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 date 6843 chosen here is of no consequence, as it refers to something that happened in Moonshine (and perhaps Poswa) instead; therefore, the daughter languages of Ogili most likely break off at different times while the "core" language, with by far the most speakers, continues to be spoken in the Swamp.

Tones can be numbered from 1 to 5, 5 being highest.

TONE shifts are in PURPLE for easy reading.

The initial consonant inventory was something such as

Rounded bilabials:       pʷ  ṗʷ  bʷ      hʷ          w
Spread bilabials:        p   ṗ   b   m   
Alveolars:               t   ṭ   d   n   s   r   l
Postalveolars:           č       ǯ       š   ž  (λ)  y   
Velars:                  k   ḳ       ŋ   x   g
Labiovelars:             kʷ  ḳʷ  ġʷ      xʷ  gʷ
Postvelars:              q               h       ʕ
Labialized postvelars:   qʷ

And the vowels /a e i o u/ on six tones: à ă ā á â a͆, where the last four were two pairs whose members differed in sandhi effects only. All four of those tones were long, the first set being high and the last set being low. The à tone was also high, but ă signifies a short mid tone, not a short low tone. The only short low tone in Leaper was that which arose from sandhi.

1. The voiced sounds r d shifted to b.
2. The vowels e i u, on all tones, shifted to ʲa ʲi i. 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.

   It is possible that /ʷu/ survives intact, not shifting to /ʷi/, since it would only shift to a plain /i/ in the end.

3. The inherited vowel o came to be spelled ɜ.
4. Labialization defeated palatalization. Ejectives also defeat palatalization, at least in the case of /ʲi/. Uvulars also did not palatalize.
5. Any syllable before a closed syllable shifted to the LOW tone a (SPELLED WITH GRAVE ACCENT LATER BUT NOT YET). This sound change had already happened in Leaper, but there is no convenient way to mark it because it was always created by sandhi. Note also that this sound change even occurred before --historical-- closed syllables, but again this was marked on the stressed syllable in Leaper and there was no way to mark it on the unstressed syllable.
6. All syllables occurring after a post-tonic ejective (/ṗʷ ṗ ṭ ḳ ḳʷ/) shifted to the BOTTOM tone â.

   This is the primary source of the 3-1 tone melody, and if vowel+ʕ is considered as a closed syllable, it will actually be the only source other than late-stage compounds.

   However, if the BOTTOM tone is always long, this shift cannot happen, and it must be LOW instead.

7. All unstressed syllables occurring after a voiced stop or voiced fricative shifted to the low tone (SPELLED WITH GRAVE ACCENT LATER BUT NOT YET). This happened even if they were pretonic.

   This sound change was added at a time when the sound change of /r/ > /b/ had not been added yet, and therefore voiced stops were very rare. It is possible that /b/ will pattern as an approximant to keep this sound change rare.

8. POSSIBLY, ALL mid tones (tone 3, Leaper ă) shifted to low tones (tone 2, no symbol in Leaper) after a stressed mid tone. This is the only way that there could be a contrast between Leaper's ăa & aă patterns; that said, this might carry little semantic information. Note also that this shift mimics the shift before it involving voiced stops, and that that shift mostly occurred after mid tones as well because after high tones there were only rare instances of such voiced sounds.
9. The ejectives ṗʷ ṗ ṭ ḳ ḳʷ shifted to plain voiceless stops pʷ p t k kʷ. These may also shift downward to a LOW or MID tone (or differ depending on what they were). Note that it was usually /ḳ ḳʷ/ that occured before stressed syllables.
10. If a word began with an unstressed g, that shifted to Ø and the tonic syllable shifted to the â tone.
11. If a word contained an unstressed h, the tonic syllable shifted to a new tone behaving as if it ended in /s/.
12. 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).
13. The voiced fricatives ž g gʷ ʕ ʕʷ shifted to Ø Ø w Ø w. (the "oo wow" shift.)
14. The velar fricatives x xʷ shifted to h hʷ.
15. The uvulars q qʷ shifted to k kʷ.
16. All labialized consonants shifted to pure bilabials. Thus pʷ bʷ kʷ ġʷ hʷ shifted to p b p b f.

    An important artifact of this shift is that words like /kan/ were replaced by /pă/, etc, due to Leaper's grammatical alternation between A and B stems. In this example, /kan/'s B-stem in post-Gold (early Leaper) was /kəŋa/, which then shifted /kəŋa > kʷŋa > ŋkʷa > kʷa/, here becoming /p/. This could also happen with coronals. Thus the pattern was CVC > BV, where B represents the labial corresponding to the original consonant, and where voiced stops serve for nasals.

17. Consonants before (or after?) a syllabic nasal were deleted.
18. The tone change from about 1,000 years earlier in which /ā/ and /á/ were merged was undone in the 1st person marker, making á into an even higher a̋. This reversal can only be explained by special pleading such as:
    1. There was a clitic only occurring after person markers, and the reversal of the tone merger operated only in this environment.
    2. Both ā and á became ultra-high a̋ in Ogili, but then THAT sound change was undone in the sandhi environment that /ā/ had always had (because it would be a very large pitch jump between syllables), while the sound change was retained in the weaker sandhi environment that /á/ had always had. This would mean the sound change also operated on other morphemes besides the 1st person marker, unless lowered later by yet another condition.
    3. An -s or -l of indeterminate origin was inserted after the 1st person marker, and the sound change to the ultra-high a̋ actually did not happen here at all, but only later, when other syllables also shifted up.
19. Tone shifts took place in tandem:
    1. The long high tone ā became the high tone á, except before a coda of either of /l s/, in which case it rose further to the extra-high tone a̋.
    2. 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.
    3. Short mid tones ă before a voiceless coda rose to the high tone á.
       1. Coda s was then deleted, meaning the only remaining voiceless codas were labialized.
    4. 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/.
    5. 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.
    6. 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.
    7. All unstressed vowels became either ā or à, depending on sandhi.
20. Syllable-final p b l shifted to pʷ bʷ w. The new /w/ can be spelled as /lʷ/ temporarily for precision.
21. 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.

    This shift was unconditional at least before /i u/ and probably before /ɜ/, but might have been conditional or even absent before /a/. This is because the sequences /ʷi ʷu/ are just IPA [y u] in the final stage of the language. An alternative to shifting /aʷ/ > /ʷa/ is to shift it to /o/ instead, and then to skip the later sound change of /ʷa/ to /o/. This is because the /o/ ends up unrounding towards the maturation date.

22. The voiced stops ǯ ġ became the voiced fricatives ž g. Note that the labiovelar had become /b/ by this time.
23. The long high tone á:merged with the extra-high tone a̋.
24. Before a coda that was any of /m n ŋ lʷ bʷ/, the mid vowel sequences ɜ ʷɜ shifted to o ʷe.

    These were originally written in the opposite order, under the assumption that both were inherently rounded at least to start with. The new idea is that /ʷɜ/ > /ʷe/ arose from dissimilation while /ɜ/ > /o/ was of little consequence at the time. Another possibility is that both shift to /o/, and that /ʷe/ simply is somewhat less common than /ʷa ʷi/.

25. The voiced velar fricative g disappeared to Ø.
26. 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/.
27. 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.)
28. All remaining codas were deleted.
29. The tone sequences 41 42 shifted to 42 43. It is possible that it was just 41 ---> 42, which would make old bisyllables finally merge with new compounds. It may be important to see whether the 43 tone pattern is needed for something else. This shift was triggered by the relative scarcity of monosyllabic tone-2 words, meaning that in compounds, 41 was common but 42 much less so.

Thus the consonant inventory had become

Bilabials:         p   m       f   b  (w)
Alveolars:         t   n   c   s       l       
Postalveolars:         ň   č   š   ž   λ     
Palatals:              ń   ć   ś      (y)
Velars:            k   ŋ       h       

And the vowel inventory was /a e i o u wa we wi/. It is possible to analyze /o u/ as a fourth pair, /u wu/, arriving at a four-vowel inventory with room for a /w/ glide before all four vowels. Unlike the parent language, this /w/ glide is not considered part of the preceding consonant.

TONES

Tone 1

The lowest tone, often found in monosyllables.

"Low verbs"

In theory, tone 1 could be the 1st person verb ending after historical ejectives if the sound change up above in which all vowels become ultra-low after an ejective is allowed to apply even to the primordial á tone, since it would not be corrected by any future sound change. The 2nd person forms of these verbs would be of tone 2 (see below), and therefore the tone pattern is the opposite of that of the wider class of verbs. All of these verbs would begin with a voiceless stop because that is the reflex of the ejectives; therefore they would be a closed class. Note however that the 3rd person form would just be homophonous with the 1st person form, so if low verbs exist they might be mostly for evidentials or some other such thing that does not often require 3rd person marking.

Tone 2

This is rarely found in monosyllabic content words, though a few examples do exist.

It is possible that this is the tone expressed on the first of two syllables of an atomic word marked with the accusative case, which would mean that the accusative case could persist in Ogili as distinguished from the genitive. This would give the tone pattern 24 for the accusative case whether the nominative was 33 (from Leaper CăCa), 42 (from Leaper CàCa), or something else. However it is most likely that the tone pattern is in fact 34, from Leaper CaCà, and therefore the accusative and genitive are merged into an oblique case. Put another way, unless changed by a later process, all tones preceding Leaper's à tone are reflected as 3 in Ogili, and all tones after it are reflected as 2. This assymetric pattern goes back to Leaper itself, and distinguished that tone from the symmetrically down-pushing ā and a͆ tones.

Tone 3

The reflex of many other forms.

Tone 4

This is the reflex of the original high short tone and the accusative. It is also the 2nd person verb marker.

IMPORTANTLY, this is also the reflex of the genitive of all five tones, because it comes from -s, -gṡ, and -hṡ, all of which could only be preceded by a mid tone, which developed into tone 3, and then into tone 4 when the /s/ was lost. Thus, even though a final -s raised the tone level of any syllable by one level, the reflexes of the genitives of 12345 are 44444.

Tone 5

Rare in content words, but still an integral part of the language, as it is the 1st person marker, and does also occur in a few content words, mostly monosyllables.

The 1st person inflection added to a verb (or noun) that ends in a tone 5 syllable is perhaps still carried over from Leaper, in which case the final syllable becomes two, with the pattern 35, and a /k/ is inserted between them. It is possible, however, that all verbs end in tone 3, in a sense matching Play, and making all verbs strong verbs.

Grammar ideas

Already in Leaper there were subsyllabic verbs like ṗ "to see" which attached to syllabic suffixes. Thus the uninflectable subsyllables remained in the language and were seen as ordinary morphemes. In theory, if the sound changes above were shuffled, this verb (and any others with an ejective before the verb inflection) could defy the tone patterns, but this could also be problematic, as a 1p would sound like a 2p on an ordinary verb, so perhaps it is best to keep the system in order.

urine

The word for urine in Leaper was just è, and although this was a homonym of a few other words, the sense of urine remained in the language and could take verbal suffixes as above as well as behaving as an ordinary noun.

The word for anus in Leaper was just ê according to an early writeup, which may or may not still be viable. Originally this was supposed to come from MRCA /àgi/, but the closest semantic match that remains is MRCA /gàḳi/ "tunnel, tube". A few such words vacillated between having initial /g-/ and having an initial vowel, so the reflex of both words could in fact be ê in Leaper. If this is carried out, then the word could change meaning in Ogili and the words corresponding to pee and poo in English would then just be é and ȅ.

Sound changes in the daughter languages

Another group

Consonants shift forward

Possibly all postalveolars shift to alveolars, and then all palatals but /y/ shift to postalveolars. The /ž/ might become /r/ instead of /z/.

Alternatively, this shift might occur in just one daughter language, while another daughter shifts all palatals but /y/ to dentals instead, leaving the postalveolars at least partly in place (if they do shift, it is to plain alveolars, as above).

Vowel shifts

A possible vowel shift is to divide the daughters between /a i u ə wa wi wu wə/ and /a e i o u wa we wi/, where these actually denote the same parent system in two different analyses. here, /a i wa wi/ represent the same phonemes in both lists, but the left inventory's /u ə wu wə/ correspond to /o e u we/ on the right side.

Vowel-consonant tandem shifts

It is possible that a second wave of palatalization will occur, in either branch but more likely with the right-hand branch, such that /e i we wi/ shift to /ʲe ʲi e i/, nearly repeating the sound change that had occurred 2-3,000 years earlier. /wa/ might simply shift to /a/, as this shift works in best in a setup where /o u/ had been [wo wu] all along (though the /o/ was *not* pre-labialized at the time of Ogili). This would leave the language with relatively few dorsal consonants.

In the /a i u ə wa wi wu wə/ clade, a similar wave of palatalization could occur, but it is also possible that all four pre-labialized vowels change the preceding consonant to a labial, even though the language had abundant labials already, perhaps under influence from Play's daughter languages or some of the remaining Dreamlandic languages (if there are any).

Further ideas II

This section refers to earlier shifts.

The palatal and postalveolar series may merge, with /λ/ shifting to /j/. Alternatively, /l/ could become a new /r/ and then /λ/ would become /l/, though the distribution may be unbalanced due to the scarcity of palatals before /ə u/ (or /e o u/ in the other analysis.

This language has pronouns and possibly person markers. Gold had person markers, but Leaper and Play both lost them; Leaper created new person markers, but they partially merged and thus could no longer serve on their own; thus Leaper created pronouns separate from the person markers. It is possible that Ogili keeps the pronouns and loses the person markers. (Play had neither; this was one of the reasons why Play was considered such a difficult language.)

Labial stops

It is possible that at some early stage, p shifts to ʔ after any vowel, and that vowel becomes a short high tone. Since this is the most common tone before /p/ anyway, it could be said that /p/ > /Ø/. It is also possible that some other consonants joined this shift, or did the same shift at different times. This would lead to more vowel sequences in the language, but they would be pronounced with glottal stops at least to begin with. The motivation for this change would be the uneven distribution of /p/ in the language, where it had been eliminated except in clusters (of which some later shifted to /p/, but these were also unevenly distributed). Then a "proper" /p/ sound comes from /kʷ ḳʷ/.

Either way, the bilabial stops p b (especially /p/) have a marked distribution, occurring mostly in monosyllables before tones 1, 3, and 4. It is possible that they will merge with some other sounds for which there is a nearly complimentary distribution, but they would be unlikely to merge with each other, because they came hand-in-hand through the same sound changes and thus occur in mostly the same places. The voiced stop /b/ has more use in unstressed syllables because it can also come from earlier /r/.

Probably the only good choice to shift /p/ would be /f/, though this would mean that the language had lost its /p/, or most of it, three times in about 2,500 years.

Alternatively, since the first shift essentially creates the impetus for the second, it could be that neither of them take place and that the language retains the distribution of /p/ from Leaper. A middle ground would be to shift /p/ > /ʔ/ but not the later /p/ > /f/, and equalize the tone distribution somewhat. If the tones cannot be equalized, this might even still be a good thing in that it would give auditory cues for which consonants were being pronounced.

Case mergers

The NOM and CIRC cases are the same in many nouns, and GEN and ACC also merge. Therefore these cases are dropped. The DAT case may emerge as a new genitive. However, in compounds, it is likely that the historical GEN would be used even though Gold and Leaper used the nominative here. This is because new compounds were not made using the old rules, but came from phrases.

Culture

This language was the official language of Baeba Swamp, originally spoken by the Iron party. The language could even be called Iron, but the Irons were just one of many groups who spoke this language, as they governed a multiparty democracy that was stable for thousands of years. It was later used anachronistically for the Iron party's early history, as they considered their nation to have been continuous throughout time. That is, they did not use "Old Ogili" to write about old happenings, except in linguistic scholarship.

This language resembles nothing in its area, but it is possible that Baeba is so populous that it would form a sprachbund of its own, and give out cultural influence rather than absorbing cultural influence.

Ogili is bounded on the west by Dreamlandic languages and on the east by Cosmopolitan Play languages. The Ghost language may have held on in the southeast border regions, but the Ghosts' center of population was much further south.

Note that the name Erala becomes Yabala here.