sirdanilot wrote:^^ Love that phonology ! Rather disappointed there were no ejectives, but that's simply because I have this ejective fetish.
I love ejectives and almost included them, but I decided that having four plosive series would just be too much for my tastes, and I wanted to stay relatively close to the Greek influence too.
Why is glottal stop not allowed at the end of a word? I find it perfectly easy to pronounce. Again, believe it or not, you hit one of my personal fetishes: word-final glottal stop... (I just think they sound really cool, I have absolutely no idea why though)
It's nothing about being easy-to-pronounce; I just like having some phonemes with marginal distribution, and South Eresian is
quite the glottalstopfest.
Now the major points: firstly, I think the dental fricative, being the only dental in the language, is a bit too much in there. It doesn't belong. Better have an /f/ or a bilabial fricative. You could also add a /x/ versus /χ/ contrast in there, which seems perfectly symmetric.
First of all, I am quite disinclined to keep the inventory symmetrical, since perfect symmetry often seems artificial, and having a /x h/ distinction is fairly uncommon too. And lots of languages have phonemes that are isolated at a PoA; hell, English dental fricatives are pretty damn isolated, as we
only have fricatives at the interdental PoA. Just like this language.
Then, phonemic /ps ks kʷsʷ qs qʷsʷ/? Really? Affricates should be at more or less the same point of articulation (note the more or less, for things like /pf/ are still affricates). You could go this route: /͡pf ͡kx ͡kʷ͡xʷ ͡qχ ͡qʷ͡χʷ/ Personally, I can pronounce all of those with ease. If you want to keep your clusters with all the sibilants, you'll need some more complicated phonotactics, which is a fun thing though.
Who said they were affricates? I never claimed them to be affricates, because they are indeed not homorganic (except /ts/, but that patterns like the others, which are not). That said, they pattern
like affricates in the language. I'm treating them as phonemes because they pattern like phonemes within the language; if I were to treat them like clusters, then I would have to explain why a) they are the only clusters that may occur syllable-finally and b) why such a cluster may occur before /s/ and be maintained as [(T)ss] within speech, while other clusters simplify, i.e. /θθ/ just becomes [θ] and, indeed, /ss/ becomes [s] (which is a rule that I should have written above; I'll edit my post), whereas clusters of non-identical things do assimilate to each other in certain ways but do not simplify, i.e. /pp/ becomes [p] but /p_hp/ becomes [pp].
Word-final consonants are devoiced and undergo fortition: /b d g/ become [p t k], /r l/ become [ʂ ɬ] and /j w/ become [ç xʷ]. The sonorants also undergo these changes preceding or following an aspirated stop.
I'm not sure how reasonable these are, but I must admit it's quite awesome.
Word-final obstruent devoicing is ridiculously common cross-linguistically. The phenomenon of sonorant spirantization is found in at least Cakchiquel, so it's definitely not unheard of.