zompist bboard

My main lang wanted a friend to borrow words from. Gonna bash some ideas together here.

The vowel inventory is small and essentially vertical, with a three-way height distinction.

• The low vowel is IPA ja (spelled ia) after a palatalized consonant, and IPA a (spelled a) elsewhere.
• The mid vowel is IPA je (spelled ie) after a palatalized consonant, and IPA ɤ (spelled e) elsewhere.
• The high vowel is IPA i (spelled i) after a palatalized consonant, IPA ɨ (also spelled i) after an alveolar or retroflex consonant, and IPA u (spelled u) after a velarized, velar, or uvular consonant.

Because palatalization affects the spelling of vowels in a consistent way, always creating or adding the letter i, some consonant letters are reused for both palatalized and unpalatalized versions of the same consonant.

Here's the consonant inventory, with some POA distinctions collapsed to keep the table tidy: The alveolopalatal consonants act like "palatalized" counterparts to the velar ones, and the velar ones act like "velarized" counterparts to the alveolopalatal ones. (IOW: pʲ is to pˠ as tɕ is to k; both pʲ and tɕ trigger the front allomorphs of vowels, and both pˠ and k trigger the back ones.)

And here are the same consonants in practical orthography. Edit: Clusters are either geminates or nasal-plus-C. Nasal-plus-C clusters can occur word-initially. Alveolar consonants can occur in codas at the end of the word, but SOME PROCESS TBD gets rid of them if material is added that would create an illicit cluster.

The most common derivational process by far is compounding, and words longer than three syllables are essentially always compounds. There are two morphophonological processes that can happen in compounds -- triggers TBD. One is gemination: the initial consonant of the second word doubles. The other is lenition: the initial consonant undergoes these changes, which amount to "deglottalize if it's glottalized, fricativize if it's a plain stop":

I like what I see so far. I like vertical vowel inventories in general , and complex but not entirely symmetric consonant inventories.

Can palatalization be contrastive within a cluster? E.g. can there be /amʲka/? As in Gaelic. If so how do the vowel allophones behave when before a palatalized consonant, as compared with after?

Ah, yeah, could have been clearer about that. What I'm imagining is homorganic nasal-plus-C clusters, not "any nasal plus any C." So that includes same secondary articulation, and clusters always match for secondary articulation.

My current thinking is that vowels' quality will only depend on their syllable's onset consonant, and not on a following consonant. I guess that means I'll need to give them each a default quality for vowel-initial syllables, or have some kind of quality-spreading rule that takes over when there isn't a consonant there.

Syllable structure, revised
Still playing with the sound changes that derive this lang from the protolang of the family. Here's a revised version of the syllable structure.

An onset can be:

• empty, for vowel-initial syllables
• any single consonant
• any NC cluster, where C is a consonant and N is a nasal at the same point of articulation

Since there aren't phonemic nasals further back than /n/, clusters like [ŋk] or [ɳʈʂ] need a special analysis: either [ɳ] and [ŋ] are allophones of /n/, or they're allophones of some placeless nasal.

A nucleus can be:

• a single vowel
• after a non-glottalized consonant, a syllabic fricative, which is realized as [x] after a velar or velarized consonant, and [s] elsewhere.

A coda can be:

• empty
• after a vowel, a non-syllabic consonant, realized in the same ways listed above
• after a vowel and before a single consonant, a copy of that consonant to form a geminate
• at the end of a word, any alveolar consonant

muth [mʷuθ] "boy"
ntzrecsti [ɳʈʂ'ɤ.ks.tɨ] "undercooked"
psthenne [ps.θɤn.nɤ] "take (polite)"
quthiv [q'u.ðiβʲ] "dog"
ntsdicia [nts.ɗɨ.tɕa] "jerk, asshole"
pugi [pʷu.ʝi] "gone"
hede [hɤ.ɖɤ] "say"
bih [ɓʲih] "bird"
cstat [ks.tat] "start"

I have...mixed feelings about this, to be honest. It's a little too comically over-the-top consonant-heavy, like Polish gone wrong or a bad parody of something Pacific Northwest.