Re: Notation for Lithuanian Morphemes (was Underlying Circumflex in

From: Sergejus Tarasovas
Message: 16134
Date: 2002-10-10

> -----Original Message-----
> From: Richard Wordingham [mailto:richard.wordingham@...]

> >...yet <à> is a common notation for stressed
> > phonologically short /a/ (stressed phonologically long /a:/
acquires
> > circumflex accent automatically) --
> > certainly not what you meant; hence my question.

> I was looking for a symbol for /a/ at the morphemic stage where the
> prosodic contrasts are acute versus circumflex (possibly
versus 'not
> relevant') and (provisionally) accenteded versus unaccented. I
don't
> know the constraints on [a:]. <à> seemed the compactest way of
> indicating acute without implying a long vowel. Given the amount
of
> correspondence it has generated, I have failed abysmally.

To avoid misunderstanding, I probably had to put it like that:

...yet <à> is a common notation for stressed phonologically
short /a/ _except in tautosyllabic /aR/ (R={/m/,/n/,/l/,/r/}) --
/aR/ is a special case here_ (stressed phonologically long /a:/
acquires circumflex accent automatically _except in
tautosyllabic /aR/_).
/a/ in /aR/ is always treated as phonologically short (/a/,
not /a:/), no matter what prosodeme is applied (/stress with acute
pitch accent/, /stress with circumflex pitch accent/ or /zero
stress/).

On the morphological level, where _every_ syllable with a long
nucleus can be associated with a pitch accent (in distinctin from
the phonological level where only one -- if any -- syllable can bear
a pitch accent), plain {láng-} would do.

Sergei