Kadi/Salty

General
Salty Kadarh is the most widely spoken of the Downstream Kadarh dialects. It itself is made up of several subdialects that vary insignificantly amongst each other. The central, core variety spoken in the middle of the dialect zone is taken as the intermediate form and is going to make up all the examples in this article unless explicitly written as not so.

Salty Kadarh is spoken in a living community. It is one of the few Kadarh dialects that actually have a written orthographic standard (based off the Ktarh open orthography), and it has a rich tradition of both oral and written literature. It has a language standard put in place by a Ktarh board, with basis in the central variety of the dialect. The use of the standard, written in a Ktarh alphabet, is actively encouraged by the Ktarh authorities.

It is a highly inflected language, with considerable polysynthetic elements. It has an argument hierarchy based on animacy where the most animate argument always fills the agent argument of the verb and the least animate argument fills the patient argument; other arguments may be filled with nouns of any animacy.

Phonology
The phonology of Salty Kadarh is typologically fairly unremarkable for its area: it has a semi-vertical vowel system and a marked scarcity of labials. It exhibits no tone, though it does have vowel length.

It has eighteen consonants, all of which can be geminated. It has an extremely unusual phoneme, the sole true retroflex /ɖˡ/, that quite thoroughly stands out from the rest of the system. Its most abundant source was the Proto-Downstream-Kadarh /*l/ and it corresponds to laterals in other dialects most of the time.

The consonants /ɰ/ and /j/ may respectively be realised as [ɣ~ɻˠ] and [ɻʲ]; these allophones are unconditioned and in free variation with the regular realisations. Otherwise, the consonants in isolation generally tend to be close to the values prescribed by their IPA transcriptions.

Allophony and Morphophonology
Salty Kadi has some thorough and widespread allophony that involves clusters with /j/ that change into a unitary phone.

Several allophones are in free variation, such as [ʒ~ʐ] for /rj/. These do not have a conditioning environment.

Salty Kadi also has a regressive voicing assimilation process: any cluster of obstruents will agree in voicing with its last member. Assimilation does not apply if an obstruent is followed by a non-obstruent: voicing is untouched there. Ejective obstruents lose their ejectivity and become tenuis voiced in environments that promote voicing.

There is also a closely correlated ejectivisation spread, by which all clusters of obstruents that have an ejectivity distinction will follow a simple distribution rule: if an ejective is present in the cluster, the entire cluster is treated as an ejective and all members of the cluster lose their ejectivity, which is transferred to the final member of the cluster. For demonstrative purposes, a cluster of /at'ka/ will be realised as [atk'a].

The glottal stop is a glottalised, ejective voiceless stop for the purposes of assimilation. Whenever it follows a tenuis consonant, it assimilates to it wherever possible and is realised as a glottalisation feature on the consonant to which it assimilated.

The phoneme /ɖˡ/ has the voiceless and ejective counterparts in [ʈˡ] and [ʈ'].