Alex Fink wrote:
> On Sat, 21 Apr 2007 22:52:32 -0400, Roger Mills <[log in to unmask]> wrote:
> >
> Very nicely done!  (Much like your sound changes, which I don't think I 
> got
> around to commenting on.)  Seems to be reasonably comprehensive; but since
> Gwr's been through numerous relays I guess I tend to think of it as
> well-developed anyway.

Thanks! Actually some of those relays were done more or less on-the-fly, 
before I had a firm grasp of developments. A lot of the words have been 
revised, replaced, or found to be impossible :-( etc. etc. One advantage is 
that the sound changes have lots of wiggle room for irregularities, due to 
dialect mixing.
> >The various "see XXX" cross refs. are the closest we have to derivation;
> >these are forms related usually by initial vs. final stress proto-forms 
> >with
> >attendant sound changes.
> Did the stress variation have any systematic function in pre-Gwr?  (Maybe
> you've mentioned this before, and I've forgotten.)

Perhaps I should have written/posted more of the grammar, but it's still 
under construction. Initial stress tends to mark nouns and adjectives 
~stative verbs, while final stress tends to mark transitive verbs. But 
neither I nor the proto-language are very systematic in that regard......

 And what's the function
> of the m-, s-, i/y- that pop up in some of the etymologies?

Ah. Those are explained in the grammar. m- often forms adjectives; s- 
indicates "place where..." (< sÿ 'place') I/y- is an old passive marker that 
now forms patient-nouns. Again, not totally systematic.

Then there are some more recent formations-- n(ei)- < neq 'person' can form 
agent nouns, and by sandhi produces variant forms.  hi, the 'personal 
article' can also be used for agent nouns (esp. professions, loosely 
> Are there any pairs of initially- and finally-stressed forms that show a 
> lot
> of semantic drift?  I'd expect some, given that as I understand it the 
> pairs
> become doublets with no synchronically evident connection between them.

Not sure how much semantic drift I've introduced so far-- concerned mainly 
with basic vocab.-- but it's to think about for the future. Particularly for 
scientific/tech. vocab, which ought to be well developed since the Gwr have 
always been #1 in those areas (and no classical Greek/Latin analogues to 
borrow from; in fact Kash borrows or calques a lot of its tech. vocab. from 
Gwr languages). Since my own scientific background is poor, I'm going to 
have to look up the etymologies of things like 'parallel, sine, vector, 
atom, nucleus, proton, cathode-ray-tube :-) etc.' for ideas.
Thanks for the info on Chinese variants; I suspected as much.