Zompist Bboard Again

dhok wrote: ↑Thu Aug 22, 2019 8:56 pm coptic is fake. only fake coptic can actually be real

Rau is a detailed account of the infinitely foul crimes against diachronics of stress timing and if you talk one more time about the fake syllabic plosives I'll add real syllabic plosives like in Kpelle

But before we get more on Rau, it would probably be a good idea to work out more of the grammar of Proto-Vengic.

So! Here is the K!ek Tbi K!oo text in Proto-Vengic. This is slightly anachronistic, but so is translating the Shquqou text into Kannow.

(ñ = *N, which was probably a retroflex nasal but which is preserved as such only in Qoa, the obscure and divergent cant of various bands of Katnahl pirates; it becomes /ɖ/ in V'eng (but /mj/ before /u/, presumably via mɳ > mɻ), /ɲ/ or /m/ in What (merging with *v, most *s, and word-initial *h), and /r/ in Bor Hlu)

Qocomati umara e Bidaadaa su ho atatumñu
qo-comati umara e bidaadaa su ho ata-tumñu
1-raise.FUT glory E Bdáát_G|ʰaa POSS COLL father-son
Glory to the family of the Emperor Bdáát G|ʰaa

This shows both Proto-Vengic strategies for possession: X e Y and Y su X. I gloss these as E and POSS respectively. The future in Proto-Vengic has the semantic function of a hortative or obligative, but in the few languages that preserve it, it serves as a future tense. Note also that egophoricity hasn't developed yet.

Bdáát G|ʰaa is borrowed here as Bidaadaa; Proto-Vengic has strict CV syllable structure, although V is permitted in word-initial position. (Vm marks nasalization.) However, it also has strong stress accent, so the vast majority of its descendants have more permissive syllable structure; the ones that don't, like Zzyxwqnp, generally developed large consonant clusters (the earliest written ancestor of pre-Zzyxwqnp allowed up to CCCVC) and then simplified them again.

Doo xode kusacubu e doogham e bikaa
doo xoda-e kusat-yubu e doogham e bikaa
year step-E two-ten E reign E glorious
In the twelfth year of his glorious reign

Xoda 'step (on a ladder)' + e is used to form ordinals; this is almost always contracted to xode. E is also used between nouns and adjectives; in some Vengic languages, it becomes possible to separate a noun from its adjective with e + pronoun, and this develops into an agreement system. (e.g. to front 'glorious', a e bikaa ... doogham)

The suppressed final consonant in kusaa "two" is realized in kusacubu < kusat-yubu "twelve". Lengthening of the vowel preceding a suppressed final is not entirely regular, and is probably a remnant of a pre-Proto-Vengic stress system; of course, given that length attracted stress in most branches of Vengic, there may be nothing pre- about it.

Qoxuña karuu ghetama-hamicii e yeqi
qo-xuña karuu ghetama_hamicii e yeqi
1-begin investigate.AUG ocean_NMLZ-fall E world
I begin my investigation of the world and its parts

"The whole and its parts" (of the world) is idiomatically expressed as "the ocean and the raindrops"; this is often contracted to ghetaamcii, which gives V'eng thangci -> Zzyxwqnp tqntchy "thoroughly", and has an exact cognate in What tînhs "health", and a near-cognate in tymñymas "total".

The augmentative stem of the verb is used for periphrastic forms, but a regular nominalization would also be accepted: qoxuña aqo su θokaru... Note that su is used for personal productions.

*θ is a very marginal part of the reconstructed Proto-Vengic inventory; it's reconstructed for a handful of grammatical particles where Vengic shows *h and all other branches show *t.

Qaa! Ñewo xuñaxuña miruya za ghetama
qaa | ñewo xuña~xuña miruya za ghetama
lo | sky pick_up-MAJ fishing_net for sea
Lo! The sky wooed the sea

Xuña "pick up" has been adapted as an auxiliary meaning "begin", a common change in the region, and one shared with most of Hathic. (Development of the minor into an auxiliary is absent from Whatic - cf. What xhui "gather berries, scavenge" - but Whatic does show this development for the major form, which in What is, somewhat irregularly, ñuanh. This may just reflect early loss of the minor form in Whatic, however, followed by reborrowing as specialized vocabulary; the major form of xhui is uxhui, which, if we assume assimilation from an earlier form *uihuüi [cf. dialectal uifui], could be regularly derived from Proto-Whatic *fwooyfwooy.)

The Proto-Vengic tribes being coastal, fishing prowess was attractive.

A na kuma nikuto e zumanaa
a na kuma nikuto e zumanaa
3 SEQ bear five E brother
And she bore five brothers

Nikuto means both "five" and "hand"; since it's sort of a noun, it takes ezafe. The third-person pronoun is just a; this is both unusually short and subject to confusion with reflexes of aqo, the first-person pronoun, so it's often replaced. Hlu, for example, has gendered animate pronouns yoeqnga "he" and yehmahl "she". It has also replaced the first-person pronoun with gwiq.

Zumanaa kam tuungo zura na xuña ngulenga
zumanaa kam tuungo zura na xuña ngulenga
brother into pass man SEQ pick_up be_yellow.AUG
The brothers grew to men and began to feel jealous

"Yellow" for "jealous" is a common semantic equation in the languages of the area, shared also with Tsi and most of Hathic.

Na dughu ii xoma qoghamto ataama lum
na dughu ii xoma qo-ghamto ataama lum
SEQ ask give.AUG right ADJ-head and parents from
And asked of their father and mother their birthright

Adjectives with qo- don't take e.

Ghetama qa ñewo ghero zii doogham e goom
ghetama qa ñewo ghero zii doogham e goom
sea and sky DISTR give.MAJ rule E land
The sea and the sky gave to each a property

The eldest, Gąą Dbé, in his property

Made himself a hot land

For he loved the sunshine

And he made from dust the first dog to be his companion

Lo! Sri Takʰ, the second

Loved the darkness, and upon his land

Made himself a shady jungle

And he made from mud the first leopard to be his companion

(etc.)

(in progress)

I was apparently working on another Hluic language's sound changes. What's it like?

Thing is a conservative Hluic language spoken in the Pqchatggvx river valley, about 50 miles inland from Gejaehl. It shows less Hathic influence than coastal Hluic languages like Hlu and Gyitha, and is less agglutinating. It is also tonal; however, unusually for a language of the Swamp Continent, its tonal contrasts developed primarily from syllable-counting processes: preinitials (and initial geminates, which presumably took prothetic schwa) conditioned rising tone, and the absence of a preinitial conditioned mid tone. Most preinitials were then lost, but new ones were developed through compounding -- Thing has strict word-final stress, with the exception of postclitics, notably including verbal person marking.

Thing is also relatively morphologically conservative: unlike Hlu and Gyitha, it preserves the minor/major distinction and the old verbal noun system, and has innovated a major form of the verbal noun.

/p b ⁿb t d ⁿd tɕ dʑ ⁿdʑ k g ⁿg ʔ/ <p bb b t dd d c jj j k gg g q>
/f θ s ɕ h/ <f th s x h>
/v~w ð z j ɣ/ <w dh ss y gh>
/m n ɲ ŋ/ <m n ny ng>
/l ɻ/ <l r>

/a e ɤ o i u ɿ ʅ v̩/ <a e v o i u z zr w>
<á ā à ǎ> <a/â ā à á>

Only /ɐ i u/ <v/a# i u> and /ɿ ʅ v̩/ can appear in unstressed syllables. Voiced and prenasalized plosives can't appear syllable-finally.

[zɿ kʊmɪˈθál jɪlˈkrɤ́ʲɲ ɪˈʐʅ́ ˈpɻámðɐ ɲǐɕ ðɤ̄ tsɿmɣɐnjiˈcǒlmǎ]
Z Kumithal yilkrvny irzr prâmdha nyíx dhv tzmghvn'yicólmá.
z Kumithal yilkrvny i=rzr pram=dha nyíx dhv tz-mghvn-yi-c\jól=má
DEF Komthag myriad of=thing create=3 mankind 3PL in_order_to-D-VN\nourish.MAJ-BEN.3

[θjū ʔɪˈɲǐɕ ʎi mɣɐngjʊⁿdɤmɤˈⁿdǐmǎ ɟǎɻðɐ]
Thyū inyíx li mghvn'ggyudvmvdhímá jjárdha.
thyū i=nyíx li mghvn-gg<y>u-dvmvdhí=má jjar=dha
nothing of=mankind REL D-<be_able>return-give.VN=BEN.3 NEG.COP.IPFV=3

Hathic has some problems - the proto-language doesn't match up with what little is currently known of the existing languages.

- Gehui has 60 consonants and 38 vowels. Rau has an unusually large segment inventory for a current-year Hathic language, with about 20 consonants and 11 vowels. Even assuming a Polynesian-tier system collapse, which we already know exists anyway, this is a little hard to justify.
- Gehui has clicks. Current-year Hathic languages generally do not.
- Gehui forms a sprachbund with pre-Tsi, but current-year Hathic languages don't look like Tsi at all and are spoken far from Tsalaysia.
- Gehui seems to be phonologically intermediate between Vengic and Tsiic, but the resemblance to Vengic is a secondary development within Vengic.
- I don't think the properties of the Rau sound system mesh very well with clicks.

I've considered separating Gehui from Hathic entirely, but Pre-Gehui already exists in some vague form. So it might be best to decrease the time-depth of Gehui and say it's just the branch of Hathic that fell under Tsi influence and gained clicks. The Gehui Urheimat is still somewhere in Tsalaysia, but Tsalaysia is a pretty large place, and we can probably postulate some sort of Sundaland.

But Gehui has a lot of clicks! So they're secondary developments from consonant clusters - Gehui has CVC as its typical root structure anyway, so compounds will naturally create clusters. The clicks in the ~classifiers can be explained with compounding from some sort of k- prefix, maybe originally marking definiteness. (Cf. Rotokas, where the "definite article" is part of the citation form of nouns and pretty much always appears, even co-occurring with the "indefinite article". Also Austronesian.)

So! What do we know about Pre-Gehui?

me in 2013 wrote:Initials:
p t c k ʔp ʔt ʔc ʔk ʔ
b d ɟ g ʔb ʔd ʔɟ ʔg
m n ɲ ŋ
f s h
l j w ʔl ʔj ʔw
ʘ ! ǁ ǂ ʔʘ ʔ! ʔǁ ʔǂ nʘ n! nǁ nǂ

Rimes:
a e o i u + length
ai au oi eu

Two suprasegmentals: vowel retroflexion (spreading over to coronals, which turn to retroflexes) written with -r, and 'retraction' applying to a syllable as a whole, written with an underdot on the vowel.

Retraction is fine; possibly this is about vowel glottalization or something, and became pharyngealization, cf. Arabic and probably Lolo-Burmese. Then it becomes tone, syllable tone simplifies into word tone, and word tone simplifies into stress accent for Rau. The clicks are of course absent.

Glottalization actually looks like length, so let's say all consonants can be geminated (although Ubghuu also has initial ʔC clusters, doesn't it?), and we get something like:

/p b t d c ɟ k g/
/m n ɲ ŋ/
/ʔ h/
/l j w/

(maybe h was actually s)

The vowel system is basically normal, aside from r-coloring, which is a little weird. So maybe these are from syllable-final coronals. Syllable structure is (C(:))V(t n) - nasals, why not. I forget if r-coloring and retraction can co-occur, but I don't think so, so let's add -k -ŋ; obviously Gehui loses coda nasals unconditionally, and then t > r, k > ʔ > ˀ. Possibly -k is lost everywhere else too.

I have a question, out of curiosity. Is the Allosphere a rebranding and expansion of the conworld I remember you used to call The World back in the day, or is it a fresh conworld started from scratch?

The former, yes. When Nort joined in with Yng and myself, we decided that it made the most sense to graft what we had onto his conworld, because from a conworlding perspective, it was the most complete -- although "complete" is of course a strong word. The name is older than that, and originates from a years-old joke on Annie.

Hallow13 and I spent most of yesterday evening fleshing out a revisèd and newly endorsèd sketch for Proto-Kangshuic.

Phonologically, Proto-Kangshuic is characterized by a distinction between "type A" and "type B" syllables, the former being unglottalized and the latter glottalized. Glottalization is best understood as a general property of the syllable and in all likelihood was mostly realized on the final (vowel + coda) as a whole. Type B syllables have a fairly restricted phonology.

Maximally, the PK syllable had a structure of C₁RVC₂, with R and C₂ being optional. We may refer to C₁ as the 'initial', R as the 'medial', and V + C₂ as the 'final'.

Initial (C₁) phonology for Type A syllables, in working orthography:
An underdot or overdot on a sonorant indicates voicelessness. <ʒ> is /dz/ and <c> /ts/; <ḫ> is /x/ and <ḥ> /ʕ/. /ć đ ś ź/ are not really phonemic, but represent a Mandarin-style merger of the alveolars (*c *ʒ *s *z) and velars (*k *g *ḫ; velar nasals are unaffected) before front vowels or yod (*y).

R may be any of *w *r *l *y *ḥ. There does not seem to have been a voicing or glottalization distinction on R. C₁ may be a glottalized or voiceless glide in type A syllables, however. (That is, *lˀát and *lát are distinguished in type A syllables, but not *plát and **plˀát.) There may have been other restrictions on R relative to initials or nuclei.

Nucleus (V) phonology for Type A syllables permitted any of seven vowels *i *e *a *ɤ *ɯ *o *u, with no other distinctions. High vowels *i *ɯ *u seem to have been disfavored before or after the uvular stop *q, but not the pharyngeal *ḥ. *ɯ *ɤ additionally do not appear after velars (including nasals and *ḫ). Resultingly, if there is no medial glide, *k *g can only appear before *a *o *u.

Any or no coda was allowed in Type A syllables. Allowed codas were *p *t *k *q *b *d *g *m *n *ŋ *s *ḫ *z *ḥ *w *r *l *y *ẉ *ṛ *ḷ *ẏ. Again, *q did not appear after the high vowels.

Type B syllables were a bit more restricted. Initials (C₁):
Same as Type A, but no glottalization distinction. Probably the plain sonorants were phonetically glottalized and patterned in sound change with Type A glottalized sonorants, not Type A plain.

R may be any of *w *r *l *y *ḥ. No voicing or glottalization distinction.

No high vowels were allowed in type B syllables (cf. Klallam, which doesn't allow them before a coda /ʔ/), which thus had only four phonemic vowels *eˀ *aˀ *ɤˀ *oˀ. As in type A syllables, *ɤ(ˀ) doesn't appear after velars, and *e(ˀ) merges velar and alveolar obstruents to the postalveolars.

Permitted Type B codas (C₂) were *p *t *k *q *b *d *g *m(ˀ) *n(ˀ) *ŋ(ˀ) *s *w(ˀ) *r(ˀ) *y(ˀ) *l(ˀ).

Orthagonal to the type-A/type-B distinction is tone. All syllables had one of two tones, high or low, regardless of glottalization or other segmental features. We are still not entirely sure how to do both glottalization and tone in the working orthography. Perhaps we will just mark the coda of type B syllables with a superscript ˀ: *dèkˀ, *mjɤ́nˀ.

Early/Pre-Proto-Kangshuic seems to have looked morphologically rather like Bantu or somewhere between Bantu and Romance, with verbs (and nouns?) preceded by a handful of lightly bound inflectional prefixes/clitics not entirely glued to the root. The prefixes clearly had tone, but it seems likely that they did not have codas, or a type-A/type-B distinction: ó-mà-ćéŋ. In most (all?) of the daughters, the prefix system does not really survive. Early during the break-up period the Kangshuic languages undergo an early change known as the "disyllabicity constraint", which forces most words into a disyllabic mold, but not necessarily in the same way. Thus from a verbal complex ó-mà-ćéŋ one daughter may bracket ó-mà | ćéŋ and another ó | mà-ćéŋ, with far-reaching implications for which original morphology is preserved.

dhok wrote: ↑Sat Dec 21, 2019 11:14 pm /ć đ ś ź/ are not really phonemic, but represent a Mandarin-style merger of the alveolars (*c *ʒ *s *z) and velars (*k *g *ḫ; velar nasals are unaffected) before front vowels or yod (*y).

Is the velar nasal being uniquely unaffected by palatalization attested?

Nortaneous wrote: ↑Mon Dec 23, 2019 11:22 pm

dhok wrote: ↑Sat Dec 21, 2019 11:14 pm /ć đ ś ź/ are not really phonemic, but represent a Mandarin-style merger of the alveolars (*c *ʒ *s *z) and velars (*k *g *ḫ; velar nasals are unaffected) before front vowels or yod (*y).

Is the velar nasal being uniquely unaffected by palatalization attested?

Sigh...the Index Diachronica is almost completely devoid of anything on Sinitic! I am almost certain the answer is "yes". In any case, obstruents and sonorants tend to operate sufficiently differently that we can hand-wave this. (Consider, for example, /w/, which is just as velar as it is labial, but considerably less likely to undergo palatalization than /k/.)

I never got around to making Zot as Shit Khmer so it is now Shit Skouic

Proto-Vengic	V'eng	Gejaehl Hlu	Thing[/cell]	Zot	What
qavengi	vʼëŋ	ʔwɤɲ	ʔwíɲ	wẽ̂	ɲʼaŋ
mekaa	mka	mɯy	mē	mɣɛ̌	məkiy
hotenga	theŋ	hɯɲ	θéŋ	ðẽ́	mənʼyah
nakasa	nkas	ŋgɔɬ	ŋgáθ	nɣǐ	ləkiy
fukee	khe	hu	fv̩̄	fɣí	wəki̤y
cim	ciŋ	se	sí	zĩ́	sĩy
cuqaca	cʼac	ʔɔc	ʔāc	yí	səʔas
miruya	mri	ywɤy	mrí	mlòza	mərɲah
mayaa	mya	mæ	máy	mlɛ̀	mah
tongaa	ŋʼa	ɗoŋ	tóŋ	ðɛ̃́	təŋiy
tenga	teŋ	ɗɯɲ	téŋ	kyɛ̃̀	tiŋ
tamnga	taŋ	ɗɔŋ	táŋ	ðɛ̃́	nʼẽŋ
pera	per	ɓi	pér	plɛ̀	pir
yam	yaŋ	yɔ	yá	zɛ̃̀	ɲẽy
kusaa	sʼa	ɠɤɬ	kɤ́θ	hɛ́	pəɲiy
maxure	mxör	ŋgwi	ŋgɤ́r	mlì	məwra̤h
pimkimra	kʼër	cwɤ	cwǎr	ɣẽ́	mʼəŋna̤h
rakaru	skor	yɯ	gɭ̩́	lɣǒ	rəkyar
terumxu	trux	tɔh	trū	ðé	tərɨ̃h
tuuba	tu	ɗɤp	tūp	wɛ̂	tup
yinapa	nyap	ynɔw	nyǎw	nɵ̀	ɲəlap
zipuNa	spëm	ywɤ	wyúr	wǒ	ləmʼãh
Nevo	drö	yɔw	ráw	ɲù	yĩm
kimu	kim	ʄim	cím	wû	kiym

I am not entirely sure what is happening with *tenga; there is probably a bug in the sound changes. Haedus is very difficult to work with; there is enough unexpected behavior that one must be superstitious. I don't use (), {}, or zero anymore, and I suspect the 'or' keyword has also been causing trouble, getting unpacked into discrete rules in unintuitive ways.

Anyway! The phoneme inventory:
/p t k ʔ/ <p t k>
/β ð z/ <b d z>
/f s h/ <f s h>
/m n/ <m n>
/w l j ɣ/ <w l y g>
/ɛ a ɔ e ɵ o i u/ <x a o e z ou i u> + nasality <-ng>, neutralized after nasal onsets
/˥ ˩ ˩˥ ˥˩ ˥˩˥/ <t p j x f> (word tone - ˥˩˥ can't appear on monosyllables)

Initial clusters are of the following forms:
- h + voiced non-fricative consonant
- consonant + w j ɣ
- bilabial + l
- tl

Coda consonants do not exist. /f/ could probably be analyzed away, I've already done this for /ɬ tɬ ɲ/. But /s/ can't all be /tj/, because morphology restores [sj] clusters.

/h/ can't appear before /ɵ u/; when morphology would demand this, it becomes /f/. /t/ can't appear before /i j/; when morphology would demand this, it becomes /s/. (TODO: haedus this)

/hml hnj/ also exist. ɣ is generally pronounced as a uvular. There are some coalescence processes: /hɣ/ is often [χ ~ q], written <q>, and /tl hl ny/ are [tɬ ɬ ɲ].

(TODO haven't coded disyllable coalescence in Haedus yet. basically ɣ > w / _u, then iyV > yV, uwV > wV, Vyi Vwu > Vi Vu > monophthongization; this is why /o/ is written <ou>. probably ɛ merges with e in this context, then ai ei oi > ɛ e ɵ, au eu ou > ɔ ɵ o. ɵi ɵu should probably just become i u. LHL probably merges with LH, whatever, any tonal information after the second syllable is discarded entirely but words shouldn't be that long, major verbs are probably lost except for a handful of lexicalized ones)

One notable characteristic of Zot is that, due to heavy lenition or loss of C2 in initial clusters, common verbs take extensive initial consonant mutation for person.The Proto-Vengic person marking system was preserved, including the plurals, and extended with distinct third-person singular markers *tem- 3SG.M and *pa- 3SG.F and an innovative first-person exclusive form *lim- (Zotic has a clusivity distinction, formed from a polite first-person plural pronoun *(te)lim preserved otherwise only in Vengic and various outlier languages), giving the following system:
So, 'do' (*mo): 'eat' (*Neme): 'go' (*kange): 'collect, gather' (*hemnare): Thankfully, there are not many such verbs, and some aren't completely irregular. 'be under' (*mako), for example, has been regularized into the paradigm of 'do', albeit with a different vowel: Tones are also mostly predictable. If the 1SG and 3SG animate forms have a high tone, so do all forms except the 2SG and 3PL, which have a rising tone; and if the 1SG and 3SG animate forms have a falling tone, so do all forms ecept the 2SG, 3PL, and 1EXCL, which have a low tone. The 3SG inanimate varies.

Complicating matters, transitive simple verbs take object marking, which has some effects. Consider *zera 'hit' - dealing only with singular objects: Note that 1>1 and 2>2 forms do not exist. At some point there will be a reflexive.

The fact that almost all conjugated forms of verbs (except for transitive verbs with plural or third-person animate objects) are monosyllabic causes a great deal of homonymy, which is dealt with through compounding. But I'll get to that later.

Nortaneous wrote: ↑Sun Dec 29, 2019 10:51 pm I never got around to making Zot as Shit Khmer so it is now Shit Skouic

It is now shit what?





Oh.





Heh. Sounds like a family for the refined Sino-Tibetan connoisseur.

Huh, I've just been reading grammars of Skou languages. (I really like the I'saka one by Donohue and San Roque.)

CV verbs with unpredictable initial consonant mutation for person marking is extremely Skou; the paradigms are p much Wutung but less so, and so far without suppletion. consider the Wutung paradigm for 'hit'

OK, let's go back to Hluic. Specifically, person marking.

Hlu is a little boring with person marking. It has:
- A marker of egophoricity (actually two, syncretic with perfective/imperfective)
- A first-person subject agreement marker (only used in questions and for unintentional actions, because suffixal person markers are incompatible with egophoricity)
- A second-person subject agreement marker (circumfixal, with a prefix inherited directly from Proto-Vengic, but this is generally lost in Gejaehl outside fossilized forms)
- A third-person subject agreement marker
- Pluractionality alternations

So your basic paradigm is: (using yu 'hold') In Bor, this is a little different: (note n-r > ⁿd) The main complication in Bor - also in Gejaehl, but much less so since the prefix is less common - is that the egophoricity marker only blocks suffixes. So you could say, for example:

wuya tau oennyim
βuja tau̯ ɤ-n-ɲim
beer EGO.PFV PST-2-drink\PL
we (inclusive) drank beer

This is a little weird! But verbal agreement acts a little derivational in Vengic in general. This also happens in Gejaehl, in the five or so verbs that still preserve a 2-form.

Anyway! Since the 1- and 3-forms were lost in Proto-Hluic, the 2-form expanded to mark the presence of a second-person subject, and frequently expanded further from there. I was going to write about Thing but then I realized that the thing I was going to use for inspiration is actually boring so I'm not sure what I'll do with that

The only subject agreement prefixes that were preserved into Proto-Hluic were the 2SG and the 1/2PL. Many Hluic languages, including Hlu, went on to lose the 1/2PL, but Thing preserved them and generalized them to the 3PL. Proto-Hluic innovated 1/2/3 suffixes, unmarked for number, straightforwardly from the pronouns, but these suffixes couldn't co-occur with the egophoric particle.

(What's the Thing form of Hlu tu? Well, the Hlu form has to be slightly irregular (loss of aspiration, which I think makes sense for common unstressed particles) from *dughem. So we have ddvū.)

As in Gejaehl Hlu, the 2SG prefix was simply lost. The interesting development is in the 1SG. Since it couldn't co-occur with the egophoric particle, its use was limited to questions, inferences ("I must have had too much to drink last night"), and verbs of unintentional action. In Thing, the use of the *1SG was expanded, and the non-volitive sense was lost, so -yeq has become a marker of indirective-reportative evidentiality in all persons, has fused with the Hluic question marker *ɤŋ to give -yíng, and is also used in if-then constructions:

(...the ZBB font is broken, so I'll use a tilde instead of a caron for rising tone)

woyā nyũma
[wʊˈjā ɲɤ̂mə]
beer drink-3
he is drinking beer

woyā nyũmyeq
[wʊˈjā ˈɲɤ̂mjɪʔ]
beer drink-INDIR
I guess he's drinking beer

woyā nyũmyíng
[wʊˈjā ˌɲɤ̂mˈjíŋ]
beer drink=Q
is he drinking beer?

woyā nyũmyeq dda gwe nyũmyeq kél
[wʊˈjā ɲɤ̂mjɪʔ də ⁿgwɪ ˈɲɤ̂mjɪʔ kél]
beer drink-INDIR if 1SG drink-INDIR also
if he's drinking beer, I'll drink too

(gwe is the unstressed form of goúy)

As for the rest of the subject agreement morphology, 2 -o / -wo after vowels, 3 -a / -xa after vowels, PL th(u)-. These still can't co-occur with the egophoric particles, which are ddvū and the other one, so you have distinctly un-European constructions like:
woyā ddvū yáy tí
[wʊˈjā ʙʉ̄ jáj tí]
beer EGO house be_in
I have beer in my house

(ddvū is often reduced to [bʉ] ~ [ʙʉ] ~ [βʉ])

First, a few new verbs:

'pick up' < *xuNa (regular form would be tvnyap but disyllabic forms are a little weird so presumably this is leveled by analogy - or maybe there should be ʔC clusters...)

'buy' < *beexe Zot allows verbal compounding. When two person-marking verbs are compounded, both take person marking:

yxngnyaf
[jɛ̃˥˧ɲã˥]
yxngj-nyaj
collect.3M-pick_up.3M
he gathers items by picking them up (e.g. from the ground)

When a person-marking verb is nominalized, it's nominalized in a person-marked form:

yxngnyamqf
[jɛ̃˥ɲã˧mɔ̃˥]
collect.3M-pick_up.3M-NMLZ
he who gathers items by picking them up

nanyxmqf
[nã˥ɲɛ˧mɔ˥]
collect.1-pick_up.1-NMLZ
I who gather items by picking them up

If a nominalized verb has an object, the nominalization is put in the construct state, which is marked by a tonal shift and (where possible) an infix <y>, a common development in Vengic:

jwvyi pimqj
[ʔwɵ˥ji˧ pi˧mɔ̃˩]
book CONSTR-buy.3M-NMLZ
he who bought the book / the buyer of the book

The oblique postposition fut may also be used: jwvyi fut pimqj. Cf. pimqp 'buyer'.

'Buy' only marks subject; only verbs whose objects are typically animate developed transitive marking. But transitive marking is also preserved in such cases:

zomqf
[zo˥˧mɔ̃˥]
hit.1/3M>2SG-NMLZ
I/he who hit you

zisingmqj
[ʑi˧sĩ˩mɔ̃˩]
hit.3M>3M-NMLZ
he who hit him

Maybe <n> for nasalization, to follow Zzyxwqnp. Anyway. Relativization on subjects is easy to handle:

zyumqf lxp swxnp
hit.MSG>1SG-NMLZ man sleep
the man who hit me is asleep

In this case, zyumqf swxnp would work just as well, since the subject is clearly indicated. But for a non-inflecting verb:

swxnmqp lxp zyuj
sleep-NMLZ man hit.MSG>1SG
the man who is asleep hit me

However, some non-inflecting verbs preserve the augmented form as an older verbal noun, and this can be compounded with 'do to', which conjugates as follows... ah, fuck, I've forgotten where the transitive forms came from and will have to reverse-engineer them

What makes very productive use of reduplication, infixation, incopyfixation, and so on. How could this have developed?

Well, we already know that the major forms (reduplicative) lexicalize at some point. So maybe there are other reduplicative constructions.

There's the Proto-Vengic "future", actually something in the ballpark of an obligative. But most branches folded other prefixal forms into this. So if minor + future makes something, let's say you have *kame > käm (or whatever, I'm just applying sound changes from memory here), *kame-cakame > kmskäm. So that's one form. Minor + minor (as opposed to major) would of course give kmkäm. For trisyllables, *kameta-kameta > kmtkmet, which is a little tedious to say, and reduplication typically targets edges anyway, so it's just initial + final; the minor + future form does the same, so ktkmet, ktskmet.

Infixes are a little harder. Maybe there's an IE deal with preposition+verb compounds, or noun+verb or whatever, and those lexicalize. Prefixes come between the preposition and the verb, so if you have *kupo-ngeera 'stand behind' + future > pŋar, psŋar. But could this analogize out to verbs with complex onsets? (There are epenthetic vowels everywhere, so /ktkmet/ is [kətəkəˈmet].)

Unhdof lup unj yutlvj xsinmqnp zxp yxtvnt.
Yxtvnt gxp pxndlxmqnp zxp unhdof gus nvnsinj.
Zxhlup. Zxhlup. Zxhlup. Zxhlup. Zxhlup. Zxhlup. Zxhlup.

a Burmeso-like system would be diachronically reasonable for Rau, but what even is Rau. let's revise the Rau phonology again, oh boy!

we know we're starting with /a e o i u/ for the vowels; my notes say /a ɒ e ʌ o i ɯ/ but this seems unreasonable, and I think the last revision was similarly bad. probably there should be diphthongs and so on that in the capital dialect merge with the reflexes of vowel + voiced stop. plausibly there's a conservative liturgical register whose orthography can basically also be used as pan-dialectal linguistic transcription (modulo morphosyntactic differences etc.), e.g. zeu ~ zeb. the conservative variety would presumably also try to maintain the qualities of the unstressed vowels, which in Bakhzon Rau are reduced to schwa if not totally deleted, so e.g. k-ema (if it's Burmeso-like, presumably verb roots are generally vowel-initial) would be [keˈma] in Liturgical Rau and [kfa] (with m > f after obstruents, cf. m > w in Seri) in Bakhzon

Consonants:

the Bakhzon Rau consonant inventory has been basically settled for a while now:
/p t k q/
/b d̪ dz g ɢ/
/ɬ s ʃ x χ/
/m n ŋ/
+ an unspecified inventory of clicks, possibly even including bilabial clicks and a back closure contrast between velars and uvular affricates. Rau mostly doesn't preserve the clicks present in earlier stages of Hathic - most of Hathic doesn't preserve the clicks and it's possible that they were a later introduction, but if they were, they would've been introduced in the ancestor of Rau - but they were occasionally preserved, and were reinforced by loanwords from Tsi, as well as the "harmonic clusters" in Amqolic and probably consonant clusters of the form stop + velar in Kangshuic

it's possible, and likely on regional grounds, that /ts tʃ/ exist. the "velars" are actually front velars with distinct palatalization before back vowels, or at least /a/

probably most of the stop series in earlier Hathic collapsed into voiced, and the voiceless series is descended from the aspirates, or something

allophony is maybe shaped by contact with Narng in the usual mode of unconscious convergence and conscious esoterogeny that one expects with two groups in close proximity who do not like each other very much. but maybe it isn't, since this is probably 200ish years after the war where the former Rau capital got deleted

Vowels:

there were long vowels at some point. my notes imply some kind of massive vowel shift the details of which I won't work out here, but in Bakhzon this is complicated by the development of strong stress accent, vowel reduction, and the phonotactic tendencies of American English combined with the phonological perversities of the Cisjumna region. namely, long vowels break into diphthongs and preferential coda syllabification develops, but the contrast between voiced plosives, voiced fricatives, and resonants does not exist

so for example *kiː > kig [cɪj], or something like that. presumably there was also open syllable lengthening. the overall effect is that Bakhton Rau sounds quite a lot like English if you ignore the uvulars and clicks, but is completely different in every other way

a related effect of this is that sequences of two voiced consonants are prohibited, at least in codas. so /ebt/ (something like [ewt]) is licit but /ebd/ isn't. I'm not sure if stop clusters in coda position should exist tho; if they do, it seems reasonable that this would be Bakhzon-specific

an inventory of /a e ʌ o i ɯ/ seems reasonable to assume here, since that's what existed in the earlier iteration, anything else would be difficult to write in ASCII, and my alt key is broken

Basic morphological plan:

some things, such as verbs and possibly nouns, take noun class prefixes which are descended from the Hathic classifiers. possibly this entails verbal marking of definiteness and/or number in addition to person and noun class

there's also a suffixing gender system, which developed later, primarily affects adjectives, and is unrelated in form to the noun classes; probably it absorbed old augmentatives or diminutives or something in addition to semantic gender

Rau forms a contact area with Narng, a handful of peripheral Amqolic languages, and probably parts of Kangshuic. they are additionally aware of the Tsi. this all implies:
- noun case (Amqolic, Kangshuic, maybe Narng - case doesn't appear in Zzyxwqnp or similar languages like Zot but does appear in Hlu)
- verbal person marking and noun class agreement (Amqolic, Kangshuic, Vengic - in Amqolic the system is slightly difficult and I haven't really worked it out)
- bipartite aspectual contrast with a distinction between inherently perfective and inherently imperfective verbs and light verb constructions for marking? (Kangshuic??? this could very easily develop in Vengic as well)
- subordinate clauses can basically be patterned on Kangshuic because I am not good at them
Bakhton Rau is nontonal, but it's also possible that there were earlier tonal contrasts that it just happened to lose - it probably depends on the details of Rau migration, because Amqolic substrate could've removed it. it's also possible that there's just a pitch contrast on the stressed vowel or something
it's also worth giving some thought to the details that might arise from the particulars of the development of this system from the totally different and mostly analytic one of earlier Hathic but I am not going to do that yet

unrelatedly I've completely forgotten how Zot person marking works in transitive verbs, will have to redo that from scratch at some point