I have to make a thread on Zokardi. Its number system is my favourite I had to share Literally put the most effort into this one, not even a lot was put towered Isyan's numbers!
So I'm starting off with numbers right after phonology:
short: <i e æ u a u o õ> /i e ɛ~æ ʉ a u o ʌ/
long: <í é â û,ú ó ô,á> /iː eː æː ʉː oː ɑː/
(long */u/ has merged with long /ʉ/; long */ʌ/ merged with long /a/)
diphthongs: <æi ai æu õi õu ou oi ei eu> /æi̯ ɑe̯ æo̯ ʌi̯ ʌʉ̯ ɔʉ̯ oi̯ ei̯ ɛʉ̯/
(ei and eu are marginal, and where thet are retained from the merger with /æi/ and /æo/ respectively, thet are either pronounced /eː/ and /iʉ~jʉ/ respectively)
Vowels can also appear in hiatus, so you can assume any romanized <eu> is /e.u/ (which is a frequent sequence).
nasals: <m n ñ ng> /m n ɲ ŋ/
plosives: <p b t d k g> /p b t d k g/
affricates: <ts dz tj j> /t͡s d͡z t͡ʃ d͡ʒ/
fricatives: <f v s z ł š ž ś ź h> /f v s z ɬ~ɮ ʃ ʒ ç ʝ h>
rhotic: <r ř r*> /r r̥ ɹ/
approximant: <w l y> /w l j/
*can someone recommend a way to distinguish /r/ from /ɹ/ in romanization? Geminates are of course a thing, so <rr> for the trill doesn't work.
There is also the letter X which is /ks/ or /gz/ (depends on what it condenses). eg, feuxa /feˈuksa/ "to dream"; aixa /ˈɑe̯gzʌ/ (from "aigtsõ").
Consonant clusters are progressive: if a voiced consonant begins the sequence, every consonant following it will be voiced. Hence /ɑegzʌ/ instead of /ɑeksʌ/.
Usually speakers are very conservative in speech and pronounce every consonant clearly, but in certain words some great deal of lenition occurs. For example, the number 100 is "aigšxib." This should be pronounced /ɑegʒˈgzib/, but this is quite a mouthful and so you say /ɑegʒˈʒib/.
That's all for phonology for now. Let's move on to those numbers