Re: Update on me: Elliott (LONG)

Elliott Lash wrote:
>Let's look at the vowels one vowel at a time.
>-----
>THE VOWEL E
>
>E has many different allophones according to place. And yes you are right,
some of the transcription I used in my previous email were...incorrect. I
have remedied that.
>
>at the beginning of the word it is mostly [I]. However, in a few words...or
more specifically, certain forms, it is pronounced [i] in this same
position.
>
> Examples:
>    ergvol "knight"  [Irgvov]
>    eshgir "he went" [iSgir]
>
>It will be seen that [i] appears when /E/ is the prefix indicating past in
some verbs. Originally however this /E/ was also pronounce [I]. It was only
under the influence of Jsm "I was" /iz@m/ (<*[@js@m]) that the sound
changed.
>
>The other time that /E/ -> at the beginning of the word (and also in
general) when it is followed by /s, n, m/
>
>Examples:
>  Em /im/
>  Es /is/
Just based on these examples, I detect a rule:  /e/ > I / #_CCC...
                                                                           >
i / #_C(C)...
but evidently that doesn't work in other environments.  No matter (but I
think it could be adjusted to work)
(snips)

>THE VOWEL A
>When /a/ precedes /l/ several possibilities occur, depending on where /l/
is in the word.
>
>When /l/ appears before vowels, then /a/ -> [o] and the /l/ -> [v] (<*[w])
>
>Examples:
> shnalal "to know" [SnovO]
>
>When /l/ appears at the end of the word, the /a/ and /l/ coellesce to form
/O/.
>
>Example:
> shnalal "to know" [SnovO]
If you were to cite these in the reverse order, a more logical and natural
progression results:
        1.  -al(#,V) > O  (presumably due to velarization/ u-vocalization of
/l/.
        2. in the sequence ...OO..., break hiatus with "homorganic" [v]--
this in fact suggests to me that there may be some kind of underlying
remnant (say [w]?) of the /l/.  Do other sequences of like vowels also
insert some hiatus breaker?  One would expect it.

Is /a/ the only vowel affected by /l/?  What about other back vowels, if
there are any?  (Aha-- I see ergvol /ergvov above.  good)


>When /l/ appears before consonants then /a/ -> [a] and the /l/ dissapears.
>
>Example:
> krvalshu "reading" (present participle) [krvaSu]
This doesn't fit with the above development; but it could happen.

(Just riding my hobby-horse here.  These are very nice developments. I have
dealt with a natlang with something similar-- /l/ > 0 between any sequence
of a/u-- i.e. a-a, a-u, u-u, u-a; and the analyst had also devised a
seemingly strange writing system  ;-)