User:Soap/Mevumep

Spoken in Mevumep. See also User:Soap/GV and User:Soap/Todrom.

Starting consonant inventory

Bilabials        p   m   b   
Coronals         t   n   d   l   r   s
Palataloids      č           y
Velars           k   ŋ   g           h

There may have been a marginal /x/ for earlier /hh/, but another possibility is that this was an allophone of /h/ and that plain /h/ had disappeared in medial position by this time.

There may have been labiovelars /kʷ gʷ w/, but these seem to behave like sequences of /ku gu u/ except in the GV clade.

Also, there may have been voiceless nasals /mʰ nʰ ŋʰ/, but so far all sound changes here rely on their having merged in the common ancestor.

NE Cluster Group I

THis is the more complicated of the two. The sound change list probably ends around 500 BCE which means that this is perhaps actually a slower-changing branch that stuck together longer rather than a more cmok,pikcated one.

1. The voiceless alveolar stops t d shifted to c ʒ (IPA /ts/).
2. Shifts involving fronting of consonants. These may be interleaved with others when the list fills out:
   1. The postalveolar affricates č ǯ shifted to s z, except possibly in some codas where the /č/ had already assimilated to the onset of the following syllable.
   2. Shifts of velars:
      1. The velars k g ŋ h shifted to č ǯ ň š before any /i/, including the sequence /Ci/ before a vowel.
      2. The velars k g ŋ h shifted to kʷ gʷ ŋʷ hʷ before any /u/. (The labials p b also shifted to pʷ bʷ.)
      3. Elsewhere, the plain velars k g ŋ h shifted to dentals t d ṇ ṣ. (The stops are spelled undotted since they had mostly disappeared earlier; it is unlikely that these will contrast with the marginal inherited stops.) Note that these never occurred before /i/ and possibly not before /u/.
   3. Finally, the labiovelars kʷ gʷ ŋʷ hʷ delabialized to k g ŋ h, becoming the only velars in the language.
3. The sequences əi əu shifted to ī ū.
4. The voiceless bilabial stops p pʷ lenited to fricatives f ɸ in many positions (unconditionally between vowels, possibly conditionally elsewhere).
5. Geminate stops (including many that were not described in this list) became single. This restores /p/ to phonemic status, since the geminates had always been voiceless (e.g. /gb/ [kp], probably shifting early to [pp] though unlisted here) and also had escaped the previous shift.
6. The labials pʲ bʲ shifted to t d. NOTE, there may need to be a new consonant deletion first, because the needed sequence /ia/ is rare in the proto-language.
7. The schwa vowel ə disappeared to Ø, creating clusters, except where rules disallowed a collapse, in which case it became i.
8. The long vowels ī ū shifted to i u, and the other long vowels may have also become short here.

NE Cluster Group II

1. The voiceless alveolar stops t d shifted to ss s. They may have been [s z] when alone, but in clusters they were always voiceless, so for example later on a cluster like /zd/ is in fact [st] even if both members are voiced on their own.
2. The postalveolar affricates č ǯ shifted to šš š, except possibly in some codas where the /č/ had already assimilated to the onset of the following syllable. These sounds do not immediately merge with either /s z/ or with /sʲ/ below (there is no voiced form), but most likely will merge eventually. In theory these could even have their own separate palatalized forms, making a four-way contrast of /s sʲ š šʲ/ as in Slavic.
3. The velars k g ŋ h shifted to kʷ gʷ ŋʷ hʷ before any /u/. (The labials p b also shifted to pʷ bʷ.)
4. The velars k g ŋ h shifted to tʲ rʲ nʲ sʲ. The last may be /ššʲ/, but the incoming /h/ is probably too weak to become a geminate.
5. The vowel i shifted to ʲi unconditionally, except that labialized consonants did not palatalize.
6. All consonants delabialized. Since new labialization is coming, these sounds may hang on after all, though they were very restricted in occurrence.
7. The sequences tʲa nʲa rʲa sʲa shifted to tʲe nʲe rʲe sʲe. It is likely that atʲ anʲ arʲ asʲ in the coda (but nowhere else) also shifted to forms with /e/.
8. The sequences tʲu nʲu rʲu sʲu shifted to tʲi nʲi rʲi sʲi. This was actually moving through schwa, while retaining palatalization, and therefore ending up the same as if they had been with /i/.
9. The sequences utʲ unʲ urʲ usʲ in the coda shifted to forms with schwa.
10. The vowel ə became i. The diphthongs əi əu shifted to ī ū (without palatalization).
11. In at least some positions, the voiced stops b g most likely lenited to w Ø as in Birch (though remember the timing is off by thousands of years).
12. Remaining voiced stops b g devoiced to p k.
13. The clusters sp tʲp rp rʲp np nʲp lp shifted to sw tw w w mp mp w.
14. The clusters sk tʲk rk rʲk nk nʲk lk shifted to x tt rr rrʲ ŋk nʲtʲ ll. However, note that these clusters which could theoretically provide new /t/ were likely assimilated before they could form (since e.g. /tʲk/ would need to have come from /kkʷ/), and may not occur after all.

    These sound changes were originally intended to apply to voiced sounds only.

    Diachronically there should never have been any coda /š/ but Birch developed it through analogy. If it exists it would follow the pattern of /s/.

15. It is possible that labialized consonants create a new vowel o in the same manner as earlier /e/.
16. Since the palatalized alveolars tʲ nʲ rʲ sʲ were only distinct before /a/, they shifted to c n r s. Note that there was still a phonemic /š ž/ untouched by this shift, though they could not occur in the coda.

    Note that it is not certain that these sounds actually depalatalized. It could be that /e i/ really phonemically equivalent to /ja ju/.

17. The clusters cl cr shifted to tt.
18. Coda c s assimilated to (that is, geminated) a following consonant, with the /s/ coda also lengthening the preceding vowel, and then disappeared to Ø.
19. Palatalized l h ɣ became y šʲ žʲ, and it may be that the palatalization is no longer phonemic on the postalveolars anyhow.
20. Any suirviving used of the voiced approximants β ɣ hardened and devoiced into stops p k. (They may have been allophonically bvoiced, but this was no longer phojnemic.)
21. Word-initial r became t.
22. Coda l lengthened a preceding vowel and then disappeared to Ø.
23. Coda ir ur > e o, disappearing elsewhere, making /n/ and gemination the only codas.

It may be that all obstruents devoice, with /s š/ and the "re-devoiced" stops becoming geminates, except in clusters. Otherwise, there will still not be any word-initial /t/.

Thus the consonant inventory aounrd 1000 BCE was

Labials                 p   m   w    
Alveolars               t   n   l   r   s    
Palataloids                     y       š   
Velars                  k   ŋ           h

The four vowels were /a e i u/. The allowable final consonants were /r n/.