Re: Labiovelar stops

Trebor wrote:

> I read on this list that it is believed that /kp/ and /gb/ might be
> related
> to /k_w/ and /g_w/. Are there any natlangs that distinguish all four (or
> only /kp/ and /k_w/ if it lacks a voice distinction)? Incidentally, are
> there any natlangs with /Nm/?
>
Some languages of Melanesia have /k_w/ or /g_w/ (in a dialect of Fijian,
IIRC, the g_w corresponds to /Ng/ in the standard.  In other languages the
origin is not always clear.)

Some Melanesian languages (perhaps even Proto-Oceanic, though not everyone
accepts this) also require the reconstruction of *N_w or perhaps *N_m,
usually deriving from earlier *m or *N in the environment of *u.  It usually
involves an irregular correspondence for *N or *m, but possibly some of the
modern languages actually have a labiovelar. Two words, especially, are
frequently cited as examples:  'house' < Austronesian *Rumaq, and 'man,
person' < *tau mataq. In _Lingua_ vol. 14-15 there's a paper by Haudricourt
on the subject.