Re: Conlang puzzles (Was: Re: New monster word in Maggel ;))))

Julien Eychenne wrote:
(re Maggel _g_ as a length indicator)
>Well, I was just wondering if a you could not imagine that it turned to a
voiced velar fricative (no problem here)then either to a voiced velar
approximant or spirant (or even spirant > approximant) or to a uvular
fricative (to me, it seems phonetically possible ) and then voiced uvular
spirant. Whichever way you choose, I think it's easy to get a voiced uvular
spirant close to french turned R (the voiced velar
fricative/spirant/approximant can become uvular). Then it is really easy to
make it disappear and get a compensatory lengthening.  For example, the
sound can become underspecified from the point of view of its articulatory
properties, so it would be close to (voiced) h....>

Very astute.  The same thing happened in Gwr historical phonology with
intervocalic and final *g, producing long vowels-- the other voiced stops
*b, *d also lenited, ultimately > **w, y, then melded/diphthongized with the
preceding vowel.  (In some cases, I think, *g > G can >....@ too, resulting
in some @-offglides.)  This is one of the ways I derive the modern 8 vowels
from the original 3, *i a u. The 9th vowel [3^] |r| is easy-- any premodern
form of the shape **#(C)Vr# > Gwr /(C)r/

This gives me an evil idea-- perhaps some related language could retain the
|g| et al. in its written forms, and we could have at least a Tibetan if not
Maggelian writing system....