Re: Historical realism and prenasalized stops

Josh Brandt-Young wrote:


>> 1. *d > nd
>> 2. *t > d
>> 3. *d > D / V_V
>> 4. *t_h > t
>> 5. *nd > d / _V{stressed}>>
>>
I see no problem with this system.  (Of course, 3 and 4 need not be ordered
with respect to each other, but that's irrelevant).  It strikes me that 5
could be a rather late, purely phonetic/surface phenomenon.  Presumably
then, stress assignment must be ordered prior to 5.

>How else do prenasalized consonants react in the language you're working
on?
>I've never even considered this issue before, so I could use some
>inspiration. :)
>
In the various Indonesian languages I'm familiar with, (pre)nasalized stops
(or nasal clusters, it's possible to analyze them either way-- most opt for
the cluster definition) are generally quite straightforward as to origin--
they occur only medially in the large Western group (Malaysia, the western
half of Indonesia, Philippines), and can be reconstructed at the
Proto-Malayo-Polynesian stage.(Some langs. have a few stray forms with
initial NC, but usually outside the normal phonologic system-- vocatives,
interjections, particles.) The problem is that there is a lot of variation
across languages:  some words reflect say *mp in all exs. of a given
cognate-- but some forms may show *mp, relatives will have *p-- and there
are lots of intra-language doublets, usually with some sematic difference
but sometimes not (at least not according to dictionary definitions, eg.
Malay/Indonesian sepit~sempit both 'narrow'). No one has ever been able to
come to a firm conclusion whether "nasal accretion" was simply an optional
feature of the proto-language or some morphemic process whose significance
is no longer evident.

Not much of interest happens to these medial NC in the west; but several
languages turn them into geminate stops-- e.g.(using "k/g" as typical stops)
Toba Batak *Nk > kk, but Ng remains; Buginese has *Nk > kk, but also *Ng >
Nk.  Some PI languages IIRC have e.g. *Ng > gg.  There may even be languages
that simplify them, *NC > C, but I can't think of any offhand.....Stress is
generally not a factor as far as can be determined.

In most languages of the Moluccas and Lesser Sundas, we find that the voiced
and voiceless NC have merged (plain stops did not), and, crucially, are also
found in initial position-- but most of those instances are fairly clearly
due to vowel deletion in the very common verbal/adjectival prefix
(unstressed) *ma-; the nasal then assimilated to the base-initial stop.
Thus an earlier (hypothetical) Eastern system like:
*p t k,  b d g, mb, nd, Ng may end up as **f t ?, w r k, b d g.  The NC tend
to develop with some variety:  they're retained in a few languages, but
typically **mb > b, p, sometimes implosive or preglottalized b, **nd > t,
plain or implosive d, r or in a couple cases k (!), **Ng > k, g, or merger
with *N.  There is a lot of variation here too; one language will show a *NC
reflex, another not.
There are also forms (nouns, which would not likely have co-occurred with
*ma-) that seem to have developed an initial NC spontaneously-- there is
usually no good explanation.

Oceanic languages (essentially everything east of a still-uncertain area
somewhere on the north coast of New Guinea) went further:  not only the NC
merge, but so did the vd. and vl. stops, so Proto Oceanic has only *p, t, {a
vl.palatal}, k, "q" (or *?), mp, nt, {n-palatal}, Nk (also written mb,nd..Ng
by some) plus the usual nasals, semivowels, *s etc. There are also a great
many more initial NC reflexes that cannot be easily explained.  And the
usual variation, *NC here, *C there.  (It's very frustrating.......) Again,
stress probably played little role, since it seems pretty clear that given
basic CVCVC structure, penultimate stress was the norm.