Re: [tied] *h3 (More deja-vu)

From: tgpedersen
Message: 15817
Date: 2002-09-30

--- In cybalist@..., "Sergejus Tarasovas" <S.Tarasovas@...> wrote:
> -----Original Message-----
> From: tgpedersen [mailto:tgpedersen@...]
> Sent: 2002 m. rugsejo 30 d. 17:02
> To: cybalist@...
> Subject: Re: [tied] *h3 (More deja-vu)
> >> --- In cybalist@..., "Sergejus Tarasovas" <S.Tarasovas@...>
> >> What are the typological grounds for the statement that uvulars
> >> trigger a-colouring (and not o-colouring, for instance)? What
> >> be a possible (neuro)physiological explanation? Just wondering.
> > With respect to Greenlandic, if I remember correctly now so many
> years after, what my phonology teacher Jørn Rischel, who was
> in the design of a spelling reform for it, said, it is a three-
> language (i, a, u), but before uvulars (eg /q/) /i/ has the
> allophone /e/, and /u/ the allophone /o/ (present ending(?) -voq, -
> poq). For convenience they decided to keep /e/ and /o/ in the
> spelling in those cases.
> Isn't [i] -> [e] a plain lowering? That's what -- as I wrote in one
> my previous messages -- exactly agrees with my language intuition.
> is [e] -> [a] just a lowering? The latter would yield something like
> [æ], which is close to [front a], but far from [back a]. If post-
> PIE *a was indeed fronted, I'm happy with the lowering effect of
> uvulars. But some state [front a] was a phonetic realization of
> PIE *e, not *a.
> Sergei

Yes, but tough question.
Long vowels in Danish are ordered this way

i y u
e ø o
æ (ö) å

On top of that, /a/ has two allophones; we migh call them 'ä'
and 'a'. The 'ä', slightly lower than the one in English /hat/ just
fits in between /æ/ and 'a' (foreigners have a hard time with Danish
vowels). For the long vowels 'a' occurs before and after the
uvular /r/, 'ä' otherwise. The question is now which allophone is
unmarked, 'a' or 'ä'? Does /r/ lower /ä/ or does it keep /a/ from
becoming /ä/?