Re: a-umlaut (was Re: Epicene words)

Ray Brown wrote:
> On Tuesday, March 1, 2005, at 06:16 , Jörg Rhiemeier wrote:

>> No ;-)  I resisted putting in e-umlaut, o-umlaut, etc.; that would
>> have reduced my nice vowel system to a stew of /2/s ;-)
>
>
> So I should think  ;)
>
> AFAIK e-umlaut & o-umlaut are not found in natlangs, only i-umlaut (the
> most common), a-umlaut and u-umlaut (apparently the least common). It
> seems that suffixes with vowels at the apexes of the vocalic triangle have
> a tendency to affect the vowel of the stem.
>
Kejeb started out with only the three vowels *a *i *u, but then a full
set of umlauts expands that to nine vowels

i   i\ u\   u
e           o
&   a       Q

which are then again reduced to six or five in the dialects
through merger.  Classical Sohlob has

i   i\   u

&   a    Q

while Kidilib has the same minus /i\/ (_e_ in the romanization
being /i\/ in Classical Sohlob and /&/ in Kidilib).  Linjeb
has

i y     u
e       o
     a

The mergers are not always the obvious ones.  Thus /e/ of the
first umlauted stage goes into /i\/ in Classical Sohlob but
into /&/ in Kidilib, while Linjeb /y/ contains both /u\/ and
/u/, Linjeb /u/ being from secondary diphthongs arisen through
the loss of intervocalic *N and *s > h.

Other dialects that I haven't worked a lot on yet acquire also
front rounded vowels from /u o Q/ when there is a palatal
vowel in the same word.

--

/BP 8^)>
--
Benct Philip Jonsson -- melroch at melroch dot se

         Solitudinem faciunt pacem appellant!
                                             (Tacitus)