On Sat, Jun 22, 2013 at 7:37 PM, Alex Fink <[log in to unmask]> wrote:

> No!  Certainly not a *lot* weirder.  The odds are slightly pretty much
> exactly fifty-fifty that a random language will have it, in WALS.  However,
> in three out of eight languages that have it, it can't appear in absolute
> initial position (and perhaps other initial positions).

fair enough. my apologies for not gathering the relevant statistics before
throwing out my usual blanket statements. nevertheless my main point was
that phonologies with /m/ and /n/ and no /N/ (with or without [N]) are all
over the place.

> That's the prevailing one, yeah, but there are other sources of [N] (and
> thus of /N/).  At times it seems to be the mòst preferred nasal in coda:
> witness e.g. Caribbean dialects of Spanish, and many Chinese varieties,
> shifting all coda nasals to [N] -- in some Spanishes even internally before
> heterorganic stops!  In Nyole, a Bantu language, /N/ comes from *p,
> seemingly by rhinoglottophilia from an intermediate [h].  In Samoyedic, [N]
> was epenthesised before initial vowels.  And of course even if there is
> [Ng] it doesn't have to come from /ng/; prenasal series can come from
> elsewhere (I seem to remember SE Asian examples of spontaneous
> prenasalisation of voiced stops).

another one i thought of is the Hindi word for 'lion' which (i don't know
Hindi) sounds to me like [sIN] but the wikipedia-supplied phonology
suggests it could also be [sINg] = /sIng/, since /N/ is not listed among
the Hindi consonant phonemes. either way, Sanskrit *siṃha* with the usual
place-assimilation of the anusvara 'm' thing, since Sanskrit /h/ somehow
counts as a velar according to the traditional grammarian(s). if anyone can
explain that to me, that'd be great.