<!>Newtlang incubator (NP: NWC) (2016-09-10 18:38:29)
Newtlang incubator (NP: glot-taolic theory)
Anthologica Universe Atlas / Forums / Department of Creativity / Newtlang incubator (NP: glot-taolic theory) / <!>Newtlang incubator (NP: NWC) (2016-09-10 18:38:29)

? twabs fair maiden
posts: 228
, Conversational Speaker, [ˈaɪwə]
message
I didn't really deal with l/r or gm/gn. So:

gN [ɰN] > ɢN > ɢ (or q)
/k k: k:' g g: g:'/ > /ĉ č k ĵ ǰ kʼ/ (idk the exact identities, but ĉ was probably a sibilant and č an affricate)
/q q: q:' ɢ ɢ: ɢ:'/ > /q q qˤ g kʼ kʼ/

the middle stage of r r' l l' is r ṛ l ɬ, kind of

Cl' > tɬ etc. Cl > tš, Cr' > ṭṣ, Cr > ts
r'C > ṭ. rC > ṭ near high vowels, otherwise > t. l(')C > šC
r(')l > l/ḷ as above, rl' > ɬ, r'l' > ɬ̣
lr(') > r, otherwise as reverse

Also, fixing up the vowel system a bit:

A better system of pairs is (i : ɨ) :: (y : u) :: (e : o) :: (ɛ : ɔ), with /a/ neutral. Thus I can just say that consonants were devoiced before the latter pair (the inverse happened in Oghuz). /a/ was back [ɑ], so also devoicing there. i/ɨ > ʲä, y/u > ʷä, e/o > ä, ɛ/ɔ > e, a > a (but probably e > a later except in some dialect or sth.) Thus:

túinde > *tʷä́idä > tʷid
bearanna > *bésaNa > ban (with irregular syncope of *a)
mairellëonda > *mairäɬä́äda > maiɬad (maiɬd?)
tumbendandur > *tʷäbä́dätʷär > tʷäbatsʷ (tt > t)
banneanhemmuean > *paNeŊä́Mʷäen > paŋamʷ

regmannulluillear > *räqaNʷä́LʷäiLer > qanʷaɬʷiɬ
giellammeramean > *gʲäLaMä́ramen > dɮʲamam

I need to fix the frequencies in my generator, but this is looking better I think