Ogili II
Ogili II is a proposed reworking of Ogili to have a shorter list of sound changes that produce nearly the same result, thus enabling the language to more believably be moved further back in time so that its speakers can have intimate contact with Poswobs and Moonshines when those empires were still young.
The language will here below be referred to simply as Ogili.
Leaper (4700) to Ogili II (6843)
It may be that the language diverged from an earlier state, especially if this is spoken in the tropics. However, the sound changes below assume that the language is still spoken on the northwest coast, near Dreamer territory. It may have also shared sound changes with Ghost, but is all in all very different from Ghost.
- The ejectives ṗ ṭ ḳ shifted to voiced b r ġ. Note that /ġ/ was a true voiced stop, even though the coronal and possibly the labial were approximants.
- The voiced sounds r d merged as r.
- The vowels e i, on all tones, shifted to ʲa ʲi. This effectively created a vertical vowel system, since the inventory was /a ʲa o ʲi i/, but the old spellings were still used for the meantime.
- The inherited vowel o came to be spelled ɜ.
- Labialization defeated palatalization.
- In initial position or after a high tone, the fricatives s š ž x g xʷ gʷ shifted to c č ǯ k ġ kʷ ġʷ. The glottal consonants did not shift, and the shift quickly reversed itself in final position (and did so nearly losslessly, because the target sounds rarely occurred in that position after high tones).
- The voiced fricatives ž g gʷ ʕ ʕʷ shifted to Ø Ø w Ø w. (the "oo wow" shift.)
- The velar fricatives x xʷ shifted to h hʷ.
- All labialized consonants shifted to pure bilabials. Thus pʷ bʷ kʷ ġʷ hʷ shifted to p b p b f.
- In word-final position, the lateral approximant l disappeared to Ø.
- In word-final position, the fricatives f s š h merged into h.
- It is not clear how /š/ and /h/ could have occurred in this position.
- Consonants before (or after?) a syllabic nasal were deleted.
- Tone shifts took place in tandem:
- The long high tones á ā merged as the high tone á, except before a coda of either of /l s/ (H BW), in which case they rose further to the extra-high tone a̋.
- The short, post-glottalized high tone à shifted to á unconditionally. Note that this is only the orthographic high tone, and not the allophone of the long tones that sounded identical to it.
- Short mid tones ă before a voiceless coda rose to the high tone á.
- Short mid tones ă before a plain /w/ coda shifted to the low-toned ʷà. Note that the coda /l/ had not yet shifted, so this change does not affect syllables that had ended in /l/.
- Any remaining short mid tones ă shifted to ā. Here, the macron denotes pitch, not length. Also note that this covers all primordially closed syllables except those with historically long vowels. That is, /tan/ had a surface-level low tone, even though it was underlyingly high in Gold.
- The long, low, pharyngealized tone â shifted to the extra-low tone ȁ, except before coda /-s/ (H), where it appeared as the ordinary low tone à.
- All unstressed vowels became either ā or à, depending on sandhi.
- Syllable-final p b l shifted to pʷ bʷ w. The new /w/ can be spelled as /lʷ/ temporarily for precision.
- Coda labialization skipped across a preceding vowel to labialize the consonant in the onset. The syllables could also be analyzed with structures like /twa/, since the inherited labialization had disappeared. Note that in some clusters, the following consonant was also labialized.
- The voiced velar stop ġ became a voiced fricative g. Note that the labialized version had become /b/ by this time.
- The long high tone á: merged with the extra-high tone a̋.
- Before a coda that was any of /m n ŋ lʷ bʷ/, the mid vowel sequences ɜ ʷɜ shifted to ʷe o.
- In theory, these could be switched, but the assumption is that both were inherently rounded at least to start with.
- The vowel sequences ɜ ʷa ʷɜ shifted to e o u. The /u/ was always rounded and had no unrounded counterpart because the rare inherited bare [u] had become /i/.
- In a few clusters like /lʷg/, the coda consonant may have become the onset of the following syllable. (But remember that they were mostly labialized.)
- All remaining codas were deleted.
- Phonology
All vowels occurring before a labialized consonant became back rounded vowels, and those that did not were pushed into front unrounded vowels. Thus the inventory of /a ɜ ɨ/ expanded to /a e i/ (front) and /ɔ o u/ (back). This was similar to Poswa, but not to Khulls: in Khulls, vowels had been colored by vowels that occurred before them, but not after.
Tones
The tone setup travels from Khulls fairly intact. There are at least five tones, "believed to be" register (flat) tones, and one neutral tone, which takes sandhi from the others:
- a̋ is descended from the Khulls á and ā tones. It is the highest tone.
- á is descended from the Khulls à tone. In formerly closed syllables, perhaps also from the á and ā tones.
- ā is descended from the Khulls ă tone. It is a medium tone with no vowel coloring.
- à is descended from the Khulls â tone and in some syllables also from the unstressed vowels.
- ȁ is descended from ????
- a, the neutral tone, is descended from unstressed vowels.
Note: It might make sense to merge even more, since the à tone generally does not occur before a non-labialized final consonant in Khulls (it is etymologically à, but due to sandhi becomes ă), and therefore there will be almost no rounded vowels on the á tone in Ogili if this system is used as given.
Must make split versions of the tones, depending on whether they end in a final consonant or not, and perhaps also varying on what that consonant is. The number of allowable finals is down to just four, /p b w h/, by the end of the sound change list above. /p b/ are rare in non-labialized form, so they might simply switch to labialized and then disappear. /l/ might also switch over, despite up above saying that it did not.
NOTE: it is perhaps better to analyze the final consonants as /p b h/ which can either carry or not carry an additional labializing influence (thus making h into /f/).
á and ā were rare in Khulls, but occurred with both plain and labialized finals, and thus do not need to be repaired as they merge in Ogili as a̋ (but see below).
à does not occur before nonlabialized finals except as an allophone of ā and á. If this allophone persists into Ogili, then the Ogili a̋ tone will need repair.
ă occurs with both types of finals in Khulls and thus does not need repair when it appears as ā in Ogili.
â occurs with both types of finals in Khulls and thus does not need repair when it appears as à in Ogili.
Thus paths leading to ȁ and a̋ are needed. Each tone could sprout a modified version when appearing with a final consonant, which then disappears. á could become a̋ before final -p, although since this is a labialzing tone, this wouldnt help. final -h could be added to the list, since it is also voiceless.
Note that in some words a syllable-final consonant will be preserved, since the next syllable will have become vowel-initial by another rule. In these cases, the consonant still has its intended affect.
Unlike the first draft of Ogili, there are probably no new vowels resulting from elision of intervocalic consonants. Thus the number of syllables in any given word is nearly always the same as it was in Khulls.
new idea for tones etc
a reworking of the above would envision the possible syllable codas at the end of the three-vowel stage as consising of three binary elements [b] [h] [w]. That is, e.g. coda /b/ is +b -h -w, coda /pʷ/ is +b +h +w, coda /hʷ/ is -b +h +w, and so on. Then three shifts would apply:
LOSS OF FINAL CONSDONANTS
First, all vowels before [w] become rounded. Thus /a ɜ ɨ/ > /ɔ o y/ (sic). There are no diphthongs because a three vowel system with diphthongs and a mostly CV syllable setup is unstable.
Then, all vowels *not* before [h] become long. (This could be problematic, since long vowels would outnumber short vowels, but vowel length can be reduced in a secondary shift, probably by shifting long vowels to the tones with gaps in their distribution.)
Next, all vowels *not* before [b] undergo a secondary shift such that /a ɜ ɨ ɔ o y/ > /a e i o u y/. (That is, some vowels are unaffected.) Possibly allow /æ/, but this language family has a strong tradition of resisting any sound changes removing [a].
Then, /ɜ/ > /ø/ unconditionally to round out the vowel space, leacing the language with a total vowel inventory of /a e i ɨ ɔ ø y o u/.
Note that in this setup, the vowels /ɨ ɔ ø/ are restricted to syllables that had previously ended in [b], and of these, /ɔ ø/ are restricted to syllables that had previously ended in [b]-[w] (that is, /b/ without /w/). This is problematic because there is no voiceless version of this coda other than primordial final /p/, which was rare.
tomne shift history
this is the same as the above chart, but from the opposite direction. "primordial" means Khulls or even pre0Khulls sandhi is responsible for the shift, not what the syllable looked like later. i.e. the distinction was not phonemic in Khulls but rather part of sandhi, but became phonemic shortly after the split.
Khulls á & ā > Ogili [á:] in primordial open syllables, but Ogili [á] in primordial closed syllables.
Khulls à > Ogili [á] in primordial open syllables, but Ogili [ă] in primordial closed syllables.
Khulls ă > Ogili [ă] in all environments.
Khulls a > Ogili [ă] or [à] depending on sandhi. (This is the Khulls neutral tone.)
Khulls â > Ogili [à:] in all environments.
Later, [á:] > [a̋], and [à:] > [ȁ], but for ease of typing both transcriptions remain valid.
Scrub area
Make aw>wa,etc... labialization skips backward to preserve CV.
Primordial open syllables can actually end in a coda /w/ due to the desyllabification of short /u/. Thus syllables ending in /w/ have a wider variety of tones than those that do not. This problem must be solved.
0bhw has eight forms:
b__, _h_, __w, bh_, b_w, _hw, bhw b h w p bʷ f pʷ d h w t b f p
Note that the above shifts do NOT lead to a clean and balanced five-tone setup because primordial closed syllables prohibited the later presence of the rounding coda element [w], which means that the topmost tone will only ever occur with unrounded vowels unless it is fed by a secondary later shift. Additionally, the [à] tone will only occur in polysyllabic words since it arose solely from unstressed neutral tones. This, too, must be fed by a later shift.
Another minor problem is that coda [b] almost always co-occurs with coda [w], so this system is not as clean as it appears. Since coda [b] is nearly in complementary distribution with the primordial closed syllables, perhaps they could cover each other: that is, primordial closed syllables all come to end in [b]. Final /n/ could be [b] since it lacks a labialized counterpart. Final /s/ must remain as /h/ however.
The language at this time was entirely CV, unless the syllabic nasals hang on and become nonsyllabic. this changes only in the daughter languages.