Palatalization-split

A palatalization-split consonant inventory is one where every or almost every non-palatal consonant has a phonemical palatalized counterpart. This setup occurs mostly in North-Central Asia, including Russian, Nenets and Khalkha Mongolian. (More limited palatalization contrasts are common in the area, too.)

More complex variations are found dispersed over the world:

• Irish has no plain consonants, only palatalized and velarized variants.
• Marshallese takes this one step further with a palatalized/velarized/labialized tripartite division.
• Many Caucasian, especially Northwest Caucasian languages have idiosyncratic variations where certain but not all series can occur palatalized or labialized.

Curiously, purely labialization-split etc. systems do not seem to exist. Languages with many emphatic consonants can be essentially pharyngealization-split however.

Palatalization splits, especially the more complex variants, go frequently together with wholly or partly vertical vowel systems, with vowel frontness filled in from palatality of the preceding consonant, eg. /mʲi/ → [mʲi], but /mi/ → [mɨ]. It is these kind of languages where palatalized palatals are possible: this will be a segment that is both palatal in its articulation and imparts a palatal coloring: /ja/ → [ja], but /jʲa/ → [jɛ] vel. sim. Interaction with vowel harmony offers similar possiblities (perhaps implying a suprasegmental interpretation of palatality).

Some corresponding pairs of consonants may differ in more than palatalization. For example, alveolars will commonly become postalveolar, or wholly palatal: /sʲ/ → [ʃ], /tʲ/ → [c]. Irish has a similar example with velarity: /vˠ/ → [w]. Also near the apex of the palatal POA, some series may merge, leading to eg. non-palatal three-way distinction /p t k/ vs. palatal two-way distinction /pʲ c/.