The Gòranú language is a language descended from Khulls currently taking its name from discarded words that mean raspberry wine. That is, it is the next in a series of fruit-flavored language names that includes Apple Pie, Mandarin Orange, and Strawberry Icecream. Of all of the fruity languages, Goranu is the only one that is part of the canonical world of planet Teppala: the others are all meant to resemble Earth languages, or are private experiments that would not make sense even in a fictional world like Teppala.
note on name: this name is probably a mistake, actually, because /nʷ/ > /l/, not /n/.
This will be yet another attempt from me to create a language with a true vertical vowel system, spelled /a ə i/ for convenience but better understood as /a ɜ ɨ/. I had previously thought that to make a language like this I needed to have contrastive labialization on most or all of my consonants, but wikipedia:Abkhaz has a vertical vowel system with its contrastive labialization confined almost entirely to stops.
Contrast with other Khul langyages
Most Khul languages radically changed Khulls' highly unstable phonology; in fact, the early sound changes in most Khul languages made the phoneme inventory even more unstable, and therefore triggered even more rapid change early on, leading to final phoneme inventories that looked nothing like the original. For example, in Khulls voiceless aspirated stops were very abundant, outnumbering in speech the ejectives and the voiced stops combined. The Ogili language early on merged the ejectives and the plain voiced stops into a single category, and reduced the amount of aspiration on the voiceless aspirates, leading to a hypothetically more stable opposition of voiced vs voiceless stops. However this system turned out to be even more unstable, because in Khulls the most common ejective sound by far had been /ḳ/, which meant that in early Ogili, even with the primordial voiced stops added in, the most common voiced stop in the language was /ġ/ (spelled with a dot to differentiate it from the more common fricative /g/). This instability triggered another change: all of the labialized consonants became bilabial, meaning that, for example, /kʷ/ became /pʷ/. Since most primordial voiced stops had been labialized to begin with, this introduced a lot of /bʷ/ into the language, meaning that now the most common voiced stops were /bʷ/ and /ġ/. This still proved unstable, so labialization was eliminated, leaving Ogili with a lot of /b/ and /ġ/ but still nearly no /d/. However, these last two changes had radically transformed the sound of the language, since of the five contrasting places of articulation for stops, three of them were now realized as bilabials; that is, primordial /pʷ p t k kʷ/ had become /p p t k p/.
By contrast, Goranu's sound changes are conservative and often polyconditional, meaning that they appeared only when certain conditions were true.
Early sound changes
Note that /y/ represents IPA /j/.
Most early vowel changes were unconditional, like those of Khulls itself. This is largely because most morphemes are one syllable long and do not respond to sandhi processes triggered by neighboring morphemes.
Velars became palatalized before /i/. (Possibly exclude /ŋ/.)
/ʷi/ > /i/ unconditionally.
Probably /a e i o u yi yu/ > /a ya i ə ʷi yi yi/ unconditionally. Thus, there is a gap of */yə/, but this is no different than Khulls' own gaps. Note that primordial /u/ and /ʷu/ merge as /ʷi/, which is generally pronounced as a flat [u] with no audible glide at the beginning.
/r/ > /ř/ ... this is actually a spelling change, but is grouped with sound changes for convenience, since a new /r/ appears later.
Likewise, the postalveolar consonants come to be spelled as /ty, sy/ etc but this is not a true sound change. Note that at this point the language still considers labialization to be a part of the consonant but palatalization to be a separate sound occuring between a consonant and a following vowel.
/ʕ/ > /0/, which means that /ʕʷ/ comes to be seen as /lʷ/. However, there is no change in the pronunciation. Possibly, free-standing /y/ comes to be seen as /ly/.
Ejectives survive, unlike most other Khul languages, and some of them even make moves on plain voiceless stops.
All labials swallow palatalization (again). thus pya > pa, pʷya > pʷa, etc.
Possibly a change of final labialized consonants .... labialization survives, but the consonants merge with their unlabialized counterparts in all other manners. That is, instead of, e.g., /s/ vs /hʷ/, it changes to /h/ (likely [x]) vs /hʷ/. The bilabials would not change since they have members on both sides of the divide.
Seven position setup
Thus, there could be argued to be seven POA's for stops at this point: pʷ p t tʲ kʲ k kʷ. Frics also probably seven: s sʲ xʲ x xʷ h hʷ (hʲ would have merged with xʲ). Thus, labials can be united with glottals.
Unlike Khulls, palatalization is considered part of the consonant now, because it can no longer co-occur with labialization.
The consonant inventory is more conservative than most other branches of the family, and did not attempt to repair Khulls' heavy bias towards peripheral POA's (i.e. labials and dorsals). However there will probably be at least some contrasts that Khulls did not have, for example perhaps expanding the nasal series from /m n ŋ/ to /m mʷ n ñ ń ŋ ŋʷ/, although the labialized forms would be rare, probably arising more commonly from clusters in which a previously existing vowel dropped than from the unconditional but rare sound change of /u/ > /ʷi/.
The Great Vowel (and Tone) Shift
This is Goranu's only period of rapid change. This is also the only time when the number of syllables in a word changed.
Unstressed low-tone /i/ disappeared unconditionally. Note that this was still a tonal language, and there were a few examples of unstressed syllables exhibiting tone contrasts. These contrasts ended with this sound change, at least for /i/. The loss of /i/ essentially brought Goranu's plain and labialized consonants into line; previously, in Khulls, the labialized consonants and their corresponding plain ones almost never formed minimal pairs at the end of a syllable. That is, Khulls /kʷ/ could occur syllable-finally (it had earlier been /kŭ/), but the plain /k/ could not. Now, with this change, this distinction was eliminated.This shift could be thought of as being simply a long-delayed second phase of the much earlier (~4000 years) Khulls shift. Note that this sound was always [ɨ], not [i], so it did not palatalize consonants as it disappeared.
Next, unstressed syllables shifted to low-tone regardless of what the vowel was. Goranu still made distinctions of tone, but they were limited to one syllable per word and thus Goranu could at this stage be described as a pitch-accent language.
However, the previous sound change caused the stress accent to become more powerful. Possibly /ə/ > /i/ in remaining unstressed syllables, which was spelled in the native alphabet with a diacritic, as if it were simply a modified consonant cluster. Thus, the "high consonants" were formed (not to be confused with palatalized consonants, since they were not).
Voice and aspiration spread across these very short /i/ syllables, so that, e.g. ḳit > kit (that is, [kʰɨtʰ]).
At this time, Goranu still retained four of the five basic tones of Khulls: ă à ā â, the IPA values of which are respectively
- ă [a]
- à [áʔ]
- ā [á:]
- â [àʕ]
With vowel length being also affected by whether or not a syllable ended in a consonant. (it is possible that the tones get "messed up" a lot by the loss of unstressed /i/. Think of what would have happened if the sandhi had crystallized as the /i/'s were dropped instead of being analogized ... e.g. primordial as and ăsi would have different tones).
Here the tonal contrasts began to get weaker, since each of the four tones was marked by a difference in the syllable coda, although one of them was only a difference of length. In particular, the contrast between word-final ă and à began to disappear, with the final glottal stop increasingly difficult to hear as the contrast in pitch began to fade.
Later sound changes
experimental sound changes
/x xʷ/ > /q qʷ/ (no ejective form) /g gʷ/ > /x xʷ/ This idea makes sense and has probably occurred in natlangs, but it leaves a gap of */q̇/.
Resolution of consonant clusters in ways that favor traditionally underrepresented sounds such as /d/ rather than the historically favored Khulls phonemes such as /p/, /pʷ/, /kʷ/, etc.
The collapse of *ALL* unstressed vowels into consonant clusters, although many of these involve syllabic consonants and thus are "easier" to pronounce.
The reshaping of consonant clusters (separate from the above shift) into new consonants and in some cases clusters. These involve new methods and places of articulation, such as perhaps dental vs alveolar and perhaps labial vs labiodental. If the language is near Poswa at this time, it may share in Poswa's shifts such as p > f, m > v, etc when similar conditions are met.
The important thing is to think of the unstressed syllables such as sa-, sɨ- and so on as single consonants, and the native speakers may have used Andanese glyphs for these. At first they could only occur before other consonants, but when the new MOA/POA distinctions began to bleed forward, they soon could touch vowels as well. An example sound change might be /sɨl/ > /ɫ/, and /sal/ > /sl/.
kēli > ćāl "rainbow"