--- In qalam@yahoogroups.com, Michael Everson <everson@...> wrote:

> At 13:16 +0000 2005-08-07, Richard Wordingham wrote:
>
> >Nasalised vowels are a bit complicated. If one interprets vowel
> >initial syllables as beginning with a glottal stop, then one can

say that:

>
> The glottal stop is not a part of the phonetic
> inventory of Vai. Why do you insert it into your
> analysis?

> >2. Nasalisation is contrastive after glottal, labiovelar and velar
> >consonants. (A form of rhinoglottophilia!)

> kp and gb aren't labiovelar; that term is used
> for kw and gw, is not not? Kp and gb have double
> labial and velar articulation, but I think that
> is a different thing.

I had assumed the IPA definition would hold sway outside from
Indo-European matters, though I have seen the IE usage used in
comparative linguistics elsewhere. The difference may not matter much
anyway - I have come across a claim that labialised velars (e.g. k_w)
have become labiovelars (e.g. k_p) under areal influence in at least
one language in the area.

> >On the basis of the above, I would say that vowel nasalisation is a
> >feature of the preceding consonant (cf. Irish mh).
>
> I thnk it mighty dangerous to derive phonetic
> analysis from the orthography. When I say lámh or
> creideamh or Samhain in Irish, I do not have a
> nasal vowel, and nasalization is not common in
> this environment.

But you're not old :)

This illustrates one of the dangers of using information from
libraries! A little hunting on the web turned up the statment that
the nasalisation for <mh> is on the way out in Irish (
http://www.daltai.com/discus/messages/20/13975.html?1121265883 ) and
restricted to the islands in Scotland (
<http://www.akerbeltz.org/beagangaidhlig/gramar/grammar_nasalisation01.htm>
).

>
> I don't really know what the rest of your analysis is meant to do.

Suzanne claimed that no more than 60 symbols were really used in
normal Vai writing and that the other 140 plus had been added by
linguists. (Personal context: Non-Cree linguists write Cree a lot
more complicatedly than the Cree of Suzanne's acquaintance.) Peter
Daniels asked how many rows she claimed they had added. I supplied an
answer, but first provided the context that many rows were almost empty.

A natural question that may arise is why there are so few rows for
nasal vowels. The answer seems to be phonetic. It is simpler to
explain if one assumes glottal stops are present. A failure to list
is not evidence of absence - of the Proto-Tai glottal stop, Fang Kuei
Li wrote in 1977, 'This is usually kept in most dialects, although it
is not written in the orthography by many authors.'. By 'orthography'
he means the transcription system of his sources. Another example is
the glottal stop in German - it is phonemic, but readily omitted from
lists - see the comment at
http://en.wikipedia.org/wiki/German_phonology . Perhaps I should rise
to Panini's level and add at the end - 'The glottal stop is always
elided.' (cf. Panini's 'a a') :)

Richard.