Re: CHAT: Importance of stress

David Bell wrote:

>Well, close enough.
>
>(3)     er.in.var
>        or.carm.ar
>        er.in.is
>
[snip]
>> If so, then Amman Iar is weirder than I thought!  I was assuming that
>> intervocalic single consonants syllabified with the following vowel, and
>> that intervocalic consonant clusters and geminates were broken up
>> (i.e. VCV is V.CV, and VCCV is VC.CV).  In other words, I was assuming
>> the following syllabification:
>>
>> (2)     e.rin.var
>>         or.car.mar
>>         e.ri.nis
>
>Nope.  (3) can be shown to be the case by looking at the derivational
>morphology of the words as well as the stress patterns.  erinvar and erinis
>are derived from the same root erin (erin-var and erin-is). orcarmar is
>derived from the root caram > carm (or-carm-ar).
I don't mean to sound confrontational, but why should derivational
morphology make any difference?  In natlangs, syllabification rules
routinely ignore morphological boundaries.

When I asked what your Amman Iar intuitions were, I was referring
to your intuitions about syllabification.  When you pronounce "erinis",
do you really treat the "n" as syllable-final, and not syllable-initial?
If you were to pronounce the word super-slowly, syllable by syllable,
would you pause before or after the "n"?

I'm sorry to be bull-headed about this.  It's just that the analysis in
(2) seems to make perfect sense of Amman Iar metrics, while the
analysis in (3), in addition to going against known universal tendencies
of syllabification, makes Amman Iar metrics rather messy.  If your
intuition is that "erinis" is really "er.in.is" instead of "e.ri.nis",
that's fine.  It's your conlang, after all.  But I just want to make
sure that you're sure of that before I abandon my analysis.

Cheers,
Matt.