Re: your opinion

On Mon, 31 Dec 2007 20:50:25 -0500, Reilly Schlaier <schlaier@...>
wrote:

>>Which is fine by itself, but if you have for /I/ before a retroflex: [I] >
>>[I\] > [M_x] > [U] (no distinction between /I Y U/ in that position, right?)
>>the epenthetic [I\] should probably get dragged along too. Probably only the
>>first step of that was conditioned; I can recall reading of a parallel of
>>sorts from Norwegian, where [E] > [&] generally before retroflexes, but
>>after [l:] >> [r`] and [s\] >> [s`], these new retroflexes do not cause the
>>same change anymore. So it seems it was the [r] in the former clusters such
>>as [rt rn] that originally triggered [E] > [&], not the retroflex [t` n`]
>>etc. that they became.
>
>im not sure why that would drag [i\] along if it never occurs before a
retroflex
>right?
>now im not sure
>confuzzled
I gathered you're trying to do this, in essence:
-1- [i I e] become [i\ I\ @] before a retroflex
-2- [i\ I\ @] become [M U 7] everywhere
-1'- [@] becomes [I\] everywhere
-2'- [I\] becomes [i\] everywhere

But change -2-, which I'd argue to be unconditioned "drift" (ie. once the
front vowels have developed the backed allophones, the retroflexes aren't
"needed" any more for the further backing) should mess up either the input
or output of changes -1'- and -2'-. OK, in theory it's possible that even
-0'- [n= l=] become [@n @l] everywhere
occurs after change -2- so the "vowel trajectories" would not strictly cross
— but in your Big Inventory here, I just don't think an epenthetic vowel
would take on a quality different from all the phonemic vowels. Altho I can
see why you wouldn't *like* that; you'd get new intervocalic [k], which
would mess up your neat "cross-allophonic" stop-system, in part.
akn [akUn] = /akUn/ not /akn/
akun [agUn] = /agUn/ not  /akUn/
agun [aGUn] = /aGUn/ not /agUn/

John Vertical