Tropical Rim/Extension

SCRATCHPAD

Sound changes [minimal]

13:32, 25 November 2022 (PST)

Tapilula to Hapag

1. The aspirated velar stop k became č before the vowel /i/. If another vowel followed, the /i/ disappeared. This happened even if the /i/ was accented.
2. When a "velaroid" consonant (k ḳ ŋ h g l) followed an accented high tone vowel, the vowel metathesized, leaving a superheavy syllable with both a two-vowel sequence and a coda consonant. Thus, for example, /àli/ > /ail/. These closed syllables were all high-toned, and are thus written without tone marks.
3. A schwa before another vowel in any syllable disappeared. Thus əa əe əi əo əu əə shifted to a e i o u ə. This happened in both open and closed syllables.
4. The sequences iu and ui shifted to ə̄.
5. The double-vowel sequences aa ee ii oo uu əə shifted to the single vowels a e i o u ə in closed syllables only.
6. The sequences ii uu əə (which now occurred only in open syllables) shifted to əi əu ə.

   It is not clear if this /əə/ is /ə̄/, but it most likely is.

7. The sequences ie uo shifted to i u in open syllables only.

   This shift did not happen in Gold, but did happen on both sides of Gold. Either Gold performed the shift and then later undid it (some knowledge of the original vowels would have been preserved in inflections), or more likely, the shift did not actually happen until much later but is presented high up on the list because it happened independently in all non-Gold branches.

8. The remaining double-vowel sequences aa ee oo, which occurred only in open syllables, shifted to the long vowels ā ē ō.
9. The sequences ai ei oi merged as ei; the sequences au eu ou merged as ou.

Hapag to Trout

1. The sequences ea eə shifted to ee; meanwhile, oa oə became oo. Then, aə shifted to aa. Thus, the sequences /ee aa oo/ once again appeared in both open and closed syllables. Note, however, that much inherited /ea oa/ had participated in grammatical alternations with /əa/, which had become a simple /a/ by this time, and this is the form that was usually generalized.
2. The sequences ia io iə shifted to ī . Then ua ue uə shifted to ū.

   This is the latest possible date for Gold to undo the shift of /ie uo/ > /i u/ above. After this point, even through analogy with inflected forms the original vowels would have been lost.

3. In absolute final position, syllable-final ŋ changed to n. (But see below.)
4. Accented vowel-initial syllables gained a pharyngeal ʕ as an onset. Then the clusters nʕ kʕ shifted to g ḳ.
5. After long vowels, all consonants became voiced. This created the new consonants v ǯ . Thus, final -h in words like hʷīh became -g. This sound change did not affect diphthongs. There was no voiced velar stop, as all four velars simply shifted to fricatives.
6. Consonants occurring after initial vowels also became voiced.
7. After initial unstressed /u/, all consonants other than palatals became labialized. This change extended even to clusters. Because of the voicing rule, however, all of these consonants were voiced. bʷ vʷ merged as w.
8. In word-initial position, the six vowels a e i o u ə were deleted to Ø unconditionally.

   A few words began with double vowels, but these were grammatical alternants of V-/g/-V, V: (that is, inherited long vowels), and /g/-V-(g)-V pairlets, meaning that in these few words, the vowel deletion could affect both vowels, only the outermost vowel, or neither of the two vowels. Then, because of this same analogy, the words that now began with g came to be pronounced with Ø onset when following a word that ended with a consonant, and for some speakers also in isolation, but never when following a word ending in a vowel. In both Leaper and Play, this shift was quickly undone, as it had never been phonemic. In Thaoa, it may have become firm, but with the expense of now having no words beginning in /g/.

   Sometimes root-initial vowels were retained due to classifier prefixes; here, too, analogy played an important role. For example, the food item prefix mi- cleaved onto the root, and ceased to be seen as a classifier prefix, meaning that these new roots began with consonants after all. This was helped by the fact that only a few such roots began with vowels, even including those that had come to be pronounced with /g/.

9. All schwas and diphthongs became low tone.
10. Labialized consonants lost their labialization when occuring after another labial or labialized consonant.

Trout to Diver

1. The voiced coronal obstruents d ǯ shifted to r ž.
2. The sequences ae ao shifted to ai au.
3. The labialized obstruents tʷ dʷ gʷ shifted to kʷ w w.
4. The velar ejective ḳ shifted to q.
5. The sequences č kč merged as s; preceding vowels retained their tones.

   The wordlist implies that this /č/ is /ky/.

6. The labialized nasals mʷ nʷ ŋʷ shifted to m ŋʷ ŋʷ.

   NOTE ON POLITICS: Sleep Flower breaks off here.

7. f fʷ v shifted to ḥ ḥʷ Ø.
8. The clusters kx kh (and their labialized counterparts) shifted to k.

Trout to Gold

See Languages of Teppala and Gold language.

Trout to Wax

1. In unstressed position, the voiceless ejective ḳ shifted to a plain k.
2. After a high tone, the voiced sounds g w shifted to G Gʷ.

   Earlier here was listed a shift of /w kʷ/ to /b p/ after a high tone.

3. The cluster ng shifted to N, unlike any other language.

   This might also include /gn/, assuming /g/ overlaid the syllable boundary.

4. All monosyllables became low tone, even freestanding content words.
5. The voiceless stops p t k ḳ phonemicized their aspirated allophones ph th kh kh when they occurred as the onset of a stressed syllable. This happened because in all other positions, they became deaspirated.

   It is possible that they were also aspirated word-initially even when unstressed, perhaps subject to certain conditions.

6. The sequences hp ht hk hḳ shifted to ph th kh kh. This could only occur after a stressed syllable.
7. The sequences tʷh tʷ dʷ nʷ shifted to kʷh kʷ ġʷ mʷ.
8. The labialized nasals mʷ nʷ ŋʷ merged as mʷ.

Trout to Pēles

1. The sequences əi shifted to ə̄.
2. All consonants occurring after the vowel /u/ (any length, any tone) became labialized.
3. The high vowels u ū shifted to ə ə̄. Then əu shifted to ə̄ as well.
4. All consonants occurring after the vowel /i/ (**NOTE this excludes earlier /əi/**) became palatalized.
5. The high vowels ə ə̄ (including earlier /u ū/) shifted to i ī.
6. The high vowels i ī shifted to yi yī .
7. When an /a/ was in an adjacent syllable, the sequences e ē ei o ō ou shifted to ya yā yai a ā au.
8. The sequences eḳ oḳ (on any tone) shifted to aḳ.
9. The sequences e ē ei shifted to yi yī yi .
10. The sequences o ō ou shifted to u ū u .
11. The sequences aa ae ao merged as ā.
12. The labial fricatives f v shifted to h Ø.
13. Any consonant that was both labialized and palatalized became labialized alone.
14. The labialized consonants kʷ ḳʷ čʷ tʷ pʷ merged as p. Then, mʷ nʷ ŋʷ shifted to m. The voiced labialized stops dʷ bʷ merged as b. Lastly, xʷ gʷ shifted to f w.
15. The sequences kʲ ḳʲ ŋʲ xʲ gʲ shifted to č č n s y.
16. The clusters kp kb km kf shifted to pp pp pm p. Then kt kd kn shifted to tt tt tn . (/ks/ did not occur.) Then kč became čč. (/kŋ/ remained, and kh, kg, kk, etc had been eliminated in the proto-language although those shifts are not listed.)
17. The voiced coronals d dʲ shifted to r ž.

Trout to Tarise

1. The voiced fricatives g gʷ shifted to Ø w. However, the fricative allophones remained, and therefore came to also replace original Ø~ʕ. For example, syllable final -u merged with original gʷ, and obtained the velar frication as an allophone after a stressed vowel.
2. New vowel sequences became rising diphthongs.
3. The stops p b d shifted to h g r unconditionally, with palatalization and labialization both preserved. The voiceless stop t acquired an allophone of s in some positions, but did not split fully from the parent phoneme. This explains the different realizations in the daughter languages.
4. The fricatives f v also shifted to h g.
5. All unstressed syllables became CV only, with no tones.
6. The sequence ʷə shifted to ʷu.
7. The vowels ə i shifted to i yi unconditionally.

Attempts to pin down location

12:38, 25 November 2022 (PST)

It is possible that Thaoa is off by itself, somewhere in the equatorial zone, perhaps replacing the unlabeled Tropical_Rim#Area_between_Amade_and_Fernland or on an island. Another possible location is Wax, in which case it would be submerged by Gold very early but might survive in outlying areas among people who refused to assimilate fully to AlphaLeap.

The other languages ("ex-Subumpamese" and "ex-Diver") would then border each other, either in the equatorial zone as well but further west, or concentrated into the northernmost part of what later became Lobexon, on which case they would be on the fringe of the tropics and would border Nama.

New names

Subumpamese's new name will derive from MRCA ḳĭḳa-t- mbògo hà-pə̀ga. It is most likely that this is the northernmost territory, so that Diver can border AlphaLeap. Subumpamese is the earliest-branching language, and its name at the time of separation (pre-Trout) would be Ḳiḳabog, assuming the two words were grouped into one.

Diver's new name will derive from MRCA ŋà gè yàta hà-pə̀ga. At the Trout stage, this name would be Ŋaeg Yàta, unless it were grouped into a single word.

Thaoa's new name will derive from MRCA mbĭha-ḳ- hù hà-pə̀ga. The first part of the name would likely be reshaped into a single ethnonym, Mbihàhu (not with a -k-) early on, and therefore Thaoa's new name is whatever the reflex of mbihàhu would be. At least at the Trout stage (1085), this name would be Bihauh.

It is possible that the hà-pə̀ga morpheme will be fronted in some of the languages. In the meantime, a name like Hapag can refer to all three languages.

Thaoa scratchpad

Early loss of tone

Both Gold and Andanese innovated a new short low tone CV monosyllable word shape, whereas in the MRCA all CV monosyllables were high tones. Likewise, both Gold and Andanese allowed longer words to end in a short low tone (stressed, because all tonic syllables were), though perhaps only in verbs. Even Dreamlandic shows evidence of having allowed final stressed low tones. This is difficult to explain. It could be that:

• Gold and Andanese separately evolved the ability to have short low tone CV monosyllables. This is no great trouble, as the words in that class show little overlap between the two families.
• Short stressed low tones in verbs arose at a stage when such verbs were typically followed by a CVC particle, which meant that the final syllable of the root was penultimate and thus could take either tone.
• Then, this variable tone on the final syllable became analogized to the tone pattern of the root, meaning that if the first syllable of the root had an underlying high tone, then the tonic syllable came to have a low tone, and vice versa. This could have happened before the breakup of Gold and Andanese, but if this is the case, it would have also occurred in Thaoa.

It is possible that Thaoa regains tonal contrasts from surrounding influence, but in a "dead" way that adds almost nothing to the language, and just "takes up time".

Descendants

Sakhi and Palli are the two main descendants of Thaoa. Both languages survived despite Thaoa's resounding military defeat and resulting occupation after the Oyster War in 2668. The two branches were separated from each other, however, and had survived the war for different reasons.

There were some languages, such as Tuq, that survived the war because they had been against Thaoa all along and were never invaded.

Hamatap

It is also possible that Hamatap is not Moonshine, but rather a descendant of Sakhi, perhaps separating as far back as 2668 itself, or even further back. This would allow the Hamatap language about 1,300 years of divergence instead of being separate from Moonshine only as a political fiction. The Hamatap tribe was living much further north than one would expect for a branch of Thaoa, however.

If Hamatap is separate for 1,300 years, it would make more sense historically, since it would give the other tribes "someone to hate" instead of the Hamataps just breaking away only after the Play party became prominent. Thus Hamatap would join the Players as a means to settle old grudges against the tribes around them. These divisions would be older than Moonshine's Feminist Compact, and would probably not center around feminism.

Other information

It is possible that Sakhi becomes extremely feministic like Moonshine, but only if they adopted these traits from Repilian languages, since Moonshine did not border Sakhi territory until 3958. At least some influence could have spread in the opposite direction ... Sakhi innovations being adopted in Moonshine, whether through Repilian intermediaries or not, but note that the Repilian languages would have been almost completely extinct by this time. The language turnover was mostly Repilian ---> Sakhi ---> Moonshine, rather than Moonshine driving out both Sakhi and the Repilian aboriginal languages.

Sharp object ergatives

The MRCA had an infix specifically for sharp objects acting on human patients: -tu-. This was used to create verbless sentences such as ŋa-tʷ-ŏti "the thorn (cut) me", and similar verbless sentences exist in Play, Poswa, and other languages that lost their classifier prefixes and then created a new setup based on suffixes instead. Thus it is an areal phenomenon. The primordial affix -tu- specifically points out a sharp object, whereas the Play and Poswa constructions are ergatives broadly speaking. Thaoa may have both paradigms.

In Thaoa the sharp object ergative might remain as -tu- before a consonant-initial noun, and change to -(p)ph- before a vowel. The geminate would appear when padded by another prefix and the singleton otherwise. Thus one could say something such as ŋapphogu for "the thorn cut me". Here, the part that means thorn is just /o/ and relies on an idea that Thaoa could preserve very short roots when appropriately padded by classifiers. Using the longer root for thorn would still work: ŋapphotigu.

Bipersonal verb prefixes

The bipersonal verb prefixes for soft inanimates were monosyllabic in at least their 1p, 2p, and 3FEM agent form, as well as some others. These were:

FREE        a   e   i   o   u   ə
1p         na      no      no  
2p             ge  go  go  go  gə
3pFEM      mo  mo  mi  mo  mu  mo
3pMASC         te  te  to  to  tə
EPICENE    po  po  pi  po  pu  po

Blanks represent areas that cannot resolve to monosyllables through diachronics alone and thus must rely on analogy. The epicene should have been eliminated by sound changes, but it may have survived as a "singular" (e.g. one man and one woman) and then come to be plural again.

Note that /pə/ probably means 1p>2p.

Reanalysis

Meaningless /-əg-/

Thaoa could make all CV roots equivalent to C-əg-V roots since the reflexes would be the same due to the /gg/ > /g/ rule. It could be that at a later stage, /əg/ > /Ø/, but this would not help anything in the early stages of Thaoa. The intent here is to get a schwa that can participate in e~ə~o even in words that never actually had it.

This would work well if the CVC roots were also refashioned as CəCV, which is trivial already except that final /-n/ would appear as /ŋ/ instead. A few very important roots might resist the tide and stay true to their etymology; note that Thaoa has quite a few more choices than Play in this task, since it has a full six-vowel inventory whereas Play (and Gold) had just four vowels of which the schwa was most often composite.

Meaningless /-g-/

Likewise, note that most and possibly all nouns end in /-g"V"/, where the vowel is meaningful but the consonant is not. This consonant nonetheless changes to other values to mark noun case in at least some nouns (but not those covered by the /-əg-/ paradigm above).

CVCV

It is even possible that all CVCV roots can be reanalysed as CVC, with the final vowel being assumed to be part of a classifier suffix. This is difficult to justify, as the morphology of CVCV roots is very different from CVC+vowel, so perhaps it could be that the suffixes never applied to CVCV words and they were simply "left alone".

Inanimate classifiers

Inanimate classifiers in Thaoa probably all either began with vowels (the "soft" inanimates) or with coronals (the "hard" inanimates). Hard inanimates needed 2-syllable prefixes to interact with agents, because the coronal consonants were never lenited by sound changes nor removed by analogy. The more common soft inanimates used monosyllablic prefixes.

Prenasals

Try to find a way to get prenasals into the language, as it will miss out by padding CVC sequences with vowel-initial suffixes. Remember nadambo .... that is, just as prenasals are eliminated from the language by morphology, perhaps a new morphological structure can appear that specifically favors prenasals.

Analogy involving schwa

Thaoa can be like Leaper in assuming a schwa was the initial vowel in CVC words like Gold kuḳ "book". This could then participate in e~ə~o alternations.

Agent prediction without pronouns

16:14, 22 March 2022 (PDT)

If Thaoa has no pronouns, it can mark the agent on the patient just like Poswa (and to some extent Play) and use an OV structure. Using the pine tree rule, assuming na-i- > no-, the listener would know that the agent was 1st person and that the patient was in the /i-/ class, which included handheld objects (Trout /gi-/) and certain other things that were lumped into the class as initial consonants were analogized out. Unique to Thaoa is that any /e ə/ in the stem of the patient would change to /o/ to agree with the prefix. Likewise, the pine tree rule also predicts gə-i- > go- for a 2nd person agent and a handheld patient. This /g-/ does not drop out.

As in Andanese, these prefixes would also be attached to the verb, as though the verb were also part of some sort of "handheld verb" class. However it is not certain that Thaoa will also do this when the S and O prefixes remain as independent syllables.

Why /-g/ can survive

Since coda /g/ deletes in most clusters, Thaoa can lose word-final /g/ when it needs to without specifically needing a sound change. Remember that /g/ was often the first letter of the next word anyhow, and /gg/ > /g/ would be the easiest change of all.

Inefficient genitive construction

Thaoa was known for its inefficient grammar. Moonshine did not exist at the time of classical Thaoa, but later Moonshine scholars looked at the Thaoa language with disdain. It was "sick", as they put it, having lost just enough of the inherited Trout grammar to miss out on the efficiency of Play and Moonshine (and Andanese), without gaining new features of its own. For example, it had to use two words to express the genitive, instead of using compounds. Andanese was sometimes slower than Thaoa, but the teachers explained this as a consequence of Andanese's much smaller phonology and its use of repetition.

Moreover, the genitive construction that Thaoa relied on was inefficient even compared to the lesser used counterparts in neighboring languages.

For example, Late Andanese transformed ituakiu "book" into nutuakiu "library" just by changing the classifier prefix from the handheld object class to the building class. Thaoa had cognates of these classes, but many classifiers had been lumped together, and so such word families were primarily old coinages. But because the morphology had not fallen off, Thaoa could not use Play-style compounds either. Thus Thaoa had ikhusso "book" and okkoga okokuso "building of books; library", with classifiers on both words. While in theory okokuso could mean library by itself, as similar ellipsis was used in related languages, this did not happen often in Thaoa because the classifier prefixes were not distinct enough on their own to reliably carry the meaning, and because since this is already a genitive form, there is no convenient way for Thaoa to make a genitive of it; the head word, okkoga, must appear.

But note that Thaoa does not die out, even after military defeat, because of its speakers' inability to pick up Play. Therefore, with Andanese, Play, and Moonshine all influencing the language, some sort of innovation that would allow words like okokuso to stand alone and yet contrast with simple genitives is almost certain.

The name of Teppala

01:50, 1 March 2022 (PST)

Note that the sound changes below assure that nothing even close to the name Teppala can arise within Thaoa, and at best it can only be explained as a distorted trade name. The original name would need to have three morphemes, such as dì nìi ndʷò, and then the infix that provides the /l/ would be added in Thaoa only.

May 16, 2021

The bipersonal classifier prefixes beginning with n- and ŋ- had probably not decoupled from their following nouns yet. In at least Gold, which also merges the vowels a e o into a, they come to be reinterpreted as a suffix on the preceding word ending with -n. Then the prefixes disappear.

It is possible that in Moonshine, the prefixes are somehow retained, or that they are sent back to their original position by subsequent grammatical changes.

GRAMMAR

YYYYY

This is the group that was once called Subumpamese.

Morphonology

Vowel harmony

Pretonic /e~ə~o/ are in harmony with the tonic vowel, but every morpheme has a basic form that appears when the tonic vowel is one of /a i u/. Also, after a labialized consonant, only /o/ appears in this position, and this causes any vowels earlier in the word to also become /o/ regardless of the tonic vowel. Note that this harmony persists in Subumpamese despite the fact that the schwa vowel /ə/ is a high vowel, not a mid vowel. Also note that sequences like /kʷe/ do still occur in the stressed syllable and in every syllable thereafter; the harmony rule only applies to syllables that occur before the stress.

This pattern is responsible for alternations like tekʷēł "his bone" vs. tokʷŏlo "his fern".

Verbal morphology

Verb prefixes indicate both the agent and the patient; for example, in tobòči "he marries her", the prefix tob- indicates a male agent and a female patient. This arose from the stacking of agent and patient classifiers; most other Lenian languages mark only the agent.

Person marking is redundant, since Subumpamese retains the pronouns lost in other branches. The suffixes -k- (1st person) and -h- (2nd person) attach to the oblique stem of the verb, whereas 3rd person is marked by no suffix.

Other inflections

Kava was isolated from the Gold language for most of its history, and therefore took most of its influence from the grammatically dissimilar Old Andanese language. This caused Kava to develop a very simple grammar, losing most of the Subumpamese suffixes, while gaining no new prefixes or infixes from Andanese. A new part of speech called an auxiliary verb or weak verb appeared, which carried the meaning of inflections and behaved like verbs except that they did not carry the classifier prefixes that full verbs did.

These auxiliary verbs were suffixes, not separate words. Therefore, they functioned like case markers, and were just like those of Gold except that they were not fusional and never carried the word's stress. They included:

ADVERBIALS

1. si ~ ši (genitive)
2. su ~ hʷù (accusative)
3. hà to be changed by

LOCATIVES

1. -m(ə) (locative of place)
2. n(ə) (locative of motion)
3. ma on top of; used as a suffix after -m
4. mo on top of; used as a suffix after -m
5. pì supported by; used as a suffix after -m
6. ši underneath; used as a suffix after -m
7. ī behind; used as a suffix after both -m and -n
8. ŋò with; next to; near
9. ga in front of
10. c̀e covering; standing over
11. tà to push on; used as a suffix after -n
12. là to pull on; used as a suffix after -n

Of these, only the genitive is cognate to it's counterpart in Gold; the other resemblances are due to convergent evolution.

There was also a new copula verb, tò.

Morphosyntactic sound changes

Nouns ending in -x usually dropped the -x because it disappeared before the three most common case endings. Thus, for example, *pipēx changed to pipē "ocean; salt water".

However, in some nouns, it survived because these nouns were originally strong.

Nouns

Noun class prefixes are augmented to CVC before vowel-initial stems. Some of these have bled into the stems and created new roots beginning with the extra consonant, which then appear in other noun classes.

Note that /s/ appears whenever any primordial /h/ is bordered by /i~e/ and /u/ in either direction.

Strong nouns

A small number of nouns retained their case marking; nominal complexity increased west to east. This applied to the whole sprachbund, shading from Kava with no inflections to Paleo-Pabappa where the entire vocabulary was strong. However, the nearby Eastern Subumpamese languages still used weak noun morphology for the majority of their vocabulary.

DIACHRONICS

wipe

Sleep Flower

The phonology of proto-Highland was

Bilabials:            p   m   b   f   v
Alveolars:            t   n   r   s   l
Palatals:                     ž       y
Velars:               k   ŋ       h   g 
Labiovelars:          kʷ  ŋʷ      hʷ  w
Uvulars:              q
Rounded uvulars:      qʷ

The vowel inventory was

Short vowels:          a  e  i  o  u  ə
Long vowels:          aa ee ii oo uu 
Falling diphthongs:         ai    au
                            ei    ou
                            əi    əu

The /qʷ/ was only in compounds.

See edit history <Nov25-2022 for old snd changes.

Šàno

This language was spoken by a coastal (southwestern or southeastern) tribe. The name is an exonym. Note that the development is very similar early on to Subumpamese.

The consonant inventory of the mainland dialect of Tapilula was

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

1. The aspirated velar stop k became č before the vowel /i/. If another vowel followed, the /i/ disappeared. This happened even if the /i/ was accented.
2. When a "velaroid" consonant (/k ḳ ŋ h g l/) followed an accented high tone vowel, the vowel metathesized, leaving a closed syllable.
3. A schwa before another vowel in any syllable disappeared. Thus əa əe əi əo əu əə shifted to a e i o u ə. This happened in both open and closed syllables.
4. The sequences iu and ui shifted to əə.
5. The double-vowel sequences aa ee ii oo uu əə shifted to the single vowels a e i o u ə in closed syllables only.
6. The sequences ii uu əə (which now occurred only in open syllables) shifted to əi əu ə.
7. The sequences ai ei oi merged as ei; the sequences au eu ou merged as ou.
8. The mid-vowel sequences eo eə shifted to ee. Meanwhile oe oə became oo. These four sequences were all rare, however, because of shifts further back in time that affected only mid vowels.
9. All consonants adjacent to an /u/ in either direction became labialized.

   NOTE ON POLITICS: Up until this point, the Šàno language is nearly identical to proto-Subumpamese.

10. The sequences ae ao shifted to ai au.
11. The voiced labiovelar fricative gʷ became bʷ.
12. All labialized consonants become rounded bilabials.
13. In absolute initial position, t >s.
14. In syllable-final position, the voiced velar fricative g disappeared and lengthened the preceding vowel. This often occurred in the second element of a diphthong.
15. Vowel sequences in which the second element was high-tone (less common) lengthened the second vowel, thus merging with the ones which had previously been followed by /g/.
16. uā>wā.
17. Velar consonants moved up: k ŋ h g > č ň š r, probably unconditionally.
18. q>k.
19. f fʷ shifted to h hʷ.
20. In absolute final position, š č ň > s t n.

Thus the consonant inventory was

Rounded bilabials:     pʷ      bʷ  hʷ  w
Spread bilabials:      p   m   b       
Alveolars:             t   n   d   s   l   r
Palataloids:           č   ň       š      
Velars:                k           h 

The language still retained a full six-vowel system and the world's largest inventory of permissible vowel sequences:

aa    ai    au
ea ee ei      
ia ie    io         
oa       oo ou
ua ue    uo
      əi    əu

All seventeen of these occurred as falling diphthongs, but only the nine beginning with /a e o/ also occurred as rising diphthongs. Sequences with two of the same vowel were distinguished by the tone pattern and, when following a labialized consonant, also by vowel color.

Additionally, long vowels were present, and were distinct from sequences of two short vowels. Thus, there were three tones: high, low, and long.

Labialized consonants carried little information, because they inherited the gaps of *ʷə ʷa and rarity of ʷe ʷi from Tapilula, filling these only when bordering a /u/. They were not distinctive in the syllable coda either because the only non-labialized coda consonants that could occur after an /u/ were those that had previously occurred after the diphthong /ao/.

Paleo-Pabappa splits into four languages at this point, but they share most of the immediately subsequent changes.

Subumpam

This may be the "Rasula" people's language family as well.

Tapilula (0) to proto-Subumpamese (~1700)

The consonant inventory of Tapilula was

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       h   g

1. The aspirated velar stop k became č before the vowel /i/. If another vowel followed, the /i/ disappeared. This happened even if the /i/ was accented.
2. When a "velaroid" consonant (k ḳ ŋ h g l) followed an accented high tone vowel, the vowel metathesized, leaving a superheavy syllable with both a two-vowel sequence and a coda consonant. Thus, for example, /àli/ > /ail/. These closed syllables were all high-toned, and are thus written without tone marks. Thus, for example, aa implies àa.
3. A schwa before another vowel in any syllable disappeared. Thus əa əe əi əo əu əə shifted to a e i o u ə. This happened in both open and closed syllables.
4. The sequences iu and ui shifted to ə̄.
5. The double-vowel sequences aa ee ii oo uu əə shifted to the single vowels a e i o u ə in closed syllables only.
6. The sequences ii uu əə (which now occurred only in open syllables) shifted to əi əu ə.
7. The sequences ie uo shifted to i u in open syllables only.
8. The remaining double-vowel sequences aa ee oo, which occurred only in open syllables, shifted to the long vowels ā ē ō.
9. The sequences ai ei oi merged as ei; the sequences au eu ou merged as ou.

   NOTE ON POLITICS: THIS IS THE STAGE AT WHICH TROUT DIVERGES FROM THE SUBUMPAMESE BRANCH.

10. The vowels /u i e/ caused adjacent consonants, in both directions, to become labialized, palatalized, and prepalatalized. The last shift applied only to velars. Labialization and palatalization could stack.
11. The sequences ìa ìo ìə shifted to ī.
12. The sequences ùa ùo ùə shifted to ū. ə̄ also shifted to ū.
13. The sequences ei ou, in both open and closed syllables, shifted to ē ō.
14. Syllable-final h shifted to x.
15. Any fw>hw,then f>h

    Note on politics: Vuʒi split off here.

16. The three syllabic nasals ṁ ṅ ŋ̇ all merged to ən.
17. The velar ejective ḳ became q. Then kq qk shifted to qq.
18. The cluster xhʷ became xʷ.
19. All tones on unstressed syllables are released by spreading the tone of the accented syllable across the word.

    In a two-syllable root, the unstressed syllable acquires the opposite tone from the accented syllable.

    Classifier prefixes and auxiliary verbs all become low tone.

    In compounds, there is no sandhi.

20. The fricative śʷ s̀ʷ šʷ shifted to s. Then ś s̀ became š.
21. The nasals ń ǹ shifted to ň. Then mʷ nʷ ňʷ ŋʷ all merged as m.
22. The sequences km qm shifted to kʷ qʷ.
23. Voiced palatal stops and fricatives all merged as y.
24. The sequences iy ey, on any tone, shifted to ī ē. <---QUESTIONABLE. most of this would have been from ĭg.
25. The labialized palataloids čʷ ǯʷ became the velars kʷ ġʷ.
26. The labialized approximants lʷ łʷ merged as w.
27. The labialized alveolar stops tʷ dʷ shifted to pʷ bʷ.
28. Unaccented final short schwas were deleted. (In nouns, they were retained because they were not always final. Therefore, this shift applies mostly to inflections.)
29. The sequences ʷe ʷi ʷə ʷu, on any tone, shifted to e i ə u. Thus labialization remained distinctive only before /a/ and /o/.
30. Mismatched diphthongs such as /eī/ shifted to /ēi/. Generally these were from a lost final -g.

Thus the proto-Subumpamese language had the consonants

Rounded bilabials:    pʷ  bʷ          w 
Bilabials:            p   b   m                   
Alveolars:            t   d   n   s   l             
Postalveolars:        č   ǯ   ň   š   ł           
Palatals:             ć               y
Prevelars:            c̀        
Velars:               k   ġ   ŋ   x   g
Labiovelars:          kʷ  ġʷ      xʷ  gʷ
Uvulars:              q           h              
Rounded uvulars:      qʷ          hʷ

All consonants were labialized before any /u/ and palatalized before any /i/. However, sequences like si~ši remained distinct. Consonants were also labialized *after* any /u/, so there is no contrast between /upwa/ vs /upa/, even over morpheme boundaries. This means that labialization was contrastive only in a very restricted environment, since the consonant, the following vowel, and the preceding vowel must all be on the list.

The voiced velar stop /ġ/ was a conditional alternant of /ġʷ/, appearing only before vowels that /ġʷ/ could not appear before.

The high vowel sequences were / yi ə yə wu/. Thus, it is almost but not quite analyzable as a single vowel /ɨ/.

See Subumpamese languages for details of the languages that do not survive the Vegetable War.

Proto-Subumpamese (1700) to Sub-Oyster (3141)

This is a substratum of the Oyster language.

Note on culture

It is possible that some Oysters actually spoke Olati-A, since this was the language of Yuenan, in western Subumpam. If so, it's possible that they spoke a very conservative dialect of it which changed little from 1700.

This idea is based on the idea that while the Oysters represented Subumpam and Subumpamese culture, they originated from a peripheral area of Subumpam rather than the capital state of Bipabum.

The designation of Oyster as an eastern Subumpamese language may have arisen from a confusion between Bipabum (the capital) and Yuenan (the most linguistically pure state), in turn caused by the fact that a third state exists somewhere that rejected the Oyster language ... and this state cannot have been Yuenan if Yuenan *is* the Oyster state (even if it had dialects).

1. The high front vowel i, on all tones, shifted to ʲi. This had already happened in the proto-language, but was not phonemic. Note that this is different from earlier shifts that moved the consonant. For example, /ki/ became /kʲi/ here, but not /ći/. Also, this shift applied to labials.
2. All consonants bordering a /u/ in either direction became labialized. That is, u > ʷuʷ. This shift had also happened in the proto-language but was not represented in the orthography. However, the simple spelling /u/ remained, so "u" implied "ʷuʷ". There was, at this time, no /u/ that occurred outside this environment.
3. The high central vowels ə ə̄ changed to i ī unconditionally.

   Note that around this time, the classifier prefix /yi-/ was dropped from the grammar except in bare form. (That is, e.g. bo-yi- became just bo-.) This was not a sound change, but expanded the environments in which palatalized consonants could occur.

4. When bordering a uvular in either direction, the vowel i (on any tone) shifted to ʉ~u, which are the same phoneme, but the ʉ spelling indicates specifically that the surrounding consonants are not labialized.
5. Syllable-final nasals ŋ ň changed to match the place of a following consonant, and changed to n if word-final.

   Note on politics: this may be 2371.

6. The prevelar stop c̀ changed to ć.
7. The high tone vowels à è ì ò ù came to be spelled á é í ó ú. (That is, they were no longer automatically followed by a glottal stop.)
8. The mid vowel sequences o ʲo shifted to ʉ ʲe.

   Plain e apparently also shifted to ʲe.

9. ea ae>ʲa ā.

If the actual Oyster language is Andanic, this language and its entire family is probably wiped out at this point and never replenished by any closely related language. Theoretical sound changes for a survivor population are below:

1. On a low tone, the high vowels i u (including all ʉ) become ultra-short and are sometimes dropped.
2. The long vowels ā ē ī ō ū shifted to á é í ó ú, thus merging with the primordial high tones. (This is why the orthography was changed.)
3. The palatalized alveolar nasal nʲ shifted to ň.
4. The sequences čʲ ǯʲ ňʲ šʲ łʲ shifted to č ǯ ň š y.
5. The sequences ŋʲ xʲ gʲ hʲ shifted to ń ś y ś. Then ġʲ shifted to ǵ.
6. The palatalized rounded bilabials pʷʲ bʷʲ mʷʲ simplified to pʷ bʷ mʷ. These had appeared from sequences like /mumi/+vowel.
7. The sequence hʷɨg shifted to xʷ.

Palatalization can be analyzed as consonant + /j/ or as a property inherent to the consonant. Since some palatalized consonants occur in the coda, this analysis is most convenient:

                      PLAIN                      PALATALIZED

Rounded bilabials:    pʷ  bʷ          w 
Bilabials:            p   b   m                  pʲ  bʲ  mʲ 
Alveolars:            t   d   n   s   l          tʲ  dʲ      sʲ   
Postalveolars:        č   ǯ   ň   š   ł           
Palatals:             ć   ǵ   ń   ś   y         (ć   ǵ   ń   ś   y)
Velars:               k   ġ   ŋ   x   g          kʲ   
Labiovelars:          kʷ  ġʷ      xʷ  gʷ         
Uvulars:              q           h                         
Rounded uvulars:      qʷ          hʷ

All consonants are labialized before and after any /u/ (not /ʉ/); the labialized consonants listed in the table above are those that can appear in other contexts. If the u~ʉ contrast is neutralized by analyzing labialization as phonemic, then all consonants would have labialized variants, even the palatalized ones.

Unlike most other languages, inflections in FILTER did not change the stress pattern, since there was no stress pattern ... e.g. kʉ́pʉ "pine", genitive kʉ́pʉs, rather than e.g. Khulls-like kàpa~kapas.

Note the four-way contrasts between t~tʲ~č~ć, d~dʲ~ǯ~ǵ, and s~sʲ~š~ś. These were distinguished by tongue shape as well as place of articulation.

There were five vowels, /a e i o u/. In major syllables, all five vowels could occur. In minor syllables, only /a i/ could occur.

Proto-Subumpamese (1700) to Pudop (2672)

The consonant inventory was

Rounded bilabials:    pʷ  bʷ          w 
Bilabials:            p   b   m   f               
Alveolars:            t   d   n   s   l             
Postalveolars:        č   ǯ   ň   š   ł           
Palatals:             ć               y
Prevelars:            c̀        
Velars:               k   ġ   ŋ   x   g
Labiovelars:          kʷ  ġʷ      xʷ  gʷ
Uvulars:              q           h              
Rounded uvulars:      qʷ          hʷ 

This is the language spoken in the capital district, Pudop, named after its cranberry harvest.

1. The high central vowel ə changed to i unconditionally.
2. Syllable-final nasals ŋ ň changed to match the place of a following consonant, and changed to n if word-final.
3. the palatalized alveolar consonants č ǯ ň ł become plain alveolars s z n l. Then c̀ ć shifted to ś š.
4. Then, the stops k ġ shifted to ś y before any /e/ or /i/.
5. All remaining affricates change to fricatives: c ʒ > s z .
6. Labialization bleeds through clusters. e.g. kʷm > kʷmʷ. This means that it was no longer phonemic.
7. Then, voiceless stops and fricatives became voiced after a low tone or a long falling vowel. ś x h hʷ > y g Ø w .
8. The coda fricatives s š ś x all became voiced to Ø i i Ø. The silent ones lengthened a preceding vowel, and sequences such as /ii uu/ shifted to long vowels as well.
9. The voiced stops d ġ ġʷ shifted to r g gʷ. However, stop allophones remained in some positions.
10. Labialized consos in syllable final position become bilabials. Thus pʷ bʷ mʷ w > p b m w; kʷ ŋʷ > p m.
11. Palatalization also bleeds though. This is sort of a compensatory shift to make up for the last one.
12. The uvular stop q shifted to k.

Thus the final consonant inventory was

Labials:              p   m   f   w   b     
Alveolars:            t   n   s   l   r   z             
Postalveolars:                š                   
Palatals:                     ś   y        
Velars:               k   ŋ   x   g
Postvelars:                   h               

This was originally intended for a longer period; it might stop partway through.

Proto-Subumpamese (1700) to Eastern Subumpamese (2672)

1. gʷ hʷ > v f.
2. The high central vowel ə changed to i unconditionally.
3. Syllable-final ŋ ň changed to match the place of a following consonant, and changed to n if word-final.
4. pʷ bʷ mʷ w > p b m v. (Possibly /ə/ > /o/ when facing a labialized consonant before this shift.)
5. ai (on any tone) became ē (perhaps not always long).
6. The lateral approximant l shifted to w.
7. Palatals č ć ǯ ň ł > c c ʒ n l.
8. Velars (but not labiovelars) shifted doubly forward:

   c̀ k ġ ŋ x g > č č ǯ ň š ž. (Possibly velars remain in some positions, as in early Proto-Indo-European. This would best be explained as labialization.)

9. The uvular stop q shifted to k. /h/ became /x/ in most positions, but the spelling remained.
10. In syllable-final position, f c shifted to p t. (Thus /k/>/t č/, /h/>/s š/, even though the shifts were not related.)
11. The labiovelars kʷ ġʷ shifted to p b.
12. The fricative h shifted to k after a high tone.

Thus the Eastern Subumpamese consonant inventory was

Bilabials:       p   b   m   f   v   w       
Alveolars:       t   d   n   s       l   c   ʒ             
Palataloids:     č   ǯ   ň   š   ž   y                   
Velars:          k       ŋ   h

For FILTER, see Lenian languages and FILTER.

Later shifts:

Labiovelars occurred mostly in the vicinity of /i u/; each branch developed them in different ways:

1. shift to velars, which were almost in complementary distribution.
2. shift to rounded bilabials.
3. a split shift combining the above two options depending on further conditions.
4. retention, with vowel mergers creating new minimal pairs.

Gold

This branch is excluded on the basis that it loses its classifier prefixes. See Gold language.