a phonetic mystery

[Pamela Munro]

All of David's impressionistic judgments sound correct to me (thus,
John, it's not the case that there's no nasalization after nasals!).
Maybe there is a contrast, as he seems to suggest, but I believe that if
someone was interested in pursuing this exhaustively stress, what
consonant follows, and syllable position within the word the
post-nasal-C vowel would combine to determine which of these variants we
hear. I do not believe there's a phonemic contrast. But if you guys
prove me wrong, great!

Pam

from David:
This is precisely why I think we've been sloppy about this phenomenon.  I
agree that the vowel in the ma- prefix is strongly nasalized, as is the
prefix in mani 'to walk'. The position of that vowel is also considerably
higher -- more schwah-like -- than the oral equivalent.   But the -na in
wana 'now' is much less obviously nasalized most of the  time, and the
position is low central, more like the stereotyical "a as in  father" than
like a schwah.  Most speakers deny any difference when I ask  them about
it, but their practice belies their intuitions.  Another set  would be the
nasal u in nupa 'two', which is so strongly nasalized that  many people
are tempted to write numpa until we train that out of them, in  contrast
with nuwaN 'to swim' or manu 'to steal', where the "u" is higher, tenser,
backer,  and far  less strongly nasalized, and no one every tries to write
a nasal consonant after the "u".  I don't think anyone has ever
suggested that manupi 'they steal' should be written manumpi.
But I do not  completely trust my own  ears at this point, and I am not
sure how much consistency there is from  speaker to speaker or place to
place, though I am convinced that any given  speaker is quite consistent
about which pronunciation goes with which word.  If there are
contradictory data out there, I'd like to hear about them.

David