Re: Historical realism and prenasalized stops

In a message dated 10Æ05Æ02 3x35x43 PM, vionau@MAC.COM writes:


> > I would've expected *t > t_h, and *d > t, not the other way around, but I
> > can't speak for universals.
> 
> Well, now...lenition tends to move the other way, no?
> 
Yes, I was just suggesting because that's what happened in English.   :)   
(Our voiced stops are now pretty much always voiceless unaspirated, according 
to John Ohala.)

<<I'm not sure where my loss of aspiration comes from (natlang examples are
momentarily escaping me. I'll keep searching), but my theory is that, once
intervocalic voiced stops fricated, there were only two stop types left, one
voiced prenasalized and the other voiceless aspirated--which leaves a *lot*
of room between on the VOT timeline. It seems reasonable for [t_h] to move
toward the middle in these circumstances, you think?>>

     What you describe is, basically, a VOT pull chain, and it makes sense, 
even if it's never been attested (and it probably has).   The only other time 
I've heard of aspiration being lost was how in Greek, for example, where 
there was dissimilation of aspiration (I think it was that if there were two 
aspirated stops in a word, the first lost its aspiration...).   And then, of 
course, aspirated stops turning to the corresponding voiceless fricatives is 
highly common (so, /t_h/ > [T], /p_h/ > [f], and /k_h/ > [x]).   Also, you 
only described the coronals.   Does similar stuff happen with the labials and 
velars?

<<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. :)>>

Let's see...   So the stop series is: [p], [p'], [mb] (labial), [t], [t'], 
[nd] (dental), and [k], [k'], [Ng] (velar).   Word initially, the 
prenasalized stops are prenasalized.   So, a word *spelled* /mba:ta/ is 
realized [mba:ta].   (Note: I don't know the SAMPA for prenasalization, so 
I'm just using the nasal + stop orthography).   Initially, though, the 
nasalization spreads to the previous vowel, and the prenasalization is lost.  
 So, for /mba:mba/, you get /mba:~ba/.   In this situation, the new [b] takes 
some creaky voice to distinguish it from [p].   So, a minimal pair would be 
the word /mba:mba/ [mba:~ba] compared to /mba:npa/ [mba:~pa].   (Post-vocalic 
/n/ only nasalizes the previous vowel.)   And the creaky voice sustains the 
voicing.   And then prenasalized stops can't occur in the coda position.

The problem will come when other things get in the way.   So, the coda 
consonants that are allowed are /j/, /w/, /n/ and /l/.   The first two 
trigger diphthongs, and the second nasalization.   The fourth, though...   
That'll be tough.   I haven't decided how to resolve tough medial 
combinations like /VlmbV/ or /VlndV/, or anything.   That's a toughie.   Any 
ideas?

-David

"imDeziZejDekp2wilDez ZejDekkinel..."
"You can celebrate anything you want..."
               -John Lennon

Reply