Keith wrote:
> > Well, I had big problems with _hiisie_. [h] -> 0, leaving _iisie_.
> > Then the vowels shortens -> _isie_. Then the vowels are supposed
> > to drop, but I didn't really know what to do with [ie].
> Well if you want to follow British > Welsh what happens is
> /-ie/ >> /-ije/ >> /-ID/ <-ydd>
Oh! {-ydd}! I like that! That would be spelt _idd_ in Cein.
So _hiisie_ would become _isidd_ [I'sID]. Nice.
What would /-ea/ become? That's a common adjective
ending in Quenya.
> > Well, I used _limbe_ 'many' and let that get a more general
> > augmentative meaning. I'm not sure if that should be _llif_
> > instead. The rules are "nasal + voiced stop -> nasal",
> > "medial [m] -> [v]" and "drop final vowel". I'm not sure in
> > which order to apply them.
> IMHO /limbe/ >> /llimm/ double /m/ resists lenition. Modern Welsh always
> writes a single <m> but the sound was double (all single /m/ having
> gone to /v/ <f>. Possibly later /mm/ >> /m/ when final, I think this has
> happened in Welsh, but not in Cornish where /mm/ sometimes went to /bm/
> or even /bb/. N.B. Mod. W. <p t c> are really /pp tt kk/.
Aha. I skipped the double letters because it looked odd
when the stress was on the final syllable: <annon>.
Perhaps I should write it as <mm> after all.
> > No, {ch} is the mutated form of {c}, which I decided should
> > occur after the definite article. I'm still not sure if there
> > should be mutation or not after _i_ (sg) and _ir_ (pl).
> /i/ might be expected to cause lenition (soft mutation).
Yes, I'll go with that.
> /ir/ might
> just possibly cause germination leading to a spirant mutation
Very good idea too.
> and maybe doubling of /l r m n/.
What do you mean by "doubling of /l r m n/? That it should
be geminated in the beginning of a word following _ir_?
_melf_ 'love' >> _i mmelf_ 'the loves' (to take an odd ex.)
> The original /r/ of ir would be lost except perhaps before vowels.
Good idea. Initial vowels are very common in Cein.
||| daniel
--
<> Mad llamgalf! <> daniel.andreasson@telia.com <>
<> Ond llam! <> www.geocities.com/conlangus <>