The Allosphere

[cedh]

Possibly one of /l r/ exist in loans or something.

Or maybe /ɬ d/ have allophones [l ɾ] in some positions, so that foreign liquids can get imported as /ɬ d/...?

[Nortaneous]

Possibly one of /l r/ exist in loans or something.

Or maybe /ɬ d/ have allophones [l ɾ] in some positions, so that foreign liquids can get imported as /ɬ d/...?

This is a good idea - I'm not sure where Rau is spoken yet, but it's probably on the mainland near Amqolic, to which it has some phonological resemblance, and interchange between liquids and plosives is common in that region: Rengni and Narng d > l (followed in Rengni by s z > θ ð > t~f d, j > z, ts dz > s z), Amqoli ɗ > r and j w > ʒ b, Kangshi b d > β ɾ intervocalically, Amqoli b d g > β ð ɣ / V_ P_ (but maybe d should > ɾ instead?)

[Hallow XIII]

There are probably dialects of Kangshi that split *b d => ɓ ɗ (from Cb Cd clusters); => β ɾ everywhere else

But anyway, let's talk about conjugation. Amqolic having a single, fusional polypersonal verb prefix is a funny coincidence -- because so did Proto-Kaam-Yerte. In Ubghuuic the agreement calcifies and becomes a marker of egophoricity -- the egophoric transitive ʔdzi- is etymologically 1s>3s; the allophoric transitive ʔgi- is similarly historically a 3s>3s prefix.

This raises the possibility that this is another DEEP AREAL FEATURE. The problem, of course, is that Kangshuic conforms not at all to this pattern. Kangshi proper, for instance, mostly has agreement prefixes, but it also has one suffix and one tone change as an agreement marker; furthermore, while the realization of the individual person morphemes can be opaque, the combination is always analytic. So how did this happen?

In Proto-Kangshuic, verbs did only have a single agreement slot in the manner of Amqoli, but with third person being always unmarked -- the third-person object affix that existed in Proto-Qoic became vowel glottalization in Proto-Kangshuic and so stopped being productive, marking transitivization in certain verb pairs instead. A syntactic innovation of Proto-Kangshuic was the development of a set of classifier words that obligatorily accompanied each noun phrase. Since these classifiers were originally bleached demonstratives, they occurred in different positions depending on the role of the noun phrase they accompanied; in particular, since Proto-Qoic was SOV, they always preceded relativized verbs. At some point, "classifier incorporation" became an acceptable way of marking third-person objects, and the inherited system started decaying in various ways, usually because of phonological changes resulting in the loss of certain affixes.

The result is that each Kangshuic language essentially has a different hodgepodge of verbal agreement affixes, as well as wildly differing TAM systems, owing to the fact that some languages experienced an Ergative split, with intransitive/imperfective/non-past (depending on language) verbs retaining the old system, and the others switching to morphology originating in verbal nominalizations.

Eahswa is not such a language! But its system is nonetheless interesting:

PERSON	AFFIX
1sO	V-¹(⁶ng)aǎ
2sO	⁸phei-V
1pO	¹mu-V-¹(⁶ng)aǎ
2pO	¹sei-V
1sA	¹go-V
2sA	⁸phei-V
1pA	¹mu-¹go-V
2pA	¹mu-⁸phei-V
1s>2s	¹gös-V
1s>2p	¹pǎ-V
2s>1s	¹ko-V
2s>1p	⁹sei-V
1p>2s	⁸ngäs-V
1p>2p	¹sei-V
2p>1s	⁸phaă-V
2p>1p	¹pǎ-V

The object affixes are used both for transitive verbs with a 3rd subject (which is zero-marked) as well as intransitive verbs where the S is a semantic O. Class markers, in general, precede all other agreement affixes except in 1p>3 and 2p>3 scenarios, where they occur between the plural prefix and the person prefix: ¹dǎ-¹mu-¹rǎ-¹go-¹ʔaǎh, "we killed them".

Unlike basically every other Kangshuic language, Eahswa also allows 3O prefixes to cooccur with portmanteau prefixes: ⁷the-⁸ngä-⁴pǎ, "we are giving it to you" (with consonant mutation from the final -s). However, such morphological ditransitives aren't possible when both the direct and indirect object are non-SAPs; in this case a serial verb construction must be used.

[Nortaneous]

H13: The Amqolic verb prefix isn't polypersonal. There's only one role that's marked; the question is which role it marks.

It occurs to me that I should probably explain the Amqoli noun class system. There are three noun classes, 'standing', 'sitting', and 'lying down', abbreviated T, S, and L. These are to some extent semantic rather than lexical; as the old joke went about I forget which language, a penis is conventionally S, but an erect penis could be described as T.

In Amqoli, these noun classes have become somewhat more semantic than in the other Amqolic languages; the L class in particular has been extended for derogatory use, and some nouns that in Proto-Amqolic were in L have been moved to T or S.

(For a while I thought it would be a good idea to have this extended to the point of being completely semantic, and have ways to shift the personal pronouns etc. between the different classes, but I have since decided that that's stupid.)

On top of this, there is mostly-natural gender. All animate nouns are either masculine or feminine. This isn't relevant to adjectival agreement, but it is relevant to verbal agreement and pronoun selection.

Maybe, to make up for the lack of a 2/3 verbal contrast, there should be some sort of speech-act-participant marker (1/2 vs. 3). But I'm not sure how to derive that.

----

Egophoricity is, possibly, another areal feature, existing also in Hlu (although allophoricity is unmarked) - and in Hlu, it blocks person agreement, with the result that 1SG agreement only appears in questions or involuntary actions.

[Nortaneous]

So! Amqoli noun classes and genders.

Noun classes, of course, govern verbal agreement:

bgul phrus tqare
the man (T) eats jerky

chegi phrus mqare
the mouse (S) eats jerky

ndrya phrus tuqare
the dog (L) eats jerky

They also govern adjectival agreement, but for adjectives the affixes are slightly different:

dgemgera bgul / damoq bgul / tsxug bgul
the green man / the fat man / the annoying man

lgemgera chegi / ndamoq chegi / ntsxug chegi
the green mouse / the fat mouse / the annoying mouse

jgemgera ndrya / jdamoq ndrya / chotsxug ndrya
the green dog / the fat dog / the annoying dog

There are three adjectival declensions, which we'll call K, T, and H. K-declension adjectives mostly begin with a non-coronal, non-aspirate consonant; T-declension adjectives mostly begin with a coronal consonant; and H-declension adjectives mostly begin with an aspirate. However, due to cluster reduction and aspirate dissimilation, these categories are not entirely reliable. Note that voicing assimilation is regressive, so ch-gemgera > jgemgera.

Natural genders, on the other hand, govern pronoun selection. This works pretty much as in English, except all animates take a gender. Masculine is the default. One irregularity, however, is that inanimate T-class nouns always take lu, which is otherwise the masculine.

bgul gemgera -> lu gemgera
the man (TM) is green -> he is green

bozhgat gemgera -> bo gemgera
the mare (TF) is green -> she is green

chxem gemgera -> lu gemgera
the tree (TI) is green -> he is green (!)

chegi gemgera -> lu gemgera
the mouse (SM) is green -> he is green

chxo gemgera -> zda gemgera
the table (SI) is green -> it is green

ndrya gemgera -> lu gemgera
the dog (LM) is green -> he is green

obzh gemgera -> zda gemgera
the swamp (LI) is green -> it is green

[Nortaneous]

v low quality draft map

this (excluding Whatic) is about the area of Vietnam + Laos + Cambodia + Thailand. I don't understand climatology yet but the coastal parts are probably like the Mid-Atlantic, although this is in the southern hemisphere so it'd be swampier up north.

Vengic isn't that much deeper than Romance, but cognates are obscured by compression - imagine Romance if everything was like French and every dialect area of late Vulgar Latin had a different stress system. Zotic + Narngic probably form a subfamily.

black is land (maybe Kangshuic? Whatic looks like it has a Kangshuic substrate), blue is sea. there are probably some Hathic relict areas; haven't mapped those yet.

[dhok]

I have been Officially Recruited for continued development of the Allosphere.

Some of you may be wondering what the Allosphere's astronomical situation looks like. The details are still being debated, but Allosphere is a bit larger than Earth (radius 7100 km) and has a year about one and a half times as long as Earth's, with a day about an hour longer than Earth's. There are two moons, which transit fairly regularly (being in the same orbital plane); it hasn't been entirely worked out how large they are, nor how often the transits will be.

Allosun is approximately 10% larger than the sun, but a little smaller in the sky because Allosphere is farther away; nevertheless, our models (which amount to "run it in Universe Sandbox and see what number it spits out lol") suggest an average surface temperature of about 14 C, a tiny bit cooler than preindustrial Earth. There are a number of rocky planets closer to the sun than Allosphere, and at least two gas giants (probably (?) more). Some of these probably have moons, but since none of them are likely to be detectable before the invention of the telescope, we have not worried overly much about them.

There are two companion stars which we have titled Proximus and Procula until we can get better names for them, both appearing as bright red pinpricks in the sky. Proximus orbits (more precisely Allosun and Proximus orbit each *other*, or a point about halfway between them) at about 37.5 AU (for comparison, Neptune orbits the Sun at about 30 AU). It is a red dwarf about 130 times as massive as Jupiter and likely has a few (undetectable) planets of its own. Because its orbital period is fairly short (only about two Earth centuries), it should be clear from time immemorial that it is a planet, in the Aristotelian sense (a heavenly body that is not a fixed star).

Procula orbits Allosun and Procula at an average distance of about 126 AU; it is a bit heavier and more luminous (though not brighter, because further away) than Proximus. Its orbital period is about 1260 earth years, so hunter-gatherers are unlikely to realize that it's a planet (...maybe, though? It would take about three and a half Earth years or just over two Allospherian years for it to move one degree in the sky, so probably, yes--it would move about 20 degrees over the course of a fairly long premodern lifetime, and would not be in the same constellations at the end of somebody's life as it was at the beginning), but early civilizations making astronomical records over decades will. Like Proximus, it probably has a few planets, but these are not detectable until the invention of rather complicated telescopes or other gadgets and are broadly irrelevent.

Thank you, that will be all. The attached image is a non-canonical work-in-progress.

[Nortaneous]

expansion

Hathic expands first - to Hathenai and along the Ketas Archipelago onto the mainland, where they settle the coastline, round the Zhjumna, and get into the ?desert.

Tsiic, which may or may not form a macrofamily with Hathic (probably not due to grammatical differences, but the phonologies are similar) expands from elsewhere in Tsalaysia, relatively recently, displacing Kaam-Yerte and (if Cuhbi is Tsiic) older Tsiic.

Kaam-Yerte is indigenous to Tsalaysia, didn't do much, and wasn't in the Hathic-Tsiic sprachbund. If it expanded from the mainland, it probably got there from Narngia rather than Ketas.

Kangshuic is in the northern Zhjumna (which should probably have a Kangshi name...?) and expands in some direction or other at some point in time.

Vengic ends up partially in Kangshuic territory but otherwise does a Vietic and crawls up from Cornwall to eat the coastline, completely displacing everything else in Narngia.

In recent historical times, the Rau (with backing from the Tsi Empire) invaded Narngia, tried to push further into the swamp, and were beaten back into the ?desert by a primarily Ziwanic army, leading to the imposition of Ziwan as a lingua franca across the Swamp Continent, although whatever was previously the prestige language of Narngia (Kangshuic unless they're sufficiently mercantile to not go in for prestige languages) probably sticks around as a classical language for a while.

The layout of Tsalaysia still needs to be fixed but the basics shouldn't be affected too much by that.

[Nortaneous]

after reading about Aghem I have decided that there should be a dialect of Zzyxwqnp where the characteristic Swamp-Cornwallish vowel raising happens twice, so ja wa > e o, e o > i u, i u > z= v=, and then z= v= > z@ v@

Kuntqx jje chip yutrqn kqtye chaxmq yep yattznt.
Jjotryp yep yattznt tq kqtye pexjje Kuntqx.
Chatvx. Chatvx. Chatvx. Chatvx. Chatvx. Chatvx. Chatvx.

kv̩ʷto ⁿdʑi tɕʑ̩ jv̩ʷtʂõ koji tɕemo ji jetɨ̃
ⁿdʑutʂɘ ji jetɨ̃ to koji piⁿdʑi kṽ̩ʷto
tɕetfɘ tɕetfɘ tɕetfɘ tɕetfɘ tɕetfɘ tɕetfɘ tɕetfɘ

presumably further developments would lead to full CCV syllable structure and/or redevelopment of coda clusters. Zzyxwqnp has (should have) a Mandarinoid disyllabicity constraint (which could be violated in certain environments... maybe noun incorporation? or a large class of pronouns/determiners?), so common root extensions could collapse and condition some things

standard Mandarin uses diminutives a lot, right? hm

pfot ⁿdʑi tɕa bjotʂompfi tɕembvi jətəŋ
ⁿdʑutʂai ji jətəŋ topfi piⁿdʑi pfot
tɕepfa tɕepfa tɕepfa tɕepfa tɕepfa tɕepfa tɕepfa

wun-mwan tywe-wen-chi ywe-te tya ddi-wan wa pu-te syi u twa. ga-ten tywe-ti wa ge-ge tyi ya-ga ddye kwu-ha wu he-chi kye mun na gwu-gwa ke
v̩ʷ.mõ tɕu.wũ.tɕə ju.tɨ tɕe ⁿdzə.wõ wo pv̩ʷ.tɨ ɕʑ̩ v̩ʷ to || ga.tɨ̃ tɕu.tɨ wo gɨ.gɨ tɕʑʷ je.ga ⁿdʑi kv̩ʷ.xa v̩ʷ xɨ.tɕə ki m̩ ⁿda gv̩ʷ.go kɨ
vɨmoŋ tɕumtɕa jutə tɕe ⁿdzoŋ wo pfɨtəɕvɨt || gatəŋ tɕut wo gəg pjə jega ⁿdʑi pfa vɨ ɨtɕa kim ⁿda bvukkɨ

(+ tones of course; presumably this is where the weird floating tone stuff would start to develop, or at least the development of a lot more tones, and then root stress -> word tone but plenty of tonal ablaut from lost near-affixal tone)

this is probably what Zzyxwqnp ends up like in the time of Enze

[Hallow XIII]

Would you like some conlang stuff? Here is some conlang stuff.

This one is a Kànnow (you remember Kànnow, right?) translation of a Hlu fragment of the Zzyxwqnp epic that tells of the founding of Zzxzzyx. Very likely an anachronism, but it was nonetheless a handy text.

Pvpchqpye Vtsznxmqpye Zzxzzyx

Řqʼoqʼy! mypʼyyksʼòqduuwow gròłsytjaawow, małebosqʼaqʼowow,
Shquqou! Demon of the north, who laid waste to all he beheld,

řqʼoqʼy

Shquqou!

mpʼy-

demon(through-

i-

NF.ACT.NP-

ksʼò-

shadow-

qdu-

go-

ow-

C1A)-

ow

C1.DEF

gw-

VEN-

ro-

NF.ACT.P-

łsytʼ-

north-

ja-

come-

ow-

C1.A.SG-

ow

C1.DEF

t-

P.PL-

malʼ-

truly-

he-

FP.IND.A-

bos-

everything-

qʼaqʼ-

destroy-

ndw-

C8.P.PL-

ow-

C1.A.PL-

ow

C1.SG

Shquqou (< Rau /ʃqʌqɯ/) is simultaneously the protagonist and the main villain of the Vtsznxmqp, somewhat like Genghis Khan in the Târikh-e Jahângoshây, if the Târikh was bad poetry rather than prose. A historical figure, Shquqou was a priest of Qapi Mongkoush who assumed command over a combined Tsi and Rau army attacking the Narngic port city of Big Yeet and subsequently turned out to be both a religious fanatic and completely insane. In consequence, the relevant part of the Vtsznxmqp is mostly a long list of his war crimes interspersed with records of the sermons he delivered to the corpses. In this verse, he is introduced with the appropriate dramatism.

Linguistically, we are introduced to several specialties of Kànnow already. The word for "demon", "evil spirit" is literally "who moves through the shadows", originally possibly a taboo term, but now firmly lexicalised. Kànnow does not like overusing its possessive construction, instead producing something like "Shquqou! Demon who came from the north, who destroyed everything".

An artefact of Kànnow's morphological system is the frequency with which epithets end in -owow (a class 1 nominalization of an active verb with a class 1 agent). The preverb malʼ- in the final epithet serves as emphasis: who laid waste to all he beheld; its realization as mał- is due to the fact that except for the final element, MOA distinctions are neutralized in clusters. In this way, /h/ (when it occurs anywhere but the beginning of a word) acts as a sort of zero consonant, whose only surface realization is that it triggers cluster mechanics.

Řqʼoqʼy sgałán třpinyłr'eqnènowow, iġytktʼaaja Qʼotʼnasòo tkìmbyyja łṇàk'àksuṛġaán
Shquqou who killed myriads, he is evilest among the red spears of Katnahl.

řqʼoqʼy

shquqou

sgał-

myriad-

nw

C4.PL

t-

P.PL-

rpi-

NF.CAUS.IND-

nal-

man-

r'éq-

die-

nen-

C4.P.PL-

ow-

C1.A.SG-

ow,

C1.DEF,

i-

POSS-

ġatk-

spear-

tʼa-

C5.D.SG

ija

C5.DEF

qʼotʼnas-

Katnahl-

uu

C5.ADJ.SG

t-

P.PL-

kì-

NF.MP.NP-

mbyh-

be_red-

ah-

C5.P.PL-

ija

C5.DEF

ł-

D.PL-

ṅk'à-

most-

ksw-

INF.MP.FP-

yṛġa-

evil-

ah-

C5.D.PL-

nw

C1.P.SG

"Killed myriads" is more explicitly specified as "myriads of men" by incorporating the word for man into the verbal complex. Note the height harmony: -rpi-nal- assimilates to -rpi-nyl-; this process goes left-to-right (with two harmony sets < e a o > vs < i y u >) and is blocked by clusters. There are morphemes that are exempt from height harmony; a prominent example is the class 5 definite marker -ija (and not *-ijy).

The superlative is constructed by adding a dative to the normally univalent middle "be evil": "he is most evil to the red spears of Katnahl" (< Rau /qʌtnas/, 'people of the qutna', a generic term for the Rau religio-cultural zone, named after a holy plant of Qapi Mongkoush).

The zero consonant nature of /h/ can be seen in the fact that it disappears between vowels. This phrase contains a particularly egregious example: |mbyh-ah-ija| being realized as mbyyja, with the vowel hiatus resolving from left to right.

Eňeew ijyņuw małġyyblaknòw, rowesenán Pʼiqįeetʼén kmaapckʼołamnòw
He transgressed sorely against great Kömthag, he reduced the great city of Big Yeet to rubble.

eňe-

Ainabe-

ow

C1.DEF

i-

NF.ACT.NP-

jáṇ-

be_great-

ow

C1.A

mal'-

truly-

ġy-

MALF-

he-

INF.ACT.FP-

blak'-

transgress-

nw-

C1.D.SG-

ow,

C1.A.SG,

ro-

NF.P.A-

wese-

great_settlement-

na-

C4.D.SG-

nw

C4.DEF

p'iqįeet'-

Big Yeet-

nw

C4.DEF

kma-

fully-

hap-

INF.CAUS.FP-

ck'ołam-

rubble-

nw-

C4.P.SG-

ow

C1.A.SG

Eňeew (< Kett Ainabe) is a borrowed name for the supreme being in the general Antipodean cultural area; it serves here to translate Zzyxwqnp Kuntqx, the Vengic personification of Nature. "Reduce to rubble", in the Kànnow, is a zero derivation of ck'ołam, "rubble", with causative conjugation.

Kkìpʼymtaán nalén tcʼiņkàpłrytìninuw, sʼaaņusbiłc'aán ġatkuu tcʼiņkàpjandòw
He gathered barbarians and men to him, he made peasants and spears come to him

k-

barbarian(NEG-

kì-

NF.MP.NP-

p'amt-

understand-

a-

C1.A.PL)-

én

C1.INDEF.PL

nal-

man-

én

C1.INDEF.PL

t-

P.PL-

c'iṇ-

DISTR-

gw-

VEN-

hap-

INF.CAUS.FP-

lrytì-

be_together-

nen-

C1.P.PL-

ow,

C1.A.PL,

s'a-

HABIT-

i-

NF.ACT.NP-

ṇus-

earth-

biłc'-

dig_creases-

a-

C1.A.PL-

én

C1.INDEF.PL

ġatk-

spear-

uu

C5.INDEF.PL

t-

P.PL-

c'iṇ-

DISTR-

gw-

VEN-

hap-

INF.CAUS.FP-

ja-

on_foot-

ndw-

C8.P.PL-

ow

C1.A.SG

It is difficult to say now who counted as a barbarian to the original author of the Vtsznxmqp. It is likely that both Rau and Tsi would have been foreign barbarians to the Vengic tribes, so the word "men" will have referred to Vengic or even Ziwanic speakers that were "recruited" by the advancing armies to swell the ranks as they prepared for siege on land.

In Kànnow, a barbarian is someone whose speech is not understood; similarly, as an originally nomadic people, there is no basic word for "farmer" or "peasant", but there are rather people who habitually dig furrows in the soil. Note the non-assimilation of /u/ caused by the deleted /i/: sʼaaṇusbiłcʼaán rather than *sʼaaṇòsbiłcʼaán.

"Coordination" of objects is often achieved by the distributive preverb c'iṇ-, whose semantic effect is, in general, one of deindividualization. Note also that the venitive preverb in Kànnow is relative to the subject of the verb rather than to the speaker.

Tsʼaapraṛkmoṛnènow, tsʼaapkomgokmoṛnènow, malʼopkròqrʼałkmoṛà kkssèṇseka
He threatened them with arrows, he threatened them with swords, they were so afraid to die that they did not scatter

t-

P.PL-

s'a-

HABIT-

hap-

INF.CAUS.FP-

raṛ-

arrow-

kmóṛ-

fear-

nen-

C1.P.PL-

ow

C1.A.SG

t-

P.PL-

s'a-

HABIT-

hap-

INF.CAUS.FP-

komgo-

sword-

kmóṛ-

fear-

nen-

C1.P.PL-

ow

C1.A.SG

mal'-

truly-

ob-

such-

krò-

NF.MP.P-

qr'ał-

death-

kmóṛ-

fear-

a

C1.A.PL

k-

NEG-

ksw-

INF.MP.FP-

seṇsek-

scatter-

a

C1.A.PL

This phrase illustrates, on the one hand, the use of the nonfinite to background information. Rather than explicitly marking the act of scattering as a consequence, in Kannow one instead says: "being so very afraid to die, they did not scatter". It also provides a nice example of Kànnow's capacity for instrumental noun incorporation.

Sła Qʼotʼnasòo kkìp'ymtaaw łecadnènòw, sła Ciłewoo nalow łecadnènòw
He spoke to the barbarians like one does in Katnahl, to the men like one does in Cihlae

sła

like

Q'ot'nas-

katnahl-

uu

C5.ADJ.SG

k-

barbarian(NEG-

kì-

NF.MP.NP-

p'amt-

understand-

a-

C1.A.SG)-

ow

C1.DEF

ł-

D.PL-

he-

INF.ACT.IND.FP-

cád-

speak-

nw-

C4.P.SG-

nén-

C1.D.PL-

ow,

C1.A.SG,

sła

like

Ciłe-

cihlae-

uu

C5.ADJ.SG

nal-

man-

ow

C1.DEF

ł-

D.PL-

he-

INF.ACT.IND.FP-

cád-

speak-

nw-

C4.P.SG-

nén-

C1.D.PL-

ow

C1.A.SG

The word sła combined with something in the adjectival case has roughly the same effect as the French à la would have: he spoke to them à la Katnahlaise.

[Nortaneous]

Some Haedus files.

Proto-Vengic to Proto-Ziwanic:

More: show

mode intelligent
open 'proto-vengic.txt' as LEXICON

C = p t c k q b d j g v z y ɣ f s x h m n ɳ ŋ l r w pʲ bʲ mʲ ʈ ɖ ʂ
A = a ã aː ãː
E = e ẽ eː ẽː
O = o õ oː õː
[IH] = ɪ ɪ̃ ɪː ɪ̃ː
[UH] = ʊ ʊ̃ ʊː ʊ̃ː
I = i ĩ iː ĩː
U = u ũ uː ũː

[VSHORT] = a ã e ẽ o õ ɪ ɪ̃ ʊ ʊ̃ i ĩ u ũ
[VLONG] = aː ãː eː ẽː oː õː ɪː ɪ̃ː ʊː ʊ̃ː iː ĩː uː ũː
[VORAL] = a aː e eː o oː ɪ ɪː ʊ ʊː i iː u uː
[VNASAL] = ã ãː ẽ ẽː õ õː ɪ̃ ɪ̃ː ʊ̃ ʊ̃ː ĩ ĩː ũ ũː

V = [VSHORT] [VLONG] ɐ ь ъ % reduced vowels

[PB] = p t c k q b d j g
P = p t c k q
B = b d j g
[SZ] = v z y ɣ f s ç x h
S = f s ç x h

% normalize ASCII
[VORAL]m > [VNASAL] / _C or _#
ng > ŋ
gh > ɣ

aa > aː
ee > eː
oo > oː
ii > iː
uu > uː

[VORAL]ɴ > [VNASAL] / _

% the sound changes in http://www.incatena.org/viewtopic.php?p ... 2#p1049472 are pretty out of date
% it's assuming sesquisyllabic (CV-)CVC, but Proto-Vengic was actually strict (C)V
% but we can't take this structure as representative of Ziwanic, because Vynyi "is" Ziwanic, by some bizarre standard
% so... let's take the *N split as representative of Ziwanic
% (and fricative reflexes of V < *B + plain-voiced reflexes of B < *MB, but that's everything except Zotic)
% ( + possibly some other early branches)

ɳ > mʲ / U_
ɳ > ɖ

% Length shift to handle CVVCVCV, which is otherwise difficult
[VLONG]C[VSHORT] > [VSHORT]C[VLONG] / _CV#

% Length rule
[VLONG] > [VSHORT] / [VLONG]C_

% now that we've defined Ziwanic, let's get Proto-Everything-But-Vynyi
% which *is* defined by the sesquisyllabic structure
% from earlier penultimate stress
% however, unlike in Hluic, long vowels attract stress
% specifically, the rightmost long vowel

% as usual, we'll use ʜ to represent stress...

% first, stress every long vowel
% if you do 0 > ʜ / [VLONG]_, it won't work word-finally for some reason
[VLONG] > [VLONG]ʜ / _

% then remove all stresses but the rightmost
ʜ > 0 / (CV)+_(CV)+ʜ

% assign regular penultimate stress
0 > ʜ / [VSHORT]_C[VSHORT]# not ʜ(CV)?_C[VSHORT]#

% stress all monosyllables (that aren't already stressed)
V > Vʜ / #C_#

% Proto-Vengic root structure:
% (C1{ă ĭ1 ŭ1})(C2){a e o i u}(C3({ă ĭ2 ŭ2}))
% C1 = p t c k ʔ f s x h m n
% C2 = any C or Pr Fr or labial + y
% C3 = any C except unvoiced non-glottal plosives
% Nasality contrast is lost in unstressed syllables.

% How can we get those clusters? Well, let's say unstressed a > schwa, shorter than all the other vowels, and is lost.
% I'm not too sure about this tho - you get height/length interactions elsewhere in Vengic:
% - in Vynyi, clusters [probably] come from lost *high* vowels
% - in Whatic, final -a is less prone to loss than other final vowels
% But I'd originally planned the C1 vowels to condition height mutation, and who needs that.
% So prestress vowels won't be e o i u, but rather a i u.

i u ĩ ũ > 0 / {[PB] [SZ]}_rVʜ
i ĩ > 0 / {p b f v m}_yVʜ

% Long vowels will interfere with this somehow, of course...
% they'll give you... disesquisyllabic words?
% So let's assign secondary stress.

V > Vʜ / _C{C}?VC{C}?Vʜ

a ã > ɐ / ʜC{C}?_
e ẽ o õ > ɐ / ʜC{C}?_
i ĩ u ũ > ь ь ъ ъ / ʜC{C}?_

a ã > ɐ / _C{C}?Vʜ
e ẽ > ь / _C{C}?Vʜ
o õ > ъ / _C{C}?Vʜ
i ĩ > ь / _C{C}?Vʜ
u ũ > ъ / _C{C}?Vʜ

% One more consonant shift...
g > k / #_
g > ŋ

% Good enough.
ʜ > 0
pʲ bʲ mʲ > py by my

% Not sure what this is. Is it contrastive?
w > v
j > z

% Oh, drop length while we're at it.
[VLONG] > [VSHORT]

close LEXICON as 'proto-ziwanic.txt'

Proto-Ziwanic to V'ëng:

More: show

mode intelligent
open 'proto-ziwanic.txt' as LEXICON

C = p t ʈ c k q b d ɖ z g v y ɣ f s ʂ x h m n ŋ l r ʔ
P = p t ʈ c k
B = b d ɖ z g
[PLOSIVE] = P B
F = f s ʂ x h
V = z ɣ
[FRICATIVE] = F V
[NASAL] = m n ŋ
[OTHER] = l q
[APPROXIMANT] = y v r

[VORAL] = a e o ɪ ʊ i u
[VNASAL] = ã ẽ õ ɪ̃ ʊ̃ ĩ ũ
V = [VORAL] [VNASAL]

A = a ã
E = e ẽ
O = o õ
[IH] = ɪ ɪ̃
[UH] = ʊ ʊ̃
I = i ĩ
U = u ũ

[VREDUCED] = ɐ ь ъ
[VHR] = ь ъ

% Umlaut
E O > [IH] [UH] / _{C}+[VHR]
I U > [IH] [UH] / _{C}+ɐ
A > E / _{C}+ь
A > O / _{C}+ъ

% Reduced vowel loss
[VREDUCED] > 0

% Nasality excrescence and loss
[VNASAL] > [VNASAL]ŋ / _#
[VNASAL] > [VORAL]

% syllabify. CR clusters are onsets, everything else is coda + onset

0 > . / CV_{P B F V}{r y}V
0 > . / CV(C)_CV

b z > v y / _# or _.

% CY loss - homorganic:
iy uv > i u
ɪy ʊv > i u
ey ov > ɪ ʊ
% and heterorganic:
iv uy > u i
ɪv ʊy > u i
ev oy > ʊ ɪ
av ay > o e

% The possible preinitial consonants are p t c k ʼ f s x h m n v y.

ʈ ɖ ʂ > t d s / _C not _[APPROXIMANT] not _q % does ʂ even develop? I don't think so
B > P / _C not _[APPROXIMANT] not _q
ɣ > x / _C not _[APPROXIMANT] not _q
ŋ > n / _C not _[APPROXIMANT] not _q
l > k / _C not _[APPROXIMANT] not _q % velar lateral affricate
r > s / _C not _[APPROXIMANT] not _q

% p t c k are only contrastive before approximants; they debuccalize to ʼ in all other environments.

[PLOSIVE] > q / _C not _[APPROXIMANT] not _q
B > P / {q h}_ or _{q h}

% f x don't appear before plosives. Younger speakers merge them with h.

f x > h / _[PLOSIVE]
x > h / _l

% m n merge for all but the oldest speakers, and don't contrast before labials.

n > m / _{p b f v m}
m > 0 / _m

% Handle final palatalized labials
[UH] U > [IH] I / _Cy
y > 0 / C_#

% Word-final stuff
l > ɣ / _# or _.
ʈ ɖ > t r / _#
B > P / _#

% Orthography
yC > Cy
vC > Cv
y > 0 / {c z y}_
v > 0 / {p b m}_
hC > Ch
qC > Cq

q > ʼ
. > 0

% Some cross-syllable assimilation
r > t / _s

% More orthography
ɣ > g
ɪ ʊ > ë ö
ʈ ɖ > tr dr

close LEXICON as 'veng.txt'

Proto-Vengic to Proto-Hluic:

More: show

mode intelligent
open 'proto-vengic.txt' as LEXICON

reserve aɴ āɴ eɴ ēɴ əɴ ə̄ɴ oɴ ōɴ iɴ īɴ uɴ ūɴ

C = p t c k q b d j g v ð z y ɣ f s ç x h m n ɳ ŋ l r w kʲ xʲ ŋʲ b̃ d̃ j̃ g̃ θ ʔ

V = æ a ɔ e ə o i u ǣ ā ɔ̄ ē ə̄ ō ī ū æɴ aɴ ɔɴ eɴ əɴ oɴ iɴ uɴ ǣɴ āɴ ɔ̄ɴ ēɴ ə̄ɴ ōɴ īɴ ūɴ
[VSHORT] = æ æɴ a aɴ ɔ ɔɴ e eɴ ə əɴ o oɴ i iɴ u uɴ
[VLONG] = ǣ ǣɴ ā āɴ ɔ̄ ɔ̄ɴ ē ēɴ ə̄ ə̄ɴ ō ōɴ ī īɴ ū ūɴ
[VORAL] = æ ǣ a ā ɔ ɔ̄ e ē ə ə̄ o ō i ī u ū
[VNASAL] = æɴ ǣɴ aɴ āɴ ɔɴ ɔ̄ɴ eɴ ēɴ əɴ ə̄ɴ oɴ ōɴ iɴ īɴ uɴ ūɴ

A = a aɴ
E = e eɴ
[SCHWA] = ə əɴ
O = o oɴ
I = i iɴ
U = u uɴ

% normalize ASCII
[VORAL]m > [VNASAL] / _C or _#
ng > ŋ
gh > ɣ
ææ aa ɔɔ ee əə oo ii uu > ǣ ā ɔ̄ ē ə̄ ō ī ū
ææɴ aaɴ ɔɔɴ eeɴ əəɴ ooɴ iiɴ uuɴ > ǣɴ āɴ ɔ̄ɴ ēɴ ə̄ɴ ōɴ īɴ ūɴ

% first syncope (maybe shared with Vynyi to some extent?)
[VLS] = p t c k q f s ç x h
I U > 0 / [VLS]_[VLS]VCV

[VORAL] > [VNASAL] / ɳ_
ɳ > r
z > y
t k > ð ɣ / V_V
[VLONG] > [VSHORT]

k x ŋ > kʲ xʲ ŋʲ / _I

A > [SCHWA] / _CI or _CU
I U > [SCHWA] / _CA

% -- Syncope --
V > Vʜ / #C_#
V > Vʜ / _CV# or _C#
V > 0 / _ not _ʜ
ʜ > 0

% -- Cluster reduction --
N = m n ŋʲ ŋ

[P] = p t c kʲ k b d j gʲ g
[MB] = b̃ d̃ j̃ g̃ʲ g̃ b̃ d̃ j̃ g̃ʲ g̃
[F] = f θ s xʲ x h
R = ð z y ɣ w l r
Y = y w

[P] > [MB] / N_
N > 0 / _[MB]
Nv Nð Nɣ > b̃ d̃ g̃
Nf Ns Nc Nx > b̃ d̃ j̃ g̃
f > h

s > θ
c > s

[F] > h / _C not _R not _Y


h[P] > [P]ː / _
C = C pː tː cː kː
h > 0 / _C not _R
θð > tː
iɴ uɴ eɴ oɴ > æɴ ɔɴ aɴ aɴ
e > æ / _kʲ or _xʲ or _ŋʲ

[VNASAL] > [VORAL]̃
q > ʔ
kʲ xʲ ŋʲ > c ç ɲ
v > w
ə ə̃ > ɤ ɤ̃

% can't figure out how to fix t:xol so screw it
x > ɣ / _V
x > h

0 > ə / CC_C
0 > ə / C_C not C_R not Y_C

close LEXICON as 'proto-hluic.txt'

Proto-Hluic to Proto-Hlu:

More: show

mode intelligent
open 'proto-hluic.txt' as LEXICON

C = p t c k ʔ b d j g pː tː cː kː ʔː bː dː jː gː ð z y ɣ s ç x h m n ɲ ŋ l r w b̃ d̃ j̃ g̃

% retracted vowel contrast emerges later
[VPLAIN] = æ a ɔ e ɤ o i ɯ u æ̃ ã ɔ̃ ɤ̃
[VRET] = æˠ aˠ ɔˠ eˠ ɤˠ oˠ iˠ ɯˠ uˠ æ̃ˠ ãˠ ɔ̃ˠ ɤˠ
V = [VPLAIN] [VRET]

æ̃ ã ɔ̃ ɤ̃ > æ a ɔ ɤ

% handle presyllables
ə > 0 / C_ʔ
ə > 0 / əC_CV
əy əw əɣ > iy uw ɤˠɣ
yə wə rə ɣə ə > i u a ɤˠ ɤ
Cʔ > ʔC

ʔv ʔð ʔz ʔy ʔɣ > ɓ ɗ ɗ ʄ ɠ
ʔs ʔç ʔx > ɗ ʄ ɠ
ð > l

[LABIAL] = p b ɓ b̃ m w
[CORONAL] = t d ɗ d̃ s z n l r ð
[PALATAL] = c j ʄ j̃ ç y ɲ
[VELAR] = k g ɠ g̃ x ɣ ŋ

i > ɯ / [CORONAL]_
æ e > e i
e[VELAR] i[VELAR] > e[PALATAL] i[PALATAL]
e[PALATAL] i[PALATAL] > ɤ[PALATAL] ɯ[PALATAL]

ɣC > Cɣ

[VPLAIN] > [VRET] / _x
x > h / _#
[VPLAIN] > [VRET] / ɣ_ or _ɣ
ɣ > 0

V = V ä å
o u > we wi / _ not _[VELAR]
w > 0 / [LABIAL]_
æˠ aˠ ɔˠ eˠ ɤˠ oˠ iˠ ɯˠ uˠ > æ ɤ o æ ɯ o ä ɯ å % Gejaehl/Bor isogloss - abstract symbols
pː tː cː kː bː dː jː gː > b̃ d̃ j̃ g̃ b̃ d̃ j̃ g̃
ʔː > ʔ
p t c k > ɓ ɗ ʄ ɠ / _V
l > 0 / C_V

[LABIAL] > w / _C
wC > Cw

% p t c k s ç x > h / _C
% dunno if this would apply to anything; shouldn't trigger before semivowels anyway

b d j g > p t c k / V_C or V_#
b d j g > pʰ tʰ cʰ kʰ

V > 0 / _V

θ > ɬ

close LEXICON as 'old-hlu.txt'

Old Hlu to Gejaehl Hlu:

More: show

mode intelligent
open 'old-hlu.txt' as LEXICON

C = p t c k ʔ b d j g ɬ y s ç x h m n ɲ ŋ l r w b̃ d̃ j̃ g̃
V = æ a ɔ e ɤ o i ɯ u
P = p t c k ʔ b d j g

ä å > æ u
ay > æ

r > 0 / P_
r > 0 / _#
r > y
l > 0 / C_

a > ɔ
uy oy > wi we

close LEXICON as 'gejaehl-hlu.txt'

Proto-Vengic to Proto-Whatic: (well-commented because batshit)

More: show

mode intelligent
open 'proto-vengic.txt' as LEXICON

reserve aa ee oo ii uu aɴ aaɴ eɴ eeɴ oɴ ooɴ ɪɴ ɪɪ ɪɪɴ ʊɴ ʊʊ ʊʊɴ iɴ iiɴ uɴ uuɴ

C = p t c k q b d j g v z y ɣ f s ç x h m n ɳ ŋ l r w
V = a e o ɪ ʊ i u aa ee oo ɪɪ ʊʊ ii uu aɴ eɴ oɴ ɪɴ ʊɴ iɴ uɴ aaɴ eeɴ ooɴ ɪɪɴ ʊʊɴ iiɴ uuɴ
A = a aɴ aa aaɴ
E = e eɴ ee eeɴ
O = o oɴ oo ooɴ
[IH] = ɪ ɪɴ ɪɪ ɪɪɴ
[UH] = ʊ ʊɴ ʊʊ ʊʊɴ
I = i iɴ ii iiɴ
U = u uɴ uu uuɴ

[VSHORT] = a aɴ e eɴ o oɴ ɪ ɪɴ ʊ ʊɴ i iɴ u uɴ
[VLONG] = aa aaɴ ee eeɴ oo ooɴ ɪɪ ɪɪɴ ʊʊ ʊʊɴ ii iiɴ uu uuɴ
[VORAL] = a aa e ee o oo ɪ ɪɪ ʊ ʊʊ i ii u uu
[VNASAL] = aɴ aaɴ eɴ eeɴ oɴ ooɴ ɪɴ ɪɪɴ ʊɴ ʊʊɴ iɴ iiɴ uɴ uuɴ

[PB] = p t c k q b d j g
P = p t c k q
B = b d j g
[SZ] = v z y ɣ f s ç x h
S = f s ç x h

% normalize ASCII
[VORAL]m > [VNASAL] / _C or _#
ng > ŋ
gh > ɣ


% I wrote a bunch of rules expecting strict CV structure so let's add a fake consonant to word-initial vowels
0 > ∅ / #_V
C = C ∅

% -- Whatic length rule --

% Insert a fake consonant where it shouldn't apply; easier than defining this all in the condition...

0 > ʜ / #CV_[PB]V % not before plosives
0 > ʜ / #CV_SV % not before voiceless fricatives
0 > ʜ / #CV_C[VLONG] % not before long vowels

[VSHORT] > [VLONG] / #C_CV#
[VSHORT] > [VLONG] / #C_#

ʜ > 0

% -- Intervocalic h loss and hiatus resolution before short vowels --

% First, nasality spreads backwards through /h/... and /y/, why not
[VORAL] > [VNASAL] / _{h y}[VNASAL]

[VSHORT]h > [VLONG]ʜ / _[VSHORT]
ʜV > 0

% -- Debuccalization --

v > h
[VORAL] > [VNASAL] / ɳ_
ɳ > h
s > h / _ not _(CV)+h % I hope this works!

% -- Rebuccalization --

h > w / _{O U}
h > y / _{A E I}

x > ç / I_

% -- Umlaut --

ee eeɴ oo ooɴ > ɪɪ ɪɪɴ ʊʊ ʊʊɴ / _ not AC_
i iɴ u uɴ > ɪ ɪɴ ʊ ʊɴ / _ not AC_ not _{y w}
aya awa > ee oo
e > a / _ not _C{I U}
ee > aa / _{k g x ɣ ŋ q}

% -- Consonant stuff --

C = C ɹ l̃ ɲ mʷ

z > ɹ
n > l̃
y w > ɲ mʷ / [VNASAL]_
c > s
ɲ > n / _ not I_

b d r j g > m n n ɲ ŋ / [VNASAL]_

k g x ɣ ŋ > c j ç y ɲ / [VFRONT]_

% -- Micronesianization --

[LABIAL] = p pⁿ b bⁿ p̃ b̃ f m̥ m
[LABIALVELAR] = pʷ pʷⁿ bʷ bʷⁿ p̃ʷ b̃ʷ fʷ m̥ʷ mʷ
[VELAR] = k kⁿ g gⁿ k̃ g̃ x ŋ̥ ŋ % needed later. gh no longer patterns as a velar - maybe it backed
[HIGHROUND] = ʊ ʊɴ ʊʊ ʊʊɴ u uɴ uu uuɴ
C = C pʷ bʷ fʷ

[LABIAL] > [LABIALVELAR] / _[HIGHROUND]

% -- Plosive nasalization --

C = C m̥ m̥ʷ n̥ ɲ̥ ŋ̥ ŋ pⁿ pʷⁿ bⁿ bʷⁿ tⁿ dⁿ cⁿ jⁿ kⁿ gⁿ p̃ p̃ʷ b̃ b̃ʷ t̃ d̃ c̃ j̃ k̃ g̃

[PB] = p pʷ b bʷ t d c j k g
P = p pʷ t c k
B = b bʷ d j g
N = m̥ m̥ʷ m mʷ n̥ n ɲ̥ ɲ ŋ̥ ŋ
[BM] = pⁿ pʷⁿ bⁿ bʷⁿ tⁿ dⁿ cⁿ jⁿ kⁿ gⁿ
[MB] = p̃ p̃ʷ b̃ b̃ʷ t̃ d̃ c̃ j̃ k̃ g̃

% Stops nasalize between nasal vowels. /l/ also does this
[PB] > N / [VNASAL]_[VNASAL]
l > l̃ / [VNASAL]_[VNASAL]

% Remaining stops become prestopped nasals before a nasal vowel
[PB] > [BM] / _[VNASAL]
% ...and prenasalized stops after a nasal vowel
[PB] > [MB] / [VNASAL]_

% -- Vowel reduction --

% Block vowel reduction in monosyllables!
V > V∅ / #C_#

% Final short high vowels are lost
[LABIAL] > [LABIALVELAR] / _[SHORTHIGHROUND]#
ɪ ɪɴ ʊ ʊɴ i iɴ u uɴ > 0 / _#

% Final short vowels are lost after voiceless consonants, with compensatory lengthening in disyllables
[CVLS] = p pʷ t c k q f fʷ s ç x
[VSHORT] > [VLONG] / #C_[CVLS]{A E O}#
[VSHORT] > 0 / [CVLS]_#

% Final short vowels are lost if the preceding syllable has a long vowel
[VSHORT] > 0 / [VLONG]C_#

% Nasality and length are lost in prefinal syllables
[VNASAL] > [VORAL] / _ not _C?#
[VLONG] > [VSHORT] / _ not _C?#

% -- The Wherestern Whatic consonant shift --
y w > ɲ mʷ / _ not _# not _{[IH] [UH] I U}C?#
j > y

% -- The Wherestern Whatic vowel shift --

reserve ɨɴ æɴ ææ ææɴ

ee eeɴ oo ooɴ > ii iiɴ uu uuɴ % merger as ie uo; i u also > ie uo (short and long)
e eɴ > ɨ ɨɴ
a aɴ > æ æɴ

[AE] = æ æɴ ææ ææɴ
I > i̯[AE]
U > u̯O
[IH] > I
[UH] > U

i̯ > 0 / [LABIALVELAR]_ or {c j cⁿ jⁿ y ç ɲ̥ ɲ c̃ j̃}_
u̯ > 0 / [LABIALVELAR]_
[VELAR]u̯ > [LABIALVELAR] % u onglide turns velars into labial-velars
i̯ > 0 / _ not _VC*#
i̯ u̯ > y 0
yy > y

% Normalize output
∅ > 0

close LEXICON as 'proto-whatic.txt'

Proto-Whatic to What:

More: show

mode intelligent
open 'proto-whatic.txt' as LEXICON

% OK, reserve doesn't do what I want, so we have to normalize all this shit into something that Haedus won't break on.

[VOS] = æ a e o i ɨ u
[VNS] = æ̃ ã ẽ õ ĩ ɨ̃ ũ
[VOL] = æː aː eː oː iː ɨː uː
[VNL] = æ̃ː ãː ẽː õː ĩː ɨ̃ː ũː

[VOSB] = æ̤ a̤ e̤ o̤ i̤ ɨ̤ ṳ
[VNSB] = æ̤̃ ã̤ ẽ̤ õ̤ ĩ̤ ɨ̤̃ ṳ̃
[VOLB] = æ̤ː a̤ː e̤ː o̤ː i̤ː ɨ̤ː ṳː
[VNLB] = æ̤̃ː ã̤ː ẽ̤ː õ̤ː ĩ̤ː ɨ̤̃ː ṳ̃ː

ææ > æː
aa > aː
ee > eː
oo > oː
ii > iː
uu > uː
[VOL]ɴ > [VNL]
[VOS]ɴ > [VNS]

[VMODAL] = [VOS] [VNS] [VOL] [VNL] ə
[VBREATHY] = [VOSB] [VNSB] [VOLB] [VNLB] ə̤
V = [VMODAL] [VBREATHY]

[AE] = æ æ̃ æː æ̃ː æ̤ æ̤̃ æ̤ː æ̤̃ː
A = a ã aː ãː a̤ ã̤ a̤ː ã̤ː
E = e ẽ eː ẽː e̤ ẽ̤ e̤ː ẽ̤ː
O = o õ oː õː o̤ õ̤ o̤ː õ̤ː
I = i ĩ iː ĩː i̤ ĩ̤ i̤ː ĩ̤ː
[IH] = ɨ ɨ̃ ɨː ɨ̃ː ɨ̤ ɨ̤̃ ɨ̤ː ɨ̤̃ː
U = u ũ uː ũː ṳ ṳ̃ ṳː ṳ̃ː

[PB] = p pʷ t c k b bʷ d j g
[MP] = p̃ p̃ʷ t̃ c̃ k̃ b̃ b̃ʷ d̃ j̃ g̃
[PM] = pⁿ pʷⁿ tⁿ cⁿ kⁿ bⁿ bʷⁿ dⁿ jⁿ gⁿ
P = [PB] [MP] [PM]

[NH] = m̥ m̥ʷ n̥ ɲ̥ ŋ̥
[NV] = m mʷ n ɲ ŋ
N = [NH] [NV] l̃

F = f fʷ s ç x
R = l r ɹ y ɣ w

C = P N F R q h ɦ

% Prestopped nasals become glottalized nasals (if voiceless) or 'low' nasals (if voiced), but there's no need to represent this.

% -- A few changes --

[MP] > N

[LABIAL] = p pⁿ b bⁿ p̃ b̃ f m̥ m
[LABIALVELAR] = pʷ pʷⁿ bʷ bʷⁿ p̃ʷ b̃ʷ fʷ m̥ʷ mʷ

æyæ > eː
[SHORTAE] = æ æ̃
[SHORTAE]y[SHORTAE] > ẽː
A[LABIALVELAR] > O[LABIAL] / _#
[IH][LABIALVELAR] > U[LABIAL] / _#

ɣ > ɦ

% -- Prefinal vowel reduction --

V > ə / _ not _C?#

% -- Breathy voice conditioning --
% Preinitials in the B- BM- series condition breathy voice on the whole word and merge with P- M-. Also, ɣ > ɦ, which also conditions.
% Fuck it, let's say voiceless nasals do the same thing.
% Later on, words with breathy voice end up with it only on the main vowel. Or whatever. That's where it's written anyway.

[VMODAL]ɦ > [VBREATHY] / _#
əɦ[VMODAL] > [VBREATHY] / _C?#

[BREATHCONPRE] = b bʷ d j g bⁿ bʷⁿ dⁿ jⁿ gⁿ m̥ m̥ʷ n̥ ɲ̥ ŋ̥ ç fʷ ɦ
[BRTHMERGEPRE] = p pʷ t c k m mʷ n ɲ ŋ m mʷ n ɲ ŋ y w ɦ

[BREATHCONPRE]ə > [BRTHMERGEPRE]ə̤
[VMODAL] > [VBREATHY] / ə̤C_
[VMODAL] > [VBREATHY] / ə̤C_
[VMODAL] > [VBREATHY] / ə̤C_
[VMODAL] > [VBREATHY] / ə̤C_
[VMODAL] > [VBREATHY] / ə̤C_
[VMODAL] > [VBREATHY] / ə̤C_
[VMODAL] > [VBREATHY] / ə̤C_
[VMODAL] > [VBREATHY] / ə̤C_
[VMODAL] > [VBREATHY] / ə̤C_
[VMODAL] > [VBREATHY] / ə̤C_
[VMODAL] > [VBREATHY] / ə̤C_ % should be enough applications

ɦ > 0
ə̤ > ə
ə > 0 / _V or #_

x > h
[VLSR] = m̥ m̥ʷ n̥ ɲ̥ ŋ̥ ç
[VCDR] = hm hmʷ hn hɲ hŋ hy
hə[VCDR] > [VLSR]

% -- The Great What Vowel Shift --

[SHORTAE] = æ æ̃ æ̤ æ̤̃
[SHORTA] = a ã a̤ ã̤
[SHORTE] = e ẽ e̤ ẽ̤
[SHORTO] = o õ o̤ õ̤
[SHORTI] = i ĩ i̤ ĩ̤
[SHORTIH] = ɨ ɨ̃ ɨ̤ ɨ̤̃
[SHORTU] = u ũ ṳ ṳ̃
[LONGAE] = æː æ̃ː æ̤ː æ̤̃ː
[LONGA] = aː ãː a̤ː ã̤ː
[LONGE] = eː ẽː e̤ː ẽ̤ː
[LONGO] = oː õː o̤ː õ̤ː
[LONGI] = iː ĩː i̤ː ĩ̤ː
[LONGU] = uː ũː ṳː ṳ̃ː

[PALATAL] = c cⁿ j jⁿ c̃ j̃ ç ɲ̥ ɲ
[VELAR] = k kⁿ g gⁿ k̃ g̃ x ŋ̥ ŋ
[LABIAL] = p pⁿ b bⁿ p̃ b̃ f m̥ m
[LABIALVELAR] = pʷ pʷⁿ bʷ bʷⁿ p̃ʷ b̃ʷ fʷ m̥ʷ mʷ

[LONGAE] > [LONGE]

[SHORTI] > [SHORTIH] / _ not [PALATAL]_ not [LABIAL]_ not y_ not _y
[LABIAL][SHORTU] > [LABIALVELAR][SHORTIH]
[VELAR][SHORTU] > [LABIALVELAR][SHORTIH]
[SHORTU] > [SHORTIH]w / _# not [LABIALVELAR]_ not [PALATAL]_
[SHORTU] > [SHORTIH] / _ not _[VELAR] not _[LABIALVELAR]

[LABIALVELAR] > [LABIAL] / _#

[LONGI] > [SHORTI]
[LONGU] > [SHORTU]

[LONGE] > [SHORTI]y
[LONGO] > [SHORTU]w

[SHORTAE] > [SHORTA] / [VELAR]_{q ɹ}
[SHORTAE] > y[SHORTAE] / [VELAR]_ or _{m n}
[SHORTAE] > [SHORTA]

[LABIAL] [VELAR] > [LABIALVELAR] / _[SHORTO]
[SHORTO] > [SHORTA]w / _[VELAR] or _#

[LONGA] > [SHORTA]y / [VELAR]_#
[LONGA] > [SHORTE]y / _#
[LONGA] > [SHORTE]

% -- Consonant simplification --

% Glottalized nasals develop from pre-stress prestopped nasals and P-N sequences

[MM] = m mʷ n ɲ ŋ
[GM] = mʼ mʷʼ nʼ ɲʼ ŋʼ % glottalized
C = C [GM] ɗ

bⁿ bʷⁿ dⁿ jⁿ gⁿ > m mʷ n ɲ ŋ
pⁿ pʷⁿ tⁿ cⁿ kⁿ > mʼ mʷʼ nʼ ɲʼ ŋʼ

ə > 0 / əC_

[PB][MM] > [GM] / ə_

% Beginning of labial-velar collapse
fʷ > ʍ

% Labial-velars merge with labials except in stressed onsets
[LABIALVELAR] > [LABIAL] / _ not _{y w}?VC*#

% Now collapse the rest of them
p̃ʷ b̃ʷ > m̥ʷ mʷ
mʷ mʷʼ m̥ʷ > m mʼ m̥
pʷ bʷ > ƥ ɓ

% Lateral restructuring
l > ɗ / _VC+#

l̃ > l / _əF or _əR or _ə[VELAR] or _əq or _əh
l̃ > m / _ə[LABIAL] or _ə[LABIALVELAR]
l̃ > l / _VC+#
l̃ > n

ɹ > r / _#
ɹ > l

% h-epenthesis to properly generate underlying forms
V > Vh / _#

% Orthography
q > ʔ



close LEXICON as 'what.txt'

[Kuchigakatai]

Huh, so Morrigan's Haedus SCA allows comments. And now I'm wondering why most other SCAs don't, that's definitely a welcome feature.

[akam chinjir]

It also allows writing intermediate forms to file. (I've mostly been playing with Phonix, and that's a feature I wish it had.)

[KathTheDragon]

Huh, so Morrigan's Haedus SCA allows comments. And now I'm wondering why most other SCAs don't, that's definitely a welcome feature.

Indeed, it's not even especially complex to support them. They're even really popular, as I learnt when I asked for adding them to my SCA.

[Nortaneous]

Messed with the SCs a little to provide RUKI-style variety in details in an early Ziwanic-Hluic-Whatic isogloss - namely, I've decided (from going through some very old notes on Vengic) that *B were actually prenasalized stops, and *V were plain voiced stops. These values are preserved in Zotic but nowhere else.

However, there are details!
- In Whatic, nothing of particular interest happens.
- In Vengic, *t *z were both dental, so *z gives Proto-Ziwanic *ð. In V'eng, this ends up merging into /r/. (I haven't implemented the Zzyxwqnp sound changes yet.)
- In Hluic, this was preceded by a collapse of the voicing distinction intervocalically - but *t was dental, whereas *z was alveolar, so *-t- gives Proto-Hluic *ð, whereas *-z- gives Proto-Hluic *z.

It is possible that, in Proto-Vengic proper, *t *z were /t̪ d/ respectively, like in Proto-Austronesian or whichever branch thereof. But I'm not going to do the *t > k thing; that's just silly.

Anyway, cognates. Zzyxwqnp (worked out by hand from the V'eng forms) - Gejaehl Hlu - Bor Hlu - What. Some of these will probably be replaced.

Proto-Ziwanic	Proto-Hluic	Proto-Whatic	Zzyxwqnp	Hlu (Gejaehl)	Hlu (Bor)	What
ʔaβéŋi	ʔwæɲ	ʔeɲeŋ	vùŋ	ʔwɤɲ	ʔwɤɲ	ʔɲaŋ
miká	meɣ	mʲeke:	gā	mɯj	mɯj	mkij
hutéŋa	hðeŋ	mʷoteŋe	tsə̄ŋ	hɯɲ	hɯɲ	mnˀiah
nakása	ŋgaθ	n̆eke:	gâ	ŋgɒɬ	ŋgaɬ	lkij
fuké	huɣ	fʷuke:	tsjə̂	hu	hɒw	wkij
cĩ	sæ̃	sẽ:	tɕīŋ	se	si	sĩj
cuʔáca	ʔaɟ	soʔes	tɕjə̄	ʔɒc	ʔac	sʔas
mirúja	mrɤj	mʲiroɲe	ɖì	jwɤj	rwɤj	mrɲah
majá	maj	mʲɨe	mbjā	mæ	maj	mah
tóŋa	tog	toge	twə̄ŋ	ɗok	ɗok	tgiah
téŋa	teŋ	ti:ŋ	tsə̄ŋ	ɗɯɲ	ɗɯɲ	tiŋ
tã́ŋa	tãg	tẽŋe	tɕjāŋ	ɗɒk	ɗak	nˀiah
péra	per	pʲi:r	pjə̂	ɓi	ɓir	pir
yã	yã	ɲãː	jāŋ	jɒ	ja	ɲẽj
kusá	kɤθ	pʷoɲe:	sà	ɠɤɬ	ɠɤɬ	pɲij
maxúra	ŋgur	mefʷore	gwû	ŋgwi	ŋgwir	mwra̤h
pikĩ́ra	wcʌ̃r	pʲĩkĩne	tɕjî	ʄwɤ	ʄwɤr	mˀŋna̤h
rakáru	rɣɤr	reker	kwə̂	jɯ	rɯr	rkiar
tirṹxu	trɔ̃h	terũx	ʈû	tɒh	tɒh	trɨ̃h
túba	tɤb	ʈū	tu:bʲe	ɗɤp	ɗɤp	tbah
jinápa	jnaw	ɲen̆epʲ	ȵdʑjá	jno	jnaw	ɲlap
ðipúmʲa	jwɤr	ɹupʷoɲẽ	pjíŋ	jwɤ	jwɤr	lmˀãh
ɖéβa	rãw	jĩ:mʷ	ʈwù	jo	raw	jĩm
kímu	cim	kæ:mʷ	tɕjīŋ	ʄim	ʄim	kijm

Orthography:

More: show

wvnp - qwoeny - qwoeny - xyñagn
ga - muey - muey - mykî
czn - hueny - hueny - mynxia
gax - gahl - gahl - lykî
chet - hu - haw - ukî
chyn - se - si - sîn
che - qac - qac - syxas
dryp - ywoey - rwoey - maña
mr - mae - may - ma
ton - ddok - ddok - tygia
tszn - ddueny - ddueny - tign
tan - ddak - ddak - nxia
pex - bbi - bbir - più
rn - ya - ya - ñên
sap - ggoehl - ggoehl - pyñî
gux - gwi - gwir - muäh
kip - jjwoe - jjwoer - mxyngnah
kox - yue - ruer - akiau
trvx - tah - taoh - tyrwn
trv - ddoep - ddoep - tyba
nyrt - yno - ynaw - ñylap
pint - ywoe - ywoer - lymxan
trup - yo - raw - yim
chin - jjim - jjim - kîm

[Nortaneous]

toqqá mo bzéagum ɳɖą ŋgigórrøm yea ccoaχojóaqq steŋŋ
to-qqa mo b-zeagum ɳ-ɖą ŋ-ki=korr-øm=yea ccoaχojoaqq s=tteŋŋ
AN.T-heaven REAL CONT-bear.MAJ INAN-thing INAN-NEG=count-NMLZ=DAT nourish.MAJ POSS=mankind

tteŋŋ mo kʉnéarr ŋgiðą́ hø̨n kiɭo kimǿduŋ toqqá
0-tteŋŋ mo ki=b-nearr ŋ-ki-dã hø̨n ki=ɭo ki-møduŋ to-qqa
AN.H-mankind REAL NEG-CONT-take INAN-NEG-thing INS RECIP pay AN.T-heaven

dull dull dull dull dull dull dull
dull
kill

[Nortaneous]

introducing

THE PAN-SWAMPISH ALPHABET

Hlu:
Ykvku yxllx ccv ttwm jjqngnill yvhhngq lliddxſx i ttfiyill
Ykvku ttwſxhhyi i ddvñ kwñ yKqtnqll myqlltwn yemqk yllvyttwñ
Kvntw gqlle fqng llitwſx yo gqs yDw nnqmi vbqnſx
ggfqk gqs ddvñ gqs bodicilq cxddw ttwñ qybolli lillq
nfqyq lillq kokkq ykvt mmi llyomllq yngfe yyi yqc hhfiyoyi
hvñ i Kqtnqll ggfqkmmun fvñq hvñ i Cillx ddvñmmun fvñq

What:
khvn nlrlif ñle ſmlvf ñqſqſvqnhvfh v mnqx
ñfvd n yñkiy gyhgx mnqx ſmlvf ppuftñfvfnhvfh v khv
dufl dufl dufl dufl dufl dufl dufl

Qoa:
tokkhaj mo bzxjgum ndranh ñgigojrrvm yx ccqhodcqjkkh ſteññ
tteññ mo kwnxjrr ñgidhajn hvnnn kilro kimvjduñ tokkhaj
dull dull dull dull dull dull dull

Kangſhi:
1. Vrot ñzvbzaj ſkae xri nax
2. Vrot ñgbodut bñdau bvñbdoxſiej
3. Vrot ñxſaex n imaeit ſiabdv nix u ññwit itog nix

diſcuſs

[Nortaneous]

It will at some point be necessary to describe Rau, since it is a major language of the area north and west of the Zhjumna, and the liturgical language of the local variety of Institutional Tengrism.

/p b~v t d~ð~ɾ z̺~dz̺~ʐ~ɖʐ~ɹ k g~j~ʝ~ɣ q ɢ~ʁ~ʕ ʔ/ <p b t d z k g q r 0>
/ɬ s̻ ɕ x~h~f ꭓ/ <l s sh h kh>
/m n ŋ/ <m n ng>
/a e ʌ o i (u)/ <a e u o i ou>
/a: e: ʌ: o: i: u:/ <â ê û ô î oû>

ʇ ᵑʇ ᶢʇ <c nc gc>
ǁ ᵑǁ ᶢǁ <x nx gx>
ǂ ᵑǂ ᶢǂ <ç nç gç>

/u u:/ are rare. Historically, /u:/ became /u/ in a few environments, otherwise generally /ib/ outside the influence of uvulars. (Non-central long vowels diphthongized in general - so e: o: i: u: > eg ob ig ub. See below. They were later reintroduced from loans.) Also, note that s z sh all have different places of articulation.

There is no contrast between voiced plosives, voiced fricatives, and sonorants; furthermore, /g/ is often fronted to [ɟ], and admits the allophone [j]. The voiced series can even be syllabic, as in bsoukhashd /b.su.ꭓa.ʃd/. Syllabic voiced plosives are usually pronounced as sonorants preceded or followed by an extra-short schwa. Initial clusters can only appear in stressed syllables, and have certain phonotactic restrictions that have yet to be defined.

Coda consonants are generally phonetically mangled, with effects on the preceding vowel; and the inventory of finals is restricted. For example, consider the stressed finals with voiced plosives:
/ab eb-ʌb ob ib-ub/ [ɔo̯ ɛw ʌw ɪw~ʏw~ʊw]
/ad ed ʌd od ið uð/ [æð eð ʌð øð ið yð]
/az ʌz oz ez-iz-uz/ [æɻ ɑɻ oɻ iɻ]
/ag eg ʌg og ig-ug/ [æe̯ ej ʌj oj~uj ɪj]
/aɢ eɢ-ʌɢ oɢ iɢ-uɢ/ [ɔˤ ʊˤ oˤ uˤ]
Voiced plosives and nasals can be followed by another consonant, although /ŋ/ affects non-front vowels similarly to /ɢ/:
zdogs [s̺ɾujs]
bdang [bɾɔˤŋ]~[bɾɔ̃ˤ]
nxeml [ᵑǁɛmɬ]~[ᵑǁɛw̃ɬ]
zart [dz̺ɔˤt]

Voiced plosives and clusters cannot occur after long vowels.

Some Kangshi loans:
irsda /iɢ.sda/ [uˤs̺ɾa] "Zhjumna" < igzda "highland"
ngzumhib /ŋ.zum.xib/ [əŋ.ndzʏ̃.fʏw] "fox" < ñzwmxox "fox"
bshtad /b.ɕtad/ [əv.ɕtæð] "market" < bstct "market"

The lexical situation of Northern Continental Hathic merits comment. There is relatively little lexical inheritance from Proto-Hathic, with many words loaned from an unknown substrate and some seemingly from para-Amqolic, and a predominantly religious or otherwise high-cultured register of loans from Kangshuic. The argument for the connection of Northern Continental Hathic to the remainder of the Hathic family rests primarily on the resemblance of the determiners, the inheritance (albeit often as elements of compounds) of a small amount of core vocabulary, and the preservation of what little morphology Proto-Hathic had, such as the formation of comparative adjectives.

In addition, the Proto-Hathic people had a tradition of lexically conservative epic poetry, known generally as r!ər ɕogoɭɨg c!aam "celestial and chthonic recitations", handed down with concomitant glosses; and Hathic languages, including Northern Continental Hathic, tend to contain an elevated stratum of loans from the Recitations, informed by the glosses. These loans are often reinforced by compounding with common-register nouns, but it is not uncommon for the glosses to be wrong, or the assumed correspondences between the glosses and the words of the Recitations to be incorrect, with the result that this lexical stratum often shows semantic shifts that would be generally considered implausible.

For example, the name of the Liminal Chariot by which the Forgers of the Æon descended from the Celestial Sphere, fʷʌʌ ŋɛr, has, in the earliest texts, an alternant fʷʌʌŋ ci, which appeared for complex metrical reasons in the set phrase fʷʌʌŋ ci nt!a rɛs "shining fʷʌʌ ŋɛr". A later gloss, which must have been composed after a number of phonological mergers, assumed that fʷʌʌŋ was related to hoŋ "moon", and that fʷʌʌŋ ci nt!a rɛs referred to the "shining skin of the moon", thus inventing a word ci "skin, exoskeleton" which was compounded with substratal *fʌl "skin" to produce Rau shtibor "skin", and compounded with Amqolic *besúg-xan "mass of critters" (preserved in Amqoli bzugxal) to produce bsoukhashd "chitin".

Complicating matters further, the conversion of the Continental and Archipelagic Hathics to various denominations of Institutional Tengrism has resulted in the loss of the tradition of the Recitations; the main language in which they are preserved is Hathe, which takes the tendency of consonantal simplification present in non-Archipelagic Hathic to an extreme, leading to exceptional difficulty of reconstruction, and which has on occasion versified its glosses, elevated them to the status of Celestial Recitation, and then developed to such an extent that it was deemed necessary to produce glosses on those. However, the Hathic range overlapped to some extent with that of the Vengic pagans, and some of the Recitations were taken down by the Narng. Also available to future Enzean philologists is a very confused but extensive Ziwan codex, containing a few Celestial Recitations that had been taken up and hopelessly garbled by a Hathic-influenced pagan sect, as well as what is presumably a mechanical translation of a lost Hlu epic consisting of a collection and re-versification of a variety of glosses, and numerous written recordings (which agree to a surprising extent) of the "Keys of Heaven", the most incomprehensible, and most sacred to the early Hathics, of the Recitations.

The content of the Celestial Recitations, as will be discovered much later by the aforementioned Enzean philologists, generally relates to the nature of magic and of the gods; the creation of the world by the gods out of metals and æthers at the beginning of the historical era; sprawling, psychohistorical ramblings about prehistory dating back to the beginning of the previous historical era, in which mankind inhabited the celestial sphere; and garbled mathematical and alchemical musings promising mankind the power to liquefy gold, create stars, ascend bodily to higher spheres, and suchlike. The most important recitation was a wildly inaccurate star chart followed by several hours of uninterpretable, formulaic, vaguely numerological babble which the earliest glosses held to be somehow the language of the celestial sphere itself, and to which structural similarities were later found in the occult incantations of certain rural Kett populations on the other side of the world.

Several millennia later, Hierarch Renxau of Enze would quietly relocate the philology departments of Enze's top universities to the imperial capital.

Anyway! Rau, being a fairly average Continental Hathic language, shares with Kangshuic and the general Zhjumna linguistic area a generally but not entirely prefixing morphology, pervasive use of noun classes, and a wide array of auxiliary verbs. In addition, it has a large class of 'bipartite verbs' in which object and past/non-past markers are infixed into the verb (diachronically, these are descended from unitized V-N compounds with markers of possession serving as person markers), Amqolic-like verbalizers of location, and a distinction between absolute, conjunct, pronominal, and auxiliary forms of verbs.

This distinction merits some comment. Proto-Rau did not have "words"; it instead had stress groups. Unfortunately, this post is too long, so the aforementioned point at which Rau will have to be described is not yet.

[Nortaneous]

Rau is a stress-timed language with vowel reduction, so it's difficult to know where even to begin. How about with the system of juncture?

So! There are five degrees of juncture in Rau:

- Open juncture, or the absence of juncture
- Semi-close juncture, in which vowel length is lost on secondary stress peaks
- Old close juncture, in which vowel length is lost on secondary stress peaks, unvoiced consonants at the juncture are voiced, and certain alternations occur
- New close juncture, in which vowel length is lost on secondary stress peaks and unvoiced consonants at the juncture are voiced
- Clitic juncture, in which one element of the juncture, the clitic, is made unstressed and its non-medial unmarked consonants are voiced, and resyllabification occurs

On top of this, each junctural structure has a rule for stress peak placement. What is stress? Well, every basic morph in Rau - including compounds that have become opaque - has either one or zero primary stress peaks. For example:
b̲soú:khashd̲ 'chitin'
b̲shtád 'market'
g̲ '1SG.M'

Basic morphs of more than three syllables, which are rare, have predictable secondary stress peaks on every other syllable from the primary stress, unless that syllable contains an unstressed schwa.

Every bound group must have at least one stress peak.

A word on orthographic conventions:
- Bound groups are written separated with vertical bars.
- Syllabic consonants are written with an underline, and long vowels are written with colons.
- Open junctures are written with a space.
- Semi-close junctures are written with a tilde.
- Old close junctures are written with one hyphen.
- New close junctures are written with two hyphens.
- Clitic junctures are written with =.
- Because the velar nasal can be syllabic, it will be written <ñ>.
- The lateral fricative will be written <lh>, and /ɢ/ resulting from the old voiced lateral fricative will be written <ll>, because I keep getting confused when I try to do it otherwise.

An example of open juncture is that between a non-lexicalized sequence of a noun and an adjective:
césh pánd "red cat"

An example of semi-close juncture is that between the two elements of an ad-hoc compound:
pánd~çigçì "cat bell" (cf. çigçí: "bell")

An example of old close juncture is that between a verb and an incorporated object:
lhòu-bánd "...sees the cat"

An example of new close juncture is that between a verb and an incorporated subject: (however, in this case the verb also becomes completely unstressed)
pànd--lle "the cat hears..."

Old close juncture takes precedence over new close juncture: (also note the stress shift and [analogical] vowel alternation)
tàgm̲--lle-bánd "the dog hears the cat"

The only case in which clitic juncture is used is between a verb and a pronominal clitic:
g̲-lhá "I hear..."

The first thing to note is that none of this has anything to do with either voicing or length. So! Historical phonology.

Rau collapsed its inherited voicing contrast. Voiced consonants mutated preceding stressed vowels: (voiced / unvoiced / unstressed)
(TODO)

There is one exception: voiced lateral affricates became <r>, whereas voiceless ones became <l> - hence the variation between lhou and lle. (Really the variation is between lhib and lle, since /u:/ became /iw/* in most cases -- cf. lhib-irsda "...hears Zhjumna". This is a little contrived, but "see" is not a basic verb. In fact, it is not just any phrasal verb, but an inverse possessive phrasal verb, meaning that, when it takes person indices, its object marking is where its subject marking should be, and it takes a completely different set of subject markers, which are both obligatory and infixed. A little too advanced for the Caecilius-est-in-horto stage.)

Old close juncture is simply the close juncture old enough for an earlier phase of junctural voicing to have applied before the collapse and redevelopment of the voicing contrast.

So! How many ways are there to say "the dog hears the cat"? It depends:

The unmarked form is tàgm̲--lle-bánd.

What happens if you want to emphasize the dog? Well, several things.
- First, tagm must not be incorporated; incorporated objects are background information.
- Since tagm is definite and not incorporated, it needs an article - tagm and pand both take the article for animals, n=.
- Also since tagm is not incorporated, it no longer forms a bound group. This is fine, because it is not a pronoun, meaning that it can form its own bound group.
- Since it is not incorporated, the bound group it forms may appear either before or after the verb.

Word order is somewhat free, but these two options have different pragmatic meanings: placing it before the verb establishes it as a new topic ("the dog heard the cat, and then..."), whereas placing it after the verb draws attention to the agency of the dog. ("The dog heard the cat!") These also have different grammatical implications: in the first case (the introductory construction), the dog must be indexed on the verb, but can then be indexed as a subject by the sequential marker, whereas in the second case (the adjunct construction), the dog needs explicit case marking. Also, for definite newly-introduced topics, the article is joined by a semi-close juncture with loss of stress on the article; otherwise, the article is usually joined by clitic juncture.

Thankfully, there is one simplification: non-human third-person indexal clitics are not distinct from cliticized articles. This is because they are from demonstratives and demonstratives also served as pronouns. So:

Introductory construction: ne~tágm̲ | n̲=lhoù-bánd DEF(ANIMAL)~dog | 3SG.A(ANIMAL)=hear-cat.P
Adjunct construction: lhoù-bánd | n̲=tágm̲=d̲ hear-cat.P | DEF(ANIMAL)=dog=NOM

([neˈtæjᵊm ᵊnɬʏwˈand] and [ɬʏwˈand ᵊnˈtæjᵊmᵊð]. Remember resyllabification? Voiced plosives and nasals are preferentially syllabified as codas. Also, secondary stress just means you get the stressed form instead of the unstressed form.)

If you want to emphasize the cat, most of the same applies, although of course you'd use the genitive; and the genitive, not being the nominative, must be marked even when it occurs before the verb. Also, lhe, being vowel-final, will resyllabify the clitic =n. So:

Introductory construction: ne~pánd=tn̲ | tágm̲--lle=n DEF(ANIMAL)~cat=GEN | dog.A--hear=3SG.P(ANIMAL)
Adjunct construction: tágm̲--lle | m̲=pánd=tn̲ dog.A--hear | DEF(ANIMAL)=cat=GEN

There is also the option of incorporating neither - perhaps to introduce a folktale featuring the characters Dog and Cat.

* dhok didn't like u: > ib even though Rau "voiced plosives" are clearly fake and actually sonorants >:|

[dhok]

coptic is fake. only fake coptic can actually be real