Sakhi

Sakhi is a language spoken by feminists who had formerly been part of Thaoa. They had previously been a very aggressive military empire, at times the most aggressive on the planet, but when the Thaoan women realized that the tribes their soldiers were attacking were ruled entirely by women, they blocked their army and forced the soldiers back home. The women then signed a peace treaty called the Feminist Compact with their enemies, and abolished their own military.

Phonology

Diachronics

Sakhi was a northern dialect of Thaoa which early on gained vowel length and lost distinctive aspiration. Most long vowels arose from vowels that had previously been followed by a glottal stop. Thus, the glottal stop was eliminated.

Sakhi's sound changes were complex and often polyconditional, unlike those of Palli, where the general trend was strongly in favor of a smaller phonology and more open syllables.

early changes

The phonology of classical Thaoa had been (consonants only)

l  j  h  k  kʰ ŋ  p  pʰ m  t  tʰ  n  s  x  š  b  ž  č  ň  ʔ  q  qʰ

The uvular stops /q qʰ/ were marginal phonemes found only in loans from Old Andanese. Andanese itself had lost the uvulars in the meantime, but they persisted in a small number of words that had remained in Thaoa and were passed down to Thaoa's daughter languages of Sakhi and Palli.

Unlike Palli, the dialect that became Sakhi reflected Thaoa's early /v/ phoneme as b, and it therefore was not included in the later sound change that devoiced all fricatives. Thus proto-Sakhi /b/ usually corresponded to proto-Palli /f/ (later /h/).

Next, all stops became voiced after long vowels, even if aspirated. Nasals also switched.

Then, Sakhi soon gained a new /f/ phoneme, because it changed all of its aspirated stops to fricatives. The postalveolar affricate č behaved as aspirated in this shift, and therefore merged with š.

Next another wave of voicing happened, as all unaspirated stops became voiced, except if following an aspirated stop or a fricative, or syllable-initially with an /h/ located across a vowel. Thus, voiceless stops had almost disappeared from the language. However, the voiced stops were on their way out as well, because the next sound change was to weaken the voiced stops into fricatives and approximants. /d/ became /r/, however, not /z/. Some of the above changes excluded clusters; e.g. /nd/ did not become /nr/, and what remained came to be spelled /nt/.

At this point the language had very few words beginning with stops, and stops were not particularly common elsewhere either.

The vowel system was still entirely intact, except for the introduction of long bowels.

creation of hiatus

However, some balance was regained when all /h/ disappeared from the language. /x/, the reflex of /kʰ/, soon became /h/.

The deletion of /h/ made this already rapidly changing language even more unstable, because for the first time in thousands of years vowels were allowed to touch each other. These new vowel sequences were spelled with the old letter for /h/, since the letter for /x/ was used for the proper /h/ sound. However, to prevent confusion, the hiatus separator is Romanized instead with /ʕ/. The Sakhis needed this letter because these vowel sequences still contrasted with diphthongs, which was unusual for their area.

monophthongization and other vowel changes

A minor change to the vowel system happened here: the diphthongs /ai au/ shifted to /ei ou/, which was more distinct from the contrasting /ae ao/. Also, /ea ia/ > /eə iə/ (spelled ey iy). Note that /ia/ had always been a falling dip, not like /ja/. This was the same as Poswa's /ia/, although this sound had not yet become common in Poswa. Unstressed /a/ began to move towards schwa as well.

At this point, a large number of vowel changes occurred. The newly formed diphthongs /ei ou eə iə/ shifted to /ē ō ē ī/ (note that diphthongs had always been about the same length as long vowels, so this change was very minor). Previously, long vowels had almost always preceded voiced fricatives or approximants, but with this change three of them came to be more evenly spread.

The only remaining diphthong in the language, /oi/, now shifted to /uj/ and began to approach [ü:] (the symbol for IPA /y/ is not used because y meant schwa in Thaoa and older Sakhi, and still means /j/ in Sakhi in general). <---however this will likely go to ī anyway. <---- NOTE, this diphthong might instead switch from falling to rising and become /wi/.

Now all of the vowel sequences became diphthongs, including, for the first time, rising diphthongs beginning with /w/ and /j/. Thus phonemic /w/ reappeared. Any vowel squence beginning with /e i/ + any dissimilar vowel became a dipthong beginning with /j/. /o u/ + dissimilar vowel > diphthong with /w/. Note, however, that /ei/ > /ē/ and /ou/ > /ō/. This shift might have actually begun before the elimination of the primordial diphthoings, since these new diphthongs were mostly rising dipthongs and would have been distinct from the primordial diphthongs all along. Thus this /ei ou/ > /ē ō/ shift could have happened only once instead of twice.

However, /ii uu/ indeed shifted to /ji wu/, rather than becoming long vowels. Any vowel + schwa shifted to a long version of the first vowel, which means that /iə uə/ > /ī ū/ and this shift is probably also the same as the one above.

All nasal+stop clusters became plain voiced stops. Doubled stops (which were rare) became single.

later changes

the dorsal nasals /ŋ ñ/ changed to /ġ ǯ/ unconditionally, except in clusters or at the end of a word, where they became /n/. <--------- NOTE: this is on the list of snd changes from an earlier draft that ended in a significantrly different outcome. now, these would be the only voiced stops in the language.

all clusters became homorganic except possibly /k/+stop.

From this point on, sound changes slowed down, and became simpler and less conditional.

h > 0. This shift thus happened twice, thousands of years apart. (on the blue and red list, it is incorrectly omitted the first time, perhaps because i just confused it with deaspiration). note that this h > 0 essentially implies primordial kʰ > x > h > 0.

f>h, but not in clusters. (? sounds perverse)

possibly ž > 0. note that there was never a /z/, so there would be no new /ž/ arising from /zj/.

Final schwa died. This change was contemporary with a similar change in Poswa. Shortly afterwards, all other schwas disappeared too, becoming /a/ if they could not. [ə] came to be an allophone of /a/.<___NOTE: this was written before i had bumped Thaoa back from ~4500 to 2668 (3700 for end of classical stage), so it might not be contempotary after all.

Possible dialect splits, where poswob areas do tw>pw and khulls do tw>kw.

final phonology

Note, voiced fricatives are missing

TARGET PHONOLOGY IS

l  j  h  k  ġ  b  p  m  t  d   r  n  s  h  š  ž  ň   + PALATALS AND LABIALS.  /f/ may or may not be present.

(this letter order is not canonical)

Note that, from the previous stage of the language where stops had almost completely disappeared, the only change that created more stops was a very minor one that only worked intervocalically. instead, balance was restored by deleting vowels and fricatives, including twice /h/ > /0/. Even so, the language still has very little /p/ & /k/.

it is even possible that /t/ is more common than all five of /p b d k ġ/ combined.

Tests of words in sound changes

Assumes no semantic change. all unsuspecting off, schwa is spelled /ə/.

• hulyt ----> ult "bear:
• sahu ---> sau "eyelash"
• beanna ---> wēna "bird"
• thapakhuŋa ---> sapauga (placename)
• luʔna ---> lūra "moon"