Andanic languages

The Andanic languages are those descended from Old Andanese. Most are very conservative; however, the Andanese language for which the family is named is not particularly conservative at all, and its speakers numerically outweighed those of all of the other languages combined.

Tapilula (0) to Old Andanese (1900)

The Andanese/Gold dialect of Tapilula had the consonants

Rounded bilabials:                     hʷ  w
Spread bilabials:      p       m   b   f  (Ø)
Alveolars:             t       n   d       l
Rounded alveolars:     tʷ      nʷ  dʷ         
Velars:                k   ḳ   ŋ   ġ   h   g

The accent pattern involved with accented schwa switched to favor the following vowel. e.g. ăpo "field", apə̀ho "field (possessive) became /ăpo apəhò/. This is important to distinguish from Gold, where stress passed to an adjacent vowel but the tone was not carried over.
The "labial" vowel ə disappeared, syllabified nearby consonants or turned to i if the nearby consonants were not possible to become syllabic. Note that it never occurred after labialized consonants. Sequences such as /pəh/ collapsed to form aspirated consonants, though these behaved as clusters.
Note that because the accented schwas had already given up their tones, even when they survived as /i/ they were no longer stressed.
Preexisting syllabic consonants were also analogized to these same rules, meaning that epenthetic /i/ was inserted where necessary. This shift might have one exception: ṁ > um, seen in one word for breast; alternatively, the word was variable even in Tapilula. There is also /mə/ > /mu/ in one word.
IT IS ALWAYS /i/ IN NEWLY GENERATED WORDS.
The velar nasal ŋ changed to n in all positions.
The voiceless bilabial stops p ph shifted to b bh.
The voiced unaspirated bilabial stop b (which included nearly all earlier /p/) became w in any position within a stressed syllable, except when following a syllabic nasal, in which case that nasal shifted to /ṁ/ and the /b/ remained.
For some reason unexplained here, it seems that monosyllabic words always end up with /bh > p/ even reflecting earlier simple /b/ and therefore dodge this shift. The only viable explanation is that /h/ was analogized to the base form from an inflected form like bə̀ha. If this happened, it may have been irregular.

Remember that words are syllabified as CVCV, even after high tones, so a word like /tàba/ would not lose its /b/.
The voiceless alveolar stops t k became strongly aspirated to th kh when occurring as the onset of a stressed syllable. This did not affect the velar ejective /ḳ/. However, the cluster ḳh (always from earlier /ḳəh/ and /həḳ/, mostly in grammatical inflections) did also shift to kh.
The aspirated stop th shifted to kh. This did not affect /dh/. HOWEVER, SINCE D IS NOT SHIFTING TO DH JUST FOR BEING STRESSED, ITS POSSIBLE IT ACTUALLY DID APPLY HERE WHEN THE RARE DH WAS FORMED.
NEWLY GENERATED WORDS ASSUME /dh/ > /kh/.
The labialized consonants tʷ dʷ nʷ shifted to kʷ ġʷ ŋʷ.
The labialized sounds kʷ ġʷ hʷ f w changed to k ġ h h g when they preceded a vowel followed by a labial consonant (including /w/).
ŋʷ> ŋ.
Tautosyllabic vowel sequences òi ài èi converged to ē. This did not affect syllable-straddling words like /tùya/. Likewise, èu àu òu in the same environment converged to ō.
This shift, and the one below, require the language to still distinguish between the rotating /g/ "B" (which could become /Ø/) and the non-rotating /g/ which is always /g/.
Duplicate vowel sequences àa èe ìi òo ùu shifted to long vowels ā ē ī ō ū. But the same sequences with the opposite tone pattern did not shift.
Likewise, sequences of a high tone followed by /g/ or /h/ shifted to long vowels. After a stressed low tone, however, g h shifted to Ø, thus creating a small class of monosyllabic low-tone CV words.
NOTE ON POLITICS: Oyster breaks away here, having no /p/ phoneme but still having /bh/.
The glottalized stop k̩ changed to a uvular q in all positions.
The voiced stops b d ġ ġʷ became voiceless p t k kʷ, but /p t/ remained voiced allophonically between vowels. Note that aspiration was still contrastive, though sequences like /ph < pəh/ behaved as clusters rather than single consonants, reflecting their origin.
Chronically unstressed syllables all became short and low tone. This even extended to shifts like ou > o that had been missed by previous shifts.
In classifiers, aw ew shifted to ow, introducing an /a/ phase into the preexisting /e~o/ alternation. This only happened when the stem began with /w/, but it was analogized sporadically. This new sound was phonetically identical to /ō/ but was spelled distinctly because it overlapped a syllable boundary.
Tones were eliminated in closed syllables, but length was preserved. Clusters like /mh nh ŋh/ (the later /mp nt ŋk/) were front-loaded, and so they did not create closed syllables.

By this time the Old Andanese language had the consonants

Bilabials:         p   m   w
Alveolars:         t   n   l
Velars:            k   ŋ   g   h
Labiovelars:       kʷ          hʷ
Uvulars:           q
Rounded uvulars:   qʷ

and the vowels /a e i o u/. There were two short tones and a single long tone.

The rounded uvular /qʷ/ was much rarer than /kʷ/, but is analyzed as a single phoneme because the language otherwise has no clusters of a consonant followed by /w/.

The labial glide /w/ is often spelled /gʷ/, but there is no phonemic contrast.

Note that the sequences /uy iw/ were preserved, whereas mid vowels paired with a following glide. /iy uw/ can be seen as merely another way to spell /ī ū/, as in many other languages.

An alternate analysis is possible, where every phoneme can either be aspirated or not, and in which case /g~h/ and /w~hʷ/ form pairs. This is unwieldy however because apart from /h hʷ/ the aspirated consonants are much rarer than the plain forms.

North Andanic languages

See Litila for the Paleo-Pabappa language often grouped with Andanic.

Old Andanese (1900) to Galà (3750 AD)

The Old Andanese language had the consonants

Bilabials:         p   m   w   f
Alveolars:         t   n   l
Velars:            k   ŋ   g   h
Labiovelars:       kʷ          
Uvulars:           q
Rounded uvulars:   qʷ

and the vowels /a e i o u/ on two tones.

The uvular stops q qʷ shifted to k kʷ in word-initial position. Most root-initial /q/ also shifted, because most roots could appear without classifiers at least in certain contexts. But roots that were fully bound retained initial /q/.
The sequence qi disappeared to q before a consonant or at the end of a word. The sequences qig qih simply shifted to qi.
The voiceless stops p t shifted to b d in word-medial position. Root-initial examples also shifted if they were padded by classifier prefixes, but not in absolute initial position (that is, they followed the inverse of the pattern of the /q/>/k/ shift).
This happened even after a high tone, and Galà continued to allow voiced geminates even as they became voiceless in neighboring languages.
The clusters qp qm qt qn qk qŋ qkʷ shifted to geminates pp mm tt nn kk ŋŋ kkʷ. Any cluster *ending* in /q/ shifted it to /k/.
Note, if the geminates are assumed to be acoustically identical to high tone + singleton, this shift can be ignored because it is the same as the next one.
All remaining q shifted to Ø and caused the preceding vowel to become high-toned. If a root had had initial /q/ and did not shift it to /k/, this /q/ now disappeared but did not raise the tone of the preceding syllable (since it would be a classifier).
The clusters mh nh ŋh (common in genitives) shifted to mp nt ŋk. Then bh dh shifted to p t.
Remaining aspirate clusters deaspirated.
Therefore, /lh/ > /s/ does not exist in this language.
Unstressed mid vowels e o were raised to i u when adjacent to a stressed vowel in either direction.
Through grammatical alternation, some stressed vowels may also shift. For example, Old Andanese wĕqa "fig" > Galà dă is only possible if it first shifted to /wèa/ through /q/>/`/ and then to /wiă/ through a stress shift and the above sound change.
Before a vowel, the sequences ti hi ki all shifted to s. Likewise in the same environment gi li bi di shifted to d y. pi shifted to t. Any other palatalized consonants then depalatalized.
This may be used in the 2nd person agent marker, which was primordially /gə-/, and could have been analogized to an infix /-əg-/, then to /-eg-/ and /-ig-/, both of which would behave as expected here. There could also be /-og-/, but the /u/ is not well-addressed below.
The labialized consonants kʷ w f shifted to k Ø h.
The voiced velar fricative g disappeared to Ø, and changed any preceding high tone into a long tone.
This contrasted with primordial low tones in word pairs such as kūi "portable wall; divider" vs. kŭi "button". It also contrasted with the much rarer inherited vowel sequences that had arisen from earlier /q/, which still retained a glottal stop as part of the effect of the high tone. Thus kàa "fin", with a glottal stop between the two vowels, contrasted with kăa "safe space; refuge", and with kāa, part of the name of a flower.
After a high tone, the voiceless fricative h shifted to q.
Long vowels became allophonically shortened before a hiatus, thus becoming a short high tone. The three-way distinction remained because of the distinctive glottal stop after the grave tone.

Thus the final consonant inventory was

Bilabials:         p   b   m        
Alveolars:         t   d   n   l   s
Palatals:                      y
Velars:            k       ŋ       h
Uvulars:           q

Udami

Spoken to the west of Galà.

The uvular stop q shifted to k in all positions.
The voiced stops b d shifted to p t.
The voiced velar fricative g shifted to x. This contrasted with the inherited /h/.

Possibly a Palli-like "staircase shift" to get phonemic /s/. But how? Palli's was /fa fi fu sa si su ša ši šu ha hi hu/ > /fa fi fu sa si su sa si fu fa si fu/. This shift could have been affected by front vowels on *both* sides of the consonant. For sure, /h/ is the "weakest" consonant, probably shifting away completely by changing to /s/, /x/, or /f/, depending on the surrounding consonants. Even with all of this, /s/ will still be rare unless /k/ is somehow pulled into the shift.

Possibly /f w/ > /h 0/ later on. Under some circumstances, maybe also /p m/ > /h 0/, but the /m/ nasalizes vowels and thus causes a final -n.

Lyugi

A language that changed more than the others. Spoken in the mountains of Repilia. Possibly a "Tarise" language.

I recently (apr 2020) changed the diachronics here after letting them stand untouched for a few years, so I may have forgotten the reasoning behind my original ideas. See history for details.

The velar stop k shifted to t before any /i/.
Possibly expand the environment for this shift, since /q/ > /k/ unconditionally below. It could be that all /k/ becomes palatalized when bordering either of /e i/ in either direction, and then this sound shifts later to /t/. This would imply that /h/ was shifted to /s/ in at least as many environments.
In unstressed syllables, the vowels i u shifted to ʲ ʷ, thus creating consonant clusters.
The classifiers i- hi- gi- in most words here contract to just a /ʲ/, which crosses the syllable boundary and palatalizes the following consonant. However, they were not common enough to trigger a new grammar rule, and these words were simply padded with more classifier prefixes in most cases. Likewise, other classifier prefixes that contracted to just a single consonant were also padded with other prefixes whose vowels had not contracted.
In unstressed syllables, the vowels e o shifted to i u.
The uvular stops q qʲ qʷ shifted to k k kʷ unconditionally. Then hʲ shifted to s.
Labialization was lost before a consonant or at the end of a word.
Word-final consonants in unstressed syllables were deleted. (This only occurred when a word ended in two unstressed syllables.)
In word-final position, the consonants pʲ bʲ mʲ shifted to t r n.
Palatalization disappeared at the end of a word.
It is possible that a separate /č/ might somehow pull through.
The sequences ʷa ʷe ʷi ʷo ʷu shifted to o o u o u.
All remaining e o shifted to ə.
This seems unlikely. A new long /ē/ is generated below but it would be very rare.
Between vowels, the voiced sounds b d shifted to w r .
The resulting vowel sequences change to single (usually long) vowels.
The consonants p g shifted to h Ø unconditionally. It may be that this /h/ shifts to /Ø/ almost immediately as well, however, as it had a restricted distribution.
Alternatively, though, /b/ > /p/ in final position and /p/ becomes like that of Galà.

Possibly /ja/ > /e/, but /ai/ > /ē/ (also as in Khulls).

Core Andanic

Late Andanese is the most important member of this branch.

Old Andanese (1900) to Late Andanese (4178)

Labialized stops pʷ tʷ kʷ qʷ all changed to p.
The sequences hi ki ti shifted to s before a vowel.
The voiced velar fricative g disappeared to Ø.
The labial fricative f changed to h in all positions.
All remaining occurrences of the labial approximant w shifted to l.
This may have been culturally influenced by Play's shift of /l/ > /w/, which had occurred a few hundred years earlier, since the two peoples lived side by side and the choice of [l] or [w] could have become a marker of tribal membership even though the two groups never spoke the same language.
The uvular stop q disappeared to Ø unconditionally, except in the rare cases where it occurred after a nasal; there, it shifted to k. Unlike in Galà, this exception to the sound shift did not spawn analogical exceptions in word-initial position.
The mid vowels e o changed to i u in all positions.
Clusters such as mt became homorganics like nt, even though the origin of all such clusters was a syllabic nasal which was held longer than the following stop.
The sequences ns ŋh shifted to n ŋ.
After a high tone, the voiceless fricative h shifted to k, as in surrounding languages such as Moonshine, Leaper, and Galà. The timing of this shift may be incorrect here, but it postdates the change of /hʷ/ > /h/, and therefore there is no /hʷ/ > /p/ shift to go with it. Since the shift in Galà happened shortly before 3750 AD, it is likely that that is about the time it happened in Late Andanese as well.
The fricatives s h occurring *before* a syllable ending in a nasal coda shifted to n ŋ.
The remaining sequences mp nt ŋk shifted to m n ŋ. (Note that most of these were from earlier /mh nh ŋh/.)
All remaining coda consonants disappeared.
Tones were eliminated except in syllables with no consonants.
Tones were eliminated. Andanese now had only 9 consonants, 3 vowels, and no tones, and was entirely CV, thus making it the most phonologically simple language in the world.

The above list does not account for the areal shift of /h/ > /k/ after a high tone.

Note that the changes in the syllabary are C>G and X>S. These are not sound changes and cannot be easily derived from the above list; essentially it happens that /g/ (spelled "X" above for clarity) disappeared first, and /s/ appeared first, therefore /s/ replaced X. Later, /ŋ/ replaced /q/. Andanese had actually had a phonemic /ŋ/ even before it had phonemic /s/, but it had not been necessary to spell it separately.

West Andanic languages

note, this may be incorrect, these are3 just "left over" languiages. however, west would be a good place to put them since they cannot be in the mtns. Spoken west of Nèye. No state name given on map.

East Andanic languages

In 1823, many Andanese people settled in Thaoa. At first they were welcome, but they soon became hostile. If East Andanic goes through the same sound changes that Thaoa did, it would mean very little, perhaps only performing /kʷ/ > /p/, and perhaps not even that (since the aspirate form did not shift). Note that Old Andanese had no voicing contrast, so Thaoa's shifts would only affect allophony, and could not create any new phonemes. However, another possibility is that East Andanic lost its tones by Thaoa's influence, perhaps shifting all high tones to long vowels, or only under certain conditions, so that the short vowels would still predominate.

These languages may go extinct in 2686 like Thaoa did, and be replaced much later with an unrelated strain of Andanese from the core (Late Andanese). However, if the minority dialects of Thaoa like Tuq survived, then East Andanic almost certainly would have survived also.

East Andanic may thus survive as late as 4151, when Creamland secedes from the Anchor Empire, and may even become the majority language of Creamland. However, the capital of Creamland would have been entirely Play-speaking. Further, Andanic would have survived only as the language of a minority (Andanese) within the underclass (Pabaps), as the Andanese and Pabaps were lumped together by the ruling class and partly also for their mutual protection.

If Andanese survived in Sala, it seems likely it would survive in at least some parts of Creamland. Again, though, it is not necessarily East Andanic that survived. Core Andanic must have at least swept as far east as Thaoa proper since it is the superstratum of Palli.

Tapilula (0) to Pre-Olati (1300)

Note: these South Andanic languages broke from the main community earlier than the others, and therefore preserve voice contrasts in stops.

The Andanese/Gold dialect of Tapilula had the consonants

Rounded bilabials:                     hʷ  w
Spread bilabials:      p       m   b   f  (Ø)
Alveolars:             t       n   d       l
Rounded alveolars:     tʷ      nʷ  dʷ         
Velars:                k   ḳ   ŋ   ġ   h   g

The accent pattern involved in certain infixes with accented schwa switched to favor the following vowel. e.g. ăpo "field", apə̀ho "field (possessive) became /ăpo apəhò/.
The "labial" vowel ə disappeared, syllabified nearby consonants or turned to i if the nearby consonants were not possible to become syllabic. Note that it never occurred after labialized consonants. Sequences such as /pəh/ collapsed to form aspirated consonants, though these behaved as clusters.
The velar nasal ŋ changed to n in all positions.
The stops p b t became w w k (the /t/ shift was allophonically [th > tx > kx > kh]) except when occurring:
After an accented or high-tone vowel (but not before);

In a consonant cluster of any kind; or

In a monosyllabic word.
tʷ dʷ nʷ > kʷ ġʷ ŋʷ.
The labial fricative f shifted to hʷ.
The labialized sounds kʷ ġʷ hʷ w changed to k ġ h g when they preceded a vowel followed by a labial consonant (including /w/).
ŋʷ> ŋ.
final /b/ > /w/. must, now, add to other list!
Tautosyllabic vowel sequences òi ài èi converged to ē. This did not affect syllable-straddling words like /tùya/. Likewise, èu àu òu in the same environment converged to ō.
Duplicate vowel sequences àa èe ìi òo ùu shifted to long vowels ā ē ī ō ū. But the same sequences with the opposite tone pattern did not shift.

Proto-Olati (1300) to Olati-A (2672)

See Oyster language.

The Olati languages are known both as South Andanic and West Subumpamese. They are Andanic by genetics, but primarily Subumpamese (and partly Naman) by culture.

This is probably the language of Čudiabu (Yuenan), and perhaps also the language of Yuez (meaning that the language crosses state borders, unlike most others). If it is spoken in Yuez also, the name of the language should be Dukelŭta. If not (and it is likely not), then a fifth language will need to be created for Yuez.

The labialized consonants kʷ ḳʷ ġʷ hʷ w shifted to p p b f v unconditionally.
The aspirate clusters bh dh shift to p t.
The velars k g ġ ŋ h shifted to č y y ň š unconditionally.
This is the source of the hiatus in the state name.
The sequences py by fy shift to t d s before a vowel.
This might not be necessary, as /p/ is historically composite, and sequences like /kʷia/ and /bhia/ would likely have not existed anywhere in the entire lexicon.
The uvular stop ḳ became k.
The sequences lh shifted to s.
Remaining aspirated clusters deaspirate.
Before a vowel, the sequences ay ey oy shift to ē. iy uy shift to ī.
Long vowels before hiatus became short, but retained their stress. (Thus, /ia/ is really /ia/, not /ja/).

Thus the consonant inventory was

Labials:     p   b   m   f   v       
Alveolars:   t   d   n   s       l
Palataloids: č   ǯ   ň   š       y
Velars:      k

And the vowel inventory was /a e i o u/, on two tones, and a long series.

Proto-Olati (1300) to Olati-B (2672)

This may be the same as the Yoy language.... note that Yoy is said to have preserved voiced stops, like Olati, and unlike the rest of Andanic.

The consonant inventory as of 1300 AD was

Labials:        p       m   b      (Ø)
Alveolars:      t       n   d       l       
Velars:         k   ḳ   ŋ   ġ   h   g
Labiovelars:    kʷ  ḳʷ      ġʷ  hʷ  w

There were relatively few sequences of two or more consecutive vowels.

The clusters bh dh ġh ġʷh gh shifted to p t k kʷ h. Then mh nh ŋh became mp nt ŋk.
The voiced sounds g ġ ġʷ b shifted to Ø Ø w w. This set up a consonant gradation in which words with hiatus in their bare form developed an oblique form with -/k/- (generalized from a choice of k~p~h).
NOTE: the /k/ alternant was almost certainly the rarest of the three by a wide margin. It should either be /p/ or /h/. Or, it could merge with gender based consonants harmony.
Remaining aspirates disappeared.
The voiced stop d shifted to r. It was still [d] word-initially.
The labialized consonants kʷ ḳʷ hʷ shifted to p p f.
Before a vowel, the sequences aw ew ow shifted to o. Then iw uw shifted to u.
Before a vowel, the sequences ay oy uy shifted to e. Then iy uy shifted to i.
The vowel sequences òe òi ài èi èe ìa shifted to ē. Then èo èu àu òu òo ùa shifted to ō. Lastly, àe èa àa òa ào shifted to ā.
Note, the name of the language is properly /jōj/, hence it comes from an earlier triple sequence.
The vowel sequences ùe ùi ìi ìe shifted to ī. Then ìo ìu ùu ùo shifted to ū.
Before a high-tone vowel, the sequences ti ni ri li shifted to s n y y. Then, ki ḳi ŋi hi in the same environment shifted to č č n s.
Remaining ḳ shifted to k .

Thus the consonant inventory was

Labials:       p   m   f   w
Alveolars:     t   n   s   l   r
Postalveolars: č           y
Velars:        k   ŋ   h

Proto-Olati (1300) to Olati-C (2672)

It is possible that this is the language spoken in Vuʒi, assuming that Vuʒi's own language was lost in the wipeout.

The consonant inventory as of 1300 AD was

Labials:        p       m   b   f   w
Alveolars:      t       n   d       l       
Velars:         k   ḳ   ŋ   ġ   h   g
Labiovelars:    kʷ  ḳʷ      ġʷ

There were relatively few sequences of two or more consecutive vowels.

The voiced stops ġʷ ġ shifted to w g.
The voiceless stops p t shifted to f s . This shift included the aspirate sequences /bh dh ph th/.
The voiced stops b d shifted to p t unconditionally.
After a low tone or word-initially, the voiceless stops k kʷ shifted to x xʷ.
The voiceless ejectives ḳ ḳʷ shifted to k kʷ unconditionally. The combination of the above shifts and grammatical levelling created a consonant gradation where words with /p t k kʷ/ as the last consonant in the word shifted it to /f s x xʷ/ to form the oblique. By analogy, some words in which historical /p t k kʷ/ had become /f s x xʷ/ now reverted to /p t k kʷ/ in order to use the gradations.

Proto-Olati (1300) to Dakʷòhi (2672)

This is the language spoken in Dakʷòhi, which is the state labeled as Mania on some maps.

The voiced sounds gʷ ġʷ ġ shifted to w w g.
The voiceless stops p t shifted to f s . This shift included the aspirate sequences /bh dh ph th/.
The voiced sounds b d g shifted to p t x unconditionally.
The velars k ḳ g ŋ x shifted to č č y n š unconditionally. Note that this /x/ is distinct from /h/.
The labialized consonants kʷ ḳʷ w shifted to k k v.

Yoy language

The Yoy language was spoken in the Thunder Empire from 3844 AD to 3884 AD, during (and only during) the time when the THunder Empire was governbed by Dreamland. Yoy was confused with Dreamlandic, and the two languages had a similar sound, but they are not closely related. Rather, the anti-Thunder policies of the Dreamers allowed minority languages such as Yoy to flourish. After the overthrow of the Dreamer government, Yoy went back into suppression; however, the people who spoke Yoy were generally anti-Dreamer by this point as they had been no better treated by the Dreamers than were the majority Thunderers.

Characteristics

Deletion of /b d g/ after a stressed vowel; before this, /ab/ > /o/. This meant that the only stops that occurred between vowels were /t k q/.
Probably, the loss of voicing contrasts in stops altogether, since there would be relatively few minimal pairs by this point (only /t/ vs /d/ after unstressed vowels). Possibly remaining /d/ > /r/ before this happens.
The growth of falling diphthongs resulting from the deletion of these consonants, without the monophthongization that characterized most languages of this area. THe name Yoy would have quicvkly become *Yē in most of the neighboring langs, even those not closely related.

Culture

Early on I had written that all Andanic languages die out by 4175 AD. This is still possibly true if West Andanic is considered to be culturally Subumpamese, Oyster, or some other label ... and because it broke off at around 1300 AD, it could be considered non-Andanic linguistically as well. However, survival of minor languages like Galà is possible.

Notes