Re: CHAT: Importance of stress

On Thu, 27 Jan 2000, The Gray Wizard wrote:

> Well, the jury seems to be in.
>
> 1) VC-avoidance is a strong tendency if not a universal in natlang
> syllabification.
>
> What then could account for the amman iar anomaly?
>
> 2) Syllabification is independent of morphological boundaries.
>
> What then could be the source of the high morpheme/syllable correlation in
> amman iar?
Let me take a stab at these two questions together, using the
analysis of Gosiute Shoshoni found in my dissertation as an
example of the same tendency (with a different outcome) from a
natlang.

In Gosiute, I found that morpheme boundaries must coincide with
syllable boundaries, which is exactly the same preference that
Amman Iar shows. In Gosiute, there is a set of morpheme-final
segments (called "final features" in the literature) which show
up when a stop or nasal follows. So the first person plural
inclusive pronoun is /tammyn/ underlyingly (/y/ is a high,
central, unrounded vowel); when a word such as /pia/ 'mother'
follows, the final feature /-n/ assimilates to the /p/:
[.tam.mym.bi.a.] 'our mother' (the /p/ also voices, but that's
another story).

However, final features are deleted when i) phrase-final, and
ii) before another morpheme beginning with a vowel. So /nywyn/
'Indian person' becomes [nywy] when phrase-final, and /tammyn/
becomes [.tam.my.] pre-vocalically in the phrase [.tam.my.a.ra.]
'our uncle'.

Phrase-final deletion is explained by the universal tendency for
syllables to end in a vowel (dubbed NoCoda), a requirement which
is more strictly enforced in phrase-final position in Gosiute.
Pre-vocalic deletion, however, can only be explained by invoking
a preference which Gosiute exhibits to have morpheme boundaries
coincide with syllable boundaries, otherwise one would expect
*[tammynara], with the /n/ final feature being syllabified with
the following morpheme: [tam.my.na.ra].

The preference for syllables to coincide with morpheme
boundaries seems also to be operative in Amman Iar, but in that
language, the resolution of the conflict is different. Rather
than delete consonants which find themselves on the wrong side
of the morpheme=syllable boundary, Amman Iar instead prefers to
violate NoCoda; that is, the congruence of syllable and morpheme
boundaries takes precedence over an unmarked syllable structure.

I think this is rather cool, myself.

Dirk

PS: When I get back to my office tomorrow, I'll see what I can
find in the way of reference material on syllabification. Do you
have access to a university library?

--
Dirk Elzinga
dirk.elzinga@m.cc.utah.edu