Help creating ablaut classes and other apophony

[Ahzoh]

I continue to have difficulties in making Vrkhazhian feel "triconsonantal" enough (and I use "triconsonantal" in the same way one might use "orientalist" to describe an aesthetic; a generalistic vibe or flavor rather than an actual and clearly-defined thing). Mostly, I just don't have the kind of ablaut grades you find in PIE or the Semitic languages. I have at most merely changes in the shape of stems due to stress placement as well as grammatically-induced consonant doubling. But it doesn't feel enough.

Most I don't know how to develop ablaut classes like these:

More: show

ablautclasses.png (30.33 KiB) Viewed 760 times

More: show

ablautclasses2.png (24.89 KiB) Viewed 760 times

(from https://escholarship.org/uc/item/7t18f7jw)

I could just handwave and say "bam! you have this ablaut class and you have that ablaut class" and be done with it, but that is dissatisfying and I also don't want to exactly copy the Semitic languages.

My constraints also make things difficult: such as prosody considerations (right-leaning stress with a preference for trochaic rhythm, stress basically always on the theme vowel of the verb) and a dispreference for non-final superheavy (CVVC) syllables.

But really, I need help in figuring out where to source the ablauting. For what reason might one verb have a u-colouring morpheme while another has a-colouring or -i-colouring morphemes? Moods that got semantically-bleached? A collapsed distinction between unergative and unaccusative verbs?
That's my major problem, figuring out where specifically the mutations come from and the specific grammatical functions of those mutation-creating morphemes. This is my problem for the verbs and my problem for deriving nouns/adjectives from verbs.

Some things I must consider:

My conlang's stems have "theme vowels" where the last vowel of the stem (e.g. CaCvC-) is any of the following vowels:
u = transitive verb (e.g. "hit", "see", "speak to", "protect")
a = dynamic intransitive verb (e.g. "run", "fall down", "get up")
i = stative intransitive verb (e.g. "be old","stand", "die", "fall apart")

The distant ancestor language of Vrkhazhian that develops these ablaut classes basically had principle parts:
Biliteral stems/principle parts
CvC- (realis)
CvC- (irrealis)
CvC-áC/áh- (jussive/imperative) [the jussive forms of verbs geminate the person suffixes in many descendants]

Triliteral stems/principle parts
CaCvC- (realis)
CaCvC- (irrealis)
CvCC-áC/áh- (jussive/imperative)

Causative/Quadraliteral stems/principle parts
CaCCvC- (realis)
CaCCvC- (irrealis)
CvCCvC-áC/áh- (jussive/imperative)
(I dunno about having a causative/quadraliteral stem, cuz I am thinking maybe I will use middle-radical gemination to turn predicative verbs into attributive relative clauses, and I'm not sure how that will work with 4-con roots.)

Lastly, the person agreement markers are also entirely suffixal (unlike Semitic which started off as entirely prefixal; the suffix conjugations are more of an innovation in descendants) which is going to affect prosody.

Reconstructed ancestor language's person suffixes:

More: show

Verbal subject suffixes:
1sg: *-in/*-ni
2sg: *-Vm/*-mV
3sg: *-Vč/*-čV
1pl: *-it/*-ti
2pl: *-Vm-Vn/*-mV-nV
3pl: *-Vč-Vn/*-čV-nV

Adnominal/Pronominal suffixes:
1sg: *-niɣ/*-ni
2sg: *-mVɣ/*-mi
3sg: *-kVɣ/*-ki
1pl: *-tiɣ/*-ti
2pl: *-mVɣ-nV/*-mi-nV
3pl: *-kVɣ-nV/*-ki-nV

Reflexes of those in Vrkhazhian:

More: show

Verbal subject suffixes:
1sg: -(n)ni
2sg: -(m)ma
3sg: -(t)ta / -sa (final-weak roots)
1pl: -(t)ti
2pl: -(m)man
3pl: -(t)tan / -san (final-weak roots)

Adnominal/Pronominal suffixes:
1sg: -ni
2sg: -mi
3sg: -ki
1pl: -ti
2pl: -min
3pl: -kin

[Almoura]

What does your consonant inventory look like? Out of what you've posted, I imagine č could be i-coloring, and ɣ could be u-coloring. Perhaps the vowels themselves cause a following vowel to assimilate when stressed, e.g. *mati > mate or mata, or something like that. Apocope (loss of final vowels) and syncope of the antipenultimate syllable make a lot of sense with your trochaic rhythm, and could be used to merge various forms together.

Other ideas for vowel coloring:

a-coloring: PIE got this from some sort of laryngeal. Uvular or pharyngeal consonants are ideal if you have them floating around.

i-coloring: Palatals, such as the aforementioned č. IIRC Tibetan has u > y after a dental stop, which could easily unround.

u-coloring: Labialized consonants would be perfect, if you have them. Velars make a lot of sense, like ɣ. Not sure if it's attested but I'd believe labials causing a shift to a rounded vowel too.

[Ahzoh]

What does your consonant inventory look like? Out of what you've posted, I imagine č could be i-coloring, and ɣ could be u-coloring. Perhaps the vowels themselves cause a following vowel to assimilate when stressed, e.g. *mati > mate or mata, or something like that. Apocope (loss of final vowels) and syncope of the antipenultimate syllable make a lot of sense with your trochaic rhythm, and could be used to merge various forms together.

Ancestor language had this inventory:
/m n ŋ/
/pʰ pʼ b tʰ tʼ d kʰ kʼ g ʡ ʔ/
/t͡sʰ t͡sʼ d͡z t͡ɬʰ t͡ɬʼ d͡ɮ t͡ʃʰ t͡ʃʼ d͡ʒ/
/f s ɬ ʃ x/
/w r l j ħ h/
/a e i o u/
/aː eː iː oː uː/

I forgot there are no voiced fricatives.

Other ideas for vowel coloring:

a-coloring: PIE got this from some sort of laryngeal. Uvular or pharyngeal consonants are ideal if you have them floating around.

i-coloring: Palatals, such as the aforementioned č. IIRC Tibetan has u > y after a dental stop, which could easily unround.

u-coloring: Labialized consonants would be perfect, if you have them. Velars make a lot of sense, like ɣ. Not sure if it's attested but I'd believe labials causing a shift to a rounded vowel too.

The person suffixes will not be a source of vowel mutation.

Also, as I said, I'm less concerned about the phonological nature of possible mutating morphemes, but much more concerned on the grammatical function of these morphemes that will trigger these mutations.

Like, why does one verb have the a-u alternation while another has the a-i alternation? Why does one verb have u-a alternation while another has i-a alternation? And still more why does one verb have u-i while another is i-u?
How can it be that three different markers come to mark the same grammatical function?

[Zju]

How can it be that three different markers come to mark the same grammatical function?

Maybe they had different etymologies/origins in the first place (see: plurals in Celtic languages and German), and then through phonetic erosion they were reduced to different ablaut patterns. Or just various allomorphs of the same origin combined differently with various sound changes.

[Ahzoh]

I came to learn/realize the kind of vowel alternation in West Semitic languages is becuase they derived their past/perfective from adjectives/the stative verb conjugation. So I imagine if they used only the prefix conjugation for all verb forms like East Semitic then there'd probably be less alternation.

I'm kinda just giving up on having the kind of patterns seen in West Semitic languages. My attempts through diachronics leaves me unsatisfied with the results.

[timharvey]

With these things, are there principles to remember?

[Zju]

What do you mean?