if a voiceless stop occurs
intervocalically, it is voiced (unlike Standard Japanese, where they
remain voiceless).

She was exceedingly retiring, so I don't know if her dissertation will
ever be available to a wider audience. I thought it was absolutely
brilliant.

Dirk

--
Dirk Elzinga
dirk.elzinga@m.cc.utah.edu