The glebst of gleb, V2.0

[Raholeun]

Seed 1605849964 for y'all:

/m n̻ ɲ ɴ/
/p t̻ k q/
/b ɓ̰ d̻ ɗ̰̻ g ɢ ʔ/
/s̻ h/
/l̻ j/
/i, u, e, o, ɛ, ɔ, a/
(C1)V(j)(C2)

Quite nice, but that laminal creaky voiced alveolar implosive is giving me a hard time.

[Man in Space]

I was working with gleb tonight and it brought this thread to mind. Some stuff I came across:

• Seed 1888463642: A vowel inventory of /a e ɨ/. That's it. As a bonus, the consonants are incredibly vanilla and balanced to contrast with this.
• Seed 1833341201: Six-vowel system, of which the only high lax vowel is /ɪ̈/. The rest is a fairly standard five-vowel setup.
• Seed 1845532535: A vowel inventory of /a e ø i u/. (I know Hopi's only front-rounded vowel is /ø/, but doesn't it at least have /o/?)
• Seed 1858770627: Palatalalized alveolar lateral obstruents occurred, but without a corresponding plain series.
• Seed 1809424268: Gave me a rather bizarre vowel inventory: /a ã e ẽ i ʉ ʉ̃ o õ u/. (Maybe ʉ̃ is allophonically lowered, but still…)
• Seed 1782520844: The only rounded consonants are /sʷ zʷ kʷ gʷ ʟʷ/. The first two are the only consonants in their series with rounding—i.e., there is no /tʷ dʷ/—and there's no plain /ʟ/ either. As a bonus, this language distinguishes /sʲ zʲ/ vs. /ʃ ʒ/. Double bonus: Coarticulated bilabial-palatals!
• Seed 1691484330: The vowel system is /a aˤ ɛ i ɔ ɔˤ o u/. I know that some Gbe languages are back-vowel heavy, but this just seems weird; I think it's the low-vowel pharyngealization.

[gestaltist]

As a bonus, this language distinguishes /sʲ zʲ/ vs. /ʃ ʒ/.

That doesn't strike me as too weird if /s z/ are dental.

[cedh]

Seed 1691484330: The vowel system is /a aˤ ɛ i ɔ ɔˤ o u/. I know that some Gbe languages are back-vowel heavy, but this just seems weird; I think it's the low-vowel pharyngealization.

This is not weird at all, it's basically just a two-column six vowel system /i e a/ :: /u o ɒ/ with /e/ slightly lowered and /ɒ/ slightly raised, plus two pharyngealized vowels that only contrast front vs. back. Quite nice actually...

[bbbosborne]

this is what all conlangers see when they ascend into heaven

[alynnidalar]

You sure you didn't get heaven mixed up with that other place?

[Xwtek]

You sure you didn't get heaven mixed up with that other place?

Just look at the allophony. Noob conlanger wouldn't bother with allophone.

[Birdlang]

I seem to remember reading about the Gbe language Jabo having pharyngealized versions of its vowels, so the pharyngeal back vowels aren’t too bad. Also, all of its voiced fricatives are breathy voiced @Man In Space.

[Darren]

I generated some of these ages ago, but they're the glebst I've seen (all from the online version):

• Seed 1418999732 has /i ɯ u a ɒ/ in the vowels, which is quite wierd in itself, and /p t k/ for the consonants. Only /p t k/. Also one of only four I've found with eight phonemes.

• Seed 1321355092 has only the consonants /k͡p t̻ k s̻/. No plain labials, just /k͡p/.

• Seed 1665396524 has the vowels /iː ɨː ɯː uː ɪ̈ː ɛː ʌː ɔː a aː/, i.e. all long apart from /a/.

• Seed 1421993047 has /i y ʉ u a/ for its vowels.

• Seeds 1518943826 and 1925090786 only have plosives for consonants; /p ɓ t̻ ɗ̻ ṯ ɗ̱ k ɠ q ʛ/ and /k͡p g͡b g͡b̰ p b b̰ t̻ d̻ k g dː̻/

• Seed 1600766948 has only sixteen possible syllables (/me ne ŋe pe te ke se e mo no ŋo po to ko so o/).

• Seed 1493228184 has syllabic /k̩̟ʼ q̩ʼ ʔˁʲ̩ ʔˁ̩/. I'm not sure those are humanly possible.

[Nortaneous]

Seed 1418999732 has /i ɯ u a ɒ/ in the vowels, which is quite wierd in itself

Not too far from /i ɨ u a o/, which is attested in Numic.

No plain labials, just /k͡p/.

/kp/ without /p/ is attested. Even weirder, /b̪/ without /b/ is attested from Kajoko. But these languages all have other labials.

Seed 1665396524 has the vowels /iː ɨː ɯː uː ɪ̈ː ɛː ʌː ɔː a aː/, i.e. all long apart from /a/.

Probably happens in Austroasiatic somewhere - Vietnamese has a length contrast in only two vowels. Khasi and Dorig have the opposite system: the only long vowel is /a:/. But if this system actually existed, it'd probably be reported as a quality contrast.

Seeds 1518943826 and 1925090786 only have plosives for consonants; /p ɓ t̻ ɗ̻ ṯ ɗ̱ k ɠ q ʛ/ and /k͡p g͡b g͡b̰ p b b̰ t̻ d̻ k g dː̻/

This is pretty weird. I can't think of any languages that only have plosives for consonants, unless you count Rotokas or Proto-Lakes Plain. But there's no necessary reason one couldn't exist - all the necessary components are attested.

Seed 1493228184 has syllabic /k̩̟ʼ q̩ʼ ʔˁʲ̩ ʔˁ̩/. I'm not sure those are humanly possible.

Syllabic voiced plosives have been claimed for Kpelle for complicated reasons involving tone. Sometimes a few Berber languages and Nuxalk are analyzed as having syllabic voiceless plosives. (Also Mansi?) But palatalization and pharyngealization are AFAIK incompatible.

[Darren]

Seed 1665396524 has the vowels /iː ɨː ɯː uː ɪ̈ː ɛː ʌː ɔː a aː/, i.e. all long apart from /a/.

Probably happens in Austroasiatic somewhere - Vietnamese has a length contrast in only two vowels. Khasi and Dorig have the opposite system: the only long vowel is /a:/. But if this system actually existed, it'd probably be reported as a quality contrast.

When I vaguely toyed with going further with this one, I reanalysed all the long vowels as short / unmarked for length and short /a/ as being a sort of ultra-short "yer" /ă/.

Seeds 1518943826 and 1925090786 only have plosives for consonants; /p ɓ t̻ ɗ̻ ṯ ɗ̱ k ɠ q ʛ/ and /k͡p g͡b g͡b̰ p b b̰ t̻ d̻ k g dː̻/

This is pretty weird. I can't think of any languages that only have plosives for consonants, unless you count Rotokas or Proto-Lakes Plain. But there's no necessary reason one couldn't exist - all the necessary components are attested.

I think the most unnaturalistic thing is that there are so many plosives. If a language has a tiny inventory like PLP or Rotokas, it's bound to have some plosives and there's likely a way of analysing them as having only plosives, but these two have ten and eleven phonemes. I'd be very surprised if the voiced plosives didn't lenide or voiceless ones become fricatives.

Seed 1493228184 has syllabic /k̩̟ʼ q̩ʼ ʔˁʲ̩ ʔˁ̩/. I'm not sure those are humanly possible.

Syllabic voiced plosives have been claimed for Kpelle for complicated reasons involving tone. Sometimes a few Berber languages and Nuxalk are analyzed as having syllabic voiceless plosives. (Also Mansi?) But palatalization and pharyngealization are AFAIK incompatible.

Actually, looking back at the inventory, there are also creaky-voiced vowels. Maybe the syllabic glottal~ejective plosives represent ultrashort creaky-voiced vowels, each just one (voiceless) creak... in that case, /ʔˁʲ̩/ would probably be something like [æ̰̆], with the pharyngealisation representing lowness and the palatalisation frontness. Or it's just gleb being gleb.

[mae]

-

[Nortaneous]

I'm not really sure why people consider (Central) Rotokas to only have phonemic plosives. It's consistently claimed that [s] only arises as an allophone of /t/ before /i/ but in Firchow & Firchow's dictionary there are a fair number of (near) minimal pairs between /t/ and /s/, such as sara 'be of the same clan' and taraa 'look upward'. Some of these words are obvious loans, but others aren't, so what gives?

IIRC t+i > s / _V

Some of the reconstructions for Proto-Lakes Plain are also strange to me for similar reasons. The word for 'sun' is reconstructed as *tio but none of the languages actually attest an /i/; we simply have a correspondence between /so/ in some languages and /to/ in others. Sikaritai has sɛja which I thought might be vowel breaking/diphthongization but Clouse claims that Sikaritai /j/ derives from *di, and /ɛ/ corresponds to /a/ elsewhere, so I guess this just isn't cognate. In cases like this I don't see what the justification is for assuming there was an *i not actually represented anywhere rather than assuming that we really do have to reconstruct *s.

The current Proto-Lakes Plain reconstruction is probably not great, but how would it be improved when there's so little published material?

[mae]

-

[Darren]

Regarding proto-Lakes Plain, Clouse says in his paper that :

Although the phone [s] occurs in all of the languages, in no instance are there cognates where the reflex *s is chosed for all languages; *t is always chosen as the reflex in one or more of the languages, but not consistently in the same languages In the modern Lakes Plain languages [s] can occur with any vowel, but the Lakes Plain languages in the Wapoga headwaters area (Deirate, Weirate, Faia, Sehudate), the sequence [si] does not occur. In addition, with the exception of one word, the sequence [tiV] does not exist in the Far West or central and western Tariku languages and in the eastern Tariku languages it is exceedingly rare. This evidence prevents the positing of *s in PFW and PT, although it is evident that its status as a full phoneme is becoming more certain.

Basically there's one branch where [tiV] doesn't occur (othe [CiV] sequences presumably do) and there is never /s/ in all branches for one word. He also says that */ɾ/ may or may not have been phonemic; suggesting that it might have lost its distinction with */d/ at some point or was never a different phoneme. However the reflexes of *VdV and *VɾV are distinguished in proto-Far West as *VɾV and *VØV.

[Tropylium]

How many of these plosive-only seeds have nasals triggered by nasal vowels or epenthetic semivowels?

[Darren]

How many of these plosive-only seeds have nasals triggered by nasal vowels or epenthetic semivowels?

I don't think any of them do. None of them have nasal vowels, or any nasal phones. I'm not sure Gleb does epenthetic semivowels, but none of these have them.

[Birdlang]

Seed 1671718241 has a consonant inventory of /m n t k f s r j/ and /i ɛ ɛ̃ a ã/+creaky voice for vowels.
Seed 1563695538 has a consonant inventory where the only fricative is /h/ and only approximant is /ʟ/ among /p b t d ʦ ʣ k g kˁ gˁ/ and a standard 5 vowel set plus long vowels. Also there are no nasals.
Seed 1536895816: every voiceless plosive has an aspirated and ejective version. Fricatives can be ejective, approximants have voiceless versions, and the vowels are /i u e/.
Seed 1509502828: no approximants, vowel set is /i ʉ u ʊ̈ ɛ ɔ/.
Seed (unknown): vowels are /i iː u uː eː oː ɛː œː ɔː a aː/. Also /gʷ/ without regular /g/.
Another unknown seed: vowels are /i y u a/ and consonants are /m n p t k q h l ɰ/.
Seed 1367719385: just look at it. Only approximants are an alveolar tap & trill.
Seed 1331003844: vowel inventory of /i ɛ a/. And a pharyngealized velar approximant /ʊ̯ˁ/.
seed 1247710007: shown as picture

[Darren]

Seed (unknown): vowels are /i iː u uː eː oː ɛː œː ɔː a aː/. Also /gʷ/ without regular /g/.

I think that the /gʷ/ thing is often used as an approximant in the syllable structure, along with other unexpected voiced plosives.

Seed 1331003844: vowel inventory of /i ɛ a/. And a pharyngealized velar approximant /ʊ̯ˁ/.

I get that type of vowel inventory wierdly often from gleb. I even made a plan for a language family where one branch had /i e a/ and the other had /u o a/ as their inventories.

seed 1247710007: shown as picture

That's big, but surprisingly normal. I think the largest I ever generated have 34 places of articulation, but I doubt I could find it now.

______________________________________________________________________________

I doubt this is the right thread for this, but I still have no idea how to actually download the offline version of gleb. Is there a simple way to do it without coding?

[DesEsseintes]

Seed 1201783956 has the vowel inventory /æ i u y/.