Re: Vowel harmony part 2

Rob Haden <magwich78@...> writes:

> Hey,
>
> Thanks for your knowledgeable answers. :)  I've done a little research
> myself, and I've found that a simple triangular system of i-a-u can become
> a much fuller vowel system, with some sound changes.  Here is what I'm
> talking about:
>
> Original system:
>
> i           u
>
>       a
This is indeed the vowel system of Pre-Proto-Q (Q standing for "Quendic",
the working title of my "Elvish" language family, to be replaced some time
by a less Tolkienesque term).  Proto-Q, however, has seven vowels,
namely (ö and y are front rounded vowels):

  i   y   u
   e  ö  o
      a

The trick is that in Pre-Proto-Q, the vowel features [front], [round] and [open]
became autosegmental, i.e. they existed on a tier of their own, binding
to morphemes rather than segments.  On the segmental tier, all there was
(besides consonants, of course) were featureless vowel positions.

There was an ablaut process in action; under this ablaut, a morpheme
could be strong grade (having the feature [+open] attached to it) or
weak grade (lacking the feature [+open]).  (There was also a zero
grade, wherein the vowel position is lost.)  In addition to that, Proto-Q
had i-umlaut (if a morpheme had [+front] and no other features, this
feature spread leftwards through the word) and u-umlaut (the same game
with the feature [+round]).  But as if all that wasn't enough, there
was total vowel harmony in case of morphemes that had a vowel
position but no features attached to them: in this case, the morpheme
borrowed all the features from the root of the word.  Such morphemes
are indeed quite frequent in Proto-Q verbal morphology.

All these processes led to the rise of seven vowels, which resulted
from the possible non-zero combinations of the three features:

           a   e   i   o   u   ö   y

[+front]   -   +   +   -   -   +   +
[+round]   -   -   -   +   +   +   +
[+open]    +   +   -   +   -   +   -

(I hope this is all neatly lines up; this mail front end doesn't let me use
a monospaced font, so I have to count the blanks.)

I haven't yet worked out the gory details of this monstrosity
(such as contributions of vowel features from laryngeals and
other consonants, rule priorities and other minutiae),
but this is the basic picture.

> ai > e
> au > o
> ia > ä
> iu > ü (high front rounded)
> ua > u or o
> ui > ï (high back unrounded)
>
> New system:
>
> i ü       ï u
> e           o
> ä           a
>
> VH comes from fronting/backing of unstressed vowels based on
> frontness/backness of stressed vowels.
>
> What do y'all think?
I have seen the same vowel system in a reconstruction of Proto-Uralic,
so there is nothing wrong with it, I think.  Nicely done.

Jörg.

______________________________________________________________________________
UNICEF bittet um Spenden fur die Kinder im Irak! Hier online an
UNICEF spenden: https://spenden.web.de/unicef/special/?mc=021101