Re: Unilang: the Phonology
From: | Roger Mills <romilly@...> |
Date: | Thursday, April 26, 2001, 1:39 |
David Peterson wrote:
<< What is a "hummed" nasal? >>
> How to explain... Well, take the sound [m], and just hold it, hum it.
If you'll notice, you can do this with all nasals: mmmmmmmm, nnnnnnnnnnnn,
NNNNNNNNN... So, in Swahili (where almost every singular form begins with a
hummed nasal), you have a word [Ngeni] (I believe), where you start with
your
tongue in the [N] position, hum it and go straight into the word. This
exact
one occurs in the word "hunger". When you get to the "ng" part, try holding
the sound; I think you'll see what happens.
Anyway, this is different from, say, Russian, where you get words
starting with [mn] and neither of them are hummed; you just kind of
pronounce
them.>
As I thought. A hummed nasal, which can be held till breath runs out, is a
syllabic nasal, right? Question: Is Swahili [Ngeni] two or three
syllables? I would guess 3, since AFAIK (and it's very little) it is not
always the case that the initial nasal (which at least in some cases is a
separate morpheme) is followed by a homorganic sound. (Is is possible to
have initial /n/ before /g/?) Swahili could also have a words like *mgeni*
or *Nzeni*, in which cases the nasal would perforce have to be pronounced
somewhat separately.
The Indonesian languages I'm familiar with differ in that 1) prenasalization
occurs only with stops (and usu. voiced ones at that, but that's accidental)
and 2) the nasal element is very brief and sometimes a little hard to hear;
and it's clearly just an onset feature. And in medial position, the
syllable boundary comes between the vowel and the prenas.stop; so if you ask
your consultant to "say the word very slowly", "samba" will come out
saaa...mbaaa, never saaam...baaa.
Interestingly, these intial prenas. stops do descend from an original
morpheme-- "Proto Eastern Indonesian" (just my hypothetical name) would have
had forms like *bérat 'heavy' (adj.), *ma-bérat 'be heavy'-- the pretonic
prefix vowel deleted (and the morphemic nature of *ma- got lost) resulting
in **berat ~ mberat, which then developed individually. Makes for some
fascinating doublets in those languages, especially if the meanings change,
as they often do.