Primordial scratchpad

this is the ancestor of almost ALL Languages of Teppala. Despite its name, it still is not the ancestor of literally all languages, because there are scattered aboriginal groups who settled the mainland ~50000BC and kept their languages as other groups took away their land. But by 4000 AD, even these surviving groups had come to be politically dominated by the descendants of the migrants who spoke languages derived from Primordial.

Scope

This language was spoken sometime between 14000 BC and 11000 BC. The "pre-primordial" language is only present to give the sound changes needed to make the asymmetrical inventory of Primordial.

Possible early non-MRCA phonology and sound changes

13:21, 16 December 2024 (PST)

This language is proto-Repilian, and also identical with paleo-Zenith and paleo-Andanese, which split apart later.

Note that /mbʷ nd ŋġʷ/ are from earlier /mʷ nʷ ŋʷ/. In theory, this shift could be pushed forward so it does not occur in this branch, which would mean that there are no prenasals in this branch. Adding these back in would actually ease the analysis of the consonant inventory since there would then be lab/plain/pal sets for all of /m n ŋ l r t k s h/ and perhaps more.

Thus a minimal consonant inventory, where the shift above did not happen, could be

GLIDING     m  n  ŋ  l  r  t  k  s  h
FIXED       x  xʷ y  w  B (d)(p)(pʷ)

This can also be analyzed as

GLIDING     m  n  ŋ  l  r  t  k  s  h  x  Ø 
FIXED       B  d (y)(w)  

If assuming there was only one /d/ and no /p/, and that xʲ=hʲ simply merging.

Śince the loss of /p/ was the last shift in the previous list, it's possible that it didnt happen here either. This would mean, however, that there are far fewer vowel sequences.

Another possibility is that Primordial had all three of /b d ġ/, and all three disappeared in the MRCA branch, with /d/ going to /r/, perhaps quite late, which in turn shifted mostly to /j/, whereas /b ġ/ would have disappeared earlier.

If /p/ is kept, it may be that /d/ was a normal consonant too, which means that the gliding/fixed distinction is meaningless and the consonant inventory is

    m  n  ŋ  l  r  t  k  s  h  x  p  d

And /w j/ are just allophones of the vowels if those vowels do exist.

Bilabials:      p   b   m   w
Alveolars:      t   d   n   l   r   s
Palatals:                   y
Velars:         k   ġ   ŋ  (Ø)      x
Glottals:                           h

Other ideas relating to consonants

However it may be possible that even in the coda, there was still some contrastive labialization, as at least /kʷ/ existed in the parent language. Still, the long vowel codas like /ūt/ that appear in pre-MRCA can come from open syllables plus a consonant and a lost final /ĭ/.

It is also possible that the coarticulation shift didnt happen in this branch, meaning that they are all just vowel sequences like Play's, but yet that /mʷ ŋʷ kʷ/ still did exist and probably occurred in the coda, because those were in the phonology even three steps back.

Vowels

The vowels might be /a i u ə/ after all, with MRCA's later shift to /a ɜ ɨ/ explained by the simple collapse of the two high vowels and encouraged by the growth of labial and palatal coarts. This four-vowel inventory would thus need to be back-dated into Primordial, where it might help explain why /w j/ were so common (since they could just be allophones of / u i/).

Tones

If tones existed, they must be lost in the MRCA branch fairly early on.

Paleo-Andanic sound changes

empty for now

Proto-Repilian sound changes

To be filled in later, but see Repilian languages for ideas.

Northeast languages sound changes

It is possible that the northeastern aboriginals are even older than the Repilians, in which case they would have their own language.

This branch might be called PC, because it's both "P" (kʷ > p) and "C" (k(ʲ) > č) leaving few dorsal consonants. It also most likely spreads across the cold climates before being displaced by Repilian and then by Moonshine.

Phonology

13:13, 16 December 2024 (PST)

The consonant inventory is unclear; it depends on what stage of the language is being reconstructed. The Languages of Teppala page as of Apr 25 2020 lists the vowel inventory as just /a ɜ ɨ/, but it should probably be something like /a ɜ ɨ ā ē ī ō ū/ or at least /a ai au ɜ ɜi ɜu ɨ ɨi ɨu/, though the last two can be spelled as ī ū for clarity.

Another possible vowel inventory is

SHORT VOWELS

a   ɜ   ɨ

LONG VOWELS AND FALLING DIPHTHONGS

ā  ai  au
ɜ̄  ɜi  ɜu
ɨ̄   ī   ū

There may have been an unlisted shift like /ɨ/ > /Ø/ in prehistory, which led to syllables ending in stops. Another possible idea is that pre-Primordial ("Mapi") may have had /a ɜ i u/.

Some consonants in the inventory could be arbitrarily duplicated, since this is only the minimal inventory needed to get to Tapilula and many mergers could have takne place.

By the time of Silatibarra (8000 BC), coda /i/ had probably disappeared except in bare form, so the inventory would have been something like /ā au ɜ̄ ɜu ī ū/. this suggests that perhaps ɜ̄ should shift to /əi/.

A possible consonant inventory (taken from Repilian languages 13:08, 16 December 2024 (PST) ) is

Rounded labials:                w   hʷ mbʷ     
Plain bilabials:            m               B
Spread bilabials:           mʲ
Rounded alveolars:      tʷ      lʷ  sʷ      rʷ         
Alveolars:              t   n   l   s  nd   r
Postalveolars:          č   ň   ł   š       ř
Palatals:               ć   ń   y   ś
Velars:                 k   ŋ  (Ø)  x
Labiovelars:            kʷ          xʷ ŋġʷ
Postvelars:                         h
  

Note that /ś/ is from /xʲ/, not a retention of the original.

It is possible that a /p/ and perhaps /pʷ/ will need to be added to fix problems with Tapilula/MRCA.

Another possible addition is a plain voiced stop /d/, separate from /r/ but quickly merging with it in the Tapilula branch.

It is possible that all of the labialized and palatalized consonants can be analyzed as clusters, if they never occur in the coda, even if /mw/ onset contrasts with /m.w/.

Older information

Source of infixes

Jun 21, 2020 (updated 07:11, 11 March 2022 (PST))

The following infixes must be present by the time of Tapilula, but could be confined to just the ATG ("Andano-Tropical-Gold") or MAP (Middlesex-Andanese-Play) branch if retention of infixes is influenced by substrate languages on the mainland whereas on the islands they die out:

vowel-only

1. -i- past tense marker; evolves to an essive/partitive case in Gold since it is used on nouns and verbs alike
2. -u- plural marker (NOT future tense; this was a very late post-Play innovation, not even used in classical Play)

schwa-dependent

1. -ək- "self"; reflexive verb marker
2. -əh- "observer". usage difficult to explain; important in Lava Bed languages but evolves to a genitive when LB morphology disappears
3. -ə(g)- "patient"; evolves to a locative in Gold and early Play, but had already merged with /Ø/ in most contexts in Gold despite maintaining its separate meaning
4. -ə- "agent"; this may have simply been /Ø/ originally, but a schwa was deposited analogically in both Andanese and Play. This could easily have been done separately in the two branches since they both did /ə/ > /Ø/ separately as well. But it is also possible that the ə~Ø allomorphy is ancient and not based on diachronics. This is listed here because it patterns like the others rather than resembling the vowel-only /i/ and /u/ infixes

Obsolete infiexes: These cannot survive in bare form through to Tapilula:

1. -i- accusative case marker. Goes into the classifier prefix, not the root. It is most likely that this was originally an ɨ (barred i), not the front vowel i, which means it can surface as labialization.

This was originally intended to be a separate morpheme from the /-ɨ/ suffix that means the exact same thing, but it now seems that it is more parsimonious to unite them.

Scraping phonology

Jun 21, 2020

It seems that /p/ and all of the labialzied consonants of Tapilula must come from sequences rather than direct inheritance. This means that very few lexical roots can be the same number of syllables in Tapilula and the MRCA ... essentially, only those with the few "pure" consonants like /m n l k/ and even then only when the vowels are right.

Ive been rolling dice etc to see which Tap words trace back to atomic roots in Primordial, and then obeying that only when possible. so e.g. Tap /nègi/ "word, sound" is one of the roots that I rolled atomic for, and it could come from a Prim sequence such as /nɜ$ɨ/, but only if I can find a way to explain the retention of the final vowel, when up until now I've been deleting all short final /ɨ/. (If it is long, it would survive, but then the root would be trimoraic, and Im not sure I want to allow atomic roots of that structure.)

Lost grammar

If an accusative suffix -ɨ exists, it could stay as such after word-final consonants but mutate when occurring after a vowel. for example, /nɨtɨ/ "duck" > /nɨsɨ/.

This would be lost in Tapilula but possibly retaiend as -i in Repilian and Zenith. Tapilula has an accusative infix /-i-/, which is probably a coincidence as there is no easy way to derive it from the suffix.

It is not clear if this is originally from an earlier /ɜ/ (as expected) or from an /ɨ/ that survived because it was padded.

Syllable structure

There was never final /p/ at any stage, but there was final /kʷ/, which goes to /p/ unconditionally in Tapilula. But this means that sequences like /ip/ cannot be inherited.

NOTE, kʷ > p might have occurred much earlier than originally planned, which would mean that there *would* be final -p in pre-Tapilula, but importantly, it would not be labialized, so if it affects vowels it must have done so at an early stage.

Short words

Jun 17, 2020

Ultrashort words like ù "flowerbud" will trace back to Primordial roots often of two syllables. e.g. the flowebud word could have come from such as /yūkʷɨ/.

Other ideas:

• Tap wŏho "bowl, cup, dish" < /wa$ɜn/. Not cognate to /wò/ "water" because the morpheme order is wrong and the accent shows that it cannot have been a later coinage.

• Note, though, that looking even further back, the morpheme order may have been head-final, meaning it would have cycled three times in some branches, from HF > HI > HF. This is 15,000 years of history, so well in reach of plausibility.

• Tap pù "tree" < /kʷɨt(V)/. This is an old idea which may need to be reworked, however. Note that the current list of sound changes has /kʷ/ > /pʷ/, not to a plain /p/.
• Tap gà "angel" < possibly /hā/. This word never had a final consonant, but it may have come from a form like /GaBa/ far back in the past.
• Tap mfʷù "mountain" could have come from something such as /muCɨn/, ... note that there is no primordial /mʷ/ phoneme. This is a prpblem. Also, this might collapse into a syllabic nasal.
• Tap yò "hand" is problematic too, as it can only have come from something like /rūan/, meaning that a vowel sequence survived for thousands of years unreduced.

• Though this only if I respect my original plan to have this word come from an earlier phonetically impossible /yʷo/. Getting /yò/ without passing through a labialized stage is much easier, and requires neither a final consonant nor a long+short vowel hiatus.

• Tap gĭku "seagull" < /hɨkkɨŋʷ/, but note that this requires that /ɨŋʷ/ did not shift to a syllabic nasal the way /ɨm ɨn ɨŋ/ all did. this can only be true if it was really /ūŋʷ/.
• Tap mbù "religious sacrifice" < /mbūŋʷ/. Never had mbʷ.
• Tap mbù (used in the word for supermodel) < /mbūkʷ/. Probably *did* have /mbʷ/.
• Tap gàyo "zewan" (a cultural concept) < something like /Gai/ + /wan/.
• Tap mà ~ fà "back of body" < /ma(m)bɨ/. May or may not have had a vfinal vowel.,
• Tap ăta "salmon, trout" < /yam tan/, no final vowels.
• Tap nì "duck" < /nɨtɨ/
• Tap păndo "boat" < /CuCan nakʷɨ/. /CuCan/ was maybe already "boat" by itself in Prim, but remember the classifiers .... it could just be /pà/ "greet" + /nò/ "to want to; to move forward".
• Tap ḳàu "to drink" < /tɨra/ + something that evolved into /(g)ù/. No final C on first word.
• Tap hùhi "bridge" < /$ū$ī/ if this is a possible root structure. /$ū$ɨ/ if a means can be found in which the final -ɨ does not disappear.

other sdgdsfgdfg

Apr 26, 2020

• The Repilian branch of the family starts evolving into something like Australian Aboriginal languages (but also sort of like Eskimo-Aleut), meaning that it starts out becoming less like Owl, and Owl-like traits evolve only later on.
• Possibly nu ~ nd- for "what" as in "book" ---> "book which...". But need to first figure out whether there is an independent /u/. If not, then this is whatever /u/ evolves into, possibly just ɨ .

Apr 25, 2020

• Note that Mumba is ultimately cognate to Saffaslujje and that the city could legitimately just be called Mumba or something like Omumba.

• IF the 2P pronoun in Silatibarra is /hət/, it could have begun with any of /s š ś h/ in Primordial, and possibly also /x/. Likewise the final consonant could have been /č/ instead, which probably functions as /tʲ/.

• The prefixes may correspond at least in part to Repilian_languages/Owl#Slot_B, which could have come from prefixes attached to the second noun in a phrase. Thus, at least some of the prefixes in Primordial might be gender markers, which would be not far off from the noun classifiers that arose in Tapilula. One possibility for a male prefix is təl-, with təlʲ- being a lower probability.

• Its possible that the original vowel inventory was overloaded, and that many original vowels could collapse into schwa. for example, the male prefix might have originally been tə̄l.

• The /-i-/ infix that marks the accusative case in Tapilula may have been already an infix in Primordial, but it is more likely that such a long-lasting phenomenon arose from a more robust form.

Grammar

There were no classifiers.

There were infixes, which survived in both Repilian languages and Tapilula for >10000 years of changes (though evolving into greatly different final forms in the two branches). There may have been prefixes, which disappeared in Tapilula because Tapilula needed the prefix slot for its classifiers. Thus, a whole slew of prefixes could be made for Primordial with absolutely no need to synchronize them with Tapilula.

THe pronouns na "I" and ge "you", whatever their original forms, may well go back all the way to Primordial. In Mumba ~8000 BC their forms would probably have been /na/ and /hə/, though perhaps there were final consonants that disappeared. e.g. the words could have been /nam/ and /hət/.

If the accusative marker is cognate to Tapilula's, it is likely to have been /ī/ (that is, /ɨj/) around 11,000 BC. /ɜi/ would also work.

It is not clear whether Tapilula's genitive infix /-əh-/ could have arisen from a standalone CV or CVC form in Primordial; at some point Tapilula evolved the practice of deriving infixes from CV prefixes spelled backwards, and therefore the original form may have been a standalone morpheme that would have evolved into /hə/ had it survived. (Although its possible Andanese did preserve the standalone morpheme, I have been assuming backwards analogy.) If the genitive derives from a CV standalone form, it can only have been /čɜ/ or that plus a consonant. It could perhaps be /čɜr/, or /čɜu/ if an analogy of /o/ > /ə/ in Tapilula is assumed. This may have been a verb in 11,000 BC, and the genitive could have been marked in an entirely different way in Primordial, but it is still a direct cognate.

the true genitive is that which would have led to hàhə in Tapilula. This may mean that it was /ča/, an idea only slightly different than before.

Culture

Four different racial types all derive their languages from Primordial, and all four of them are the aboriginals of the land they inherit. Yet this language was spoken only around 14000~11000 BC, so it must have been a trade language of some sort; three of the four races were well-formed by this date.

Racially, the Zeniths may be a branch of the paleo-Andanese, and the Repilians a branch of the Zenith. The dark-skinned Pejo people would have remained on the islands, and it is possible that they did not speak Primordial and only inherited its descendant, Tapilula. The same could also be true of the Mumba people, but note that the Mumba were only a subgroup of a wider Lenian race that included many highlanders.

Lagi languages

NOTE: It is not certain that this is actually the aboriginal Lenian grouping. They may have only borrowed these languages, and if that is the case, an entirely unrelated family must be created for the true aboriginal Lenians.

These are the languages of the proto-Lenian uplands. In the core of the area, people were short in stature and had extreemely light skin and always blonde hair and blue eyes. Also, they did not have the Andanic immunity to sunburn, so they commonly suffered sunburn even though their climate was cloudier than that of the lowlands.

They were the source of the wider Lenian (Mumba) phenotype but did not have all of the same characteristics. They were also more diverse, as they had areas near their borders (but still within Lagi territory) where people had traits of the Pejo, the paleo-Andanese, etc, and possibly some other groups that never got off the islands.

These people did not refer to themselvbes as Mumba or even with a term cognate to Mumba.

The consonant inventory is similar to

Rounded labials:      pʷ        mbʷ  hʷ   w
Plain bilabials:      p    m
Spread bilabials:     pʲ   mʲ
Rounded alveolars:    tʷ   nʷ        sʷ   lʷ   rʷ
Alveolars:            t    n    nd   s    l    r
Postalveolars:        č    ň         š    ł    ř
Palatals:             ć    ń         ś  
Velars:               k    ŋ         x
Labiovelars:          kʷ        ŋġʷ  xʷ
Glottals:                            h

but probably not exact. The vowels may be

SHORT VOWELS

a   ɜ   ɨ

LONG VOWELS AND FALLING DIPHTHONGS

ā  ai  au
ɜ̄  ɜi  ɜu
ɨ̄   ī   ū

Primordial to proto-Lagi

1. The sequences mbʷ mʲ ŋġʷ shifted to bʷ b bʷ.
2. The lateral lʷ shifted to gʷ.
3. The sequences pʲ tʷ shifted to ps kt. (/ps/ might later become /ts/)