Assorted (and rather theoretical): VSO languages, articles, ergativity, and morae

Matt wrote:
>
> In response to Doug's query:  I've been racking my brains, thinking of all
> of the verb-initial language families I've been exposed to (Polynesian,
> Western Austronesian, Salishan, Mayan, Celtic, Semitic, Nilotic) and in
> all of these except one (see below), noun phrases may be preceded by
> something which could loosely be called a "determiner".  In some languages
> these determiners mark definiteness (Hebrew, Malagasy), in others they
> mark case (Tagalog), in others they mark noun class (Tz'utujil), and in
> others they mark all three (Thompson Salish).  So, if you interpret
> "determiner" broadly, then there may be some sort of correlation.
>
> The one counterexample I was able to come up with was the Oto-
> Manguean languages (includes the various languages commonly grouped
> together as "Zapotec" and "Mixtec"):  These are strictly VSO, but have
> no determiners, articles, or prenominal case markers of any kind.  So
> if there is a correlation, it's not a perfect one.
>
> Tokana, incidentally, goes with the flow:  It has prenominal determiners,
> which mark case, person/number, animacy, and definiteness.  They also
> function as pronouns when there's no noun following them.
>
Well, with the above (and other people's contributions--thanks to all), I
think I'll go the Tagalog route--mark case with prenominal determiners, but
skip definiteness, noun class, or plurality.  I know this idea is about the
third in three days.  Consider it a prime example of the sort of rapid
changes I tend to institute (apart from the rest of you).  The reason for
this appears to be that I have such an emotional attachment to Skerre (it's
_my language_) that I try to make it as perfect as possible as fast as
possible (I have sketches of other langs, but they probably could never
reach the _my language_ level).
-------------
Ray wrote:
> Another solution would, of course, be not to use 'ta' as the definite
> article and, indeed, not to use anything with CV pattern as an article.
> Would a VC article avoid the 'reduplicative plural' problem?
I did think that this was a possible solution.  In all the other cases
besides the absolutive the article was VC, taken from the non-definite case
marker + t.  However, the non-definite absolutive marker was zero, and [t]
can't stand on it's own, thus the problem arose:  how to handle the
absolutive definite article.  However, as you can see above, the problem has
been fixed satisfactorily.
------------
Onto the topic of ergativity:
BP Johnson wrote:
> I read somewhere that Mayan is VOS.  I suppose this really means it's
> ergative, doesn't it?
I think Mayan is ergative, but not necessarily because of the VOS.
>
> What other VOS languages are there?  Are they ergative or non-ergative?
Malagasy is the usual VOS example, and I think it is accusative (Am I right,
Matt?). So certainly not all VOS languages are ergative.

In another post BP wrote:
> The ergative constructions in New Indo-Aryan go back to constructions using
> a passive past participle instead of a finite verb and the agent in the
> instrumental case.
>
> Are there any other languages where the instrumental case is used as the
> agental case in ergative constructions?
In Tongan, the ergative marker phonetically resembles in the instrumental
marker in other non-ergative Polynesian languages.  But there is no
certainty as to whether Tongan got its ergativity like the New Indo-Aryan
languages did.  And there is no consensus among the syntacticians as to how
to analyze ergatives.  One theory out there is that all the ergative
sentences are like passive sentences (hence they get an instrumental or
ergative marking), but IMO I don't think that is right for ALL ergative
languages.
-------------
On morae:
According to what I learned in my phonology/phonetics class this semester,
the linguistic definition of a mora as W.S. Allen (from Ray's post) defines
it:  a syllabic nucleus.  Within a syllable, onsets (the first consonant) is
disregarded.  Thus a single short vowel is a light syllable and counts as
one mora (CV, V= 1 mora).  I believe a syllabic sonorant is also one mora. A
vowel followed by a consonant, or a long vowel are heavy and counts as two
morae (CVC, CVV=2 morae).  This system of morae counting is used lots of
languages (including Latin, Greek, and Japanese) to determine stress or
pitch accent, and indeed the linguistic term is borrowed from Classic Greek.
One weirdness is that some languages count a short vowel plus consonant as
one mora, and only a long vowel plus a consonant or a short vowel plus a
geminate (long) consonant counts as two morae (so in these langs CVC=1 mora;
CVVC, CVCC=2 morae). The other weirdness is that some languages have
syllables that are extrametrical (not counted in assignment of stress), and
thus you get alternation between the penult and antepenult as in Latin.  I
must admit that my knowledge on this matter is a bit shaky, so sorry if that
contributes to any vagueness in the above explanation.
--Doug