Zompist Bboard Again

bradrn wrote: ↑Sun Sep 04, 2022 8:42 am

Rounin Ryuuji wrote: ↑Sun Sep 04, 2022 8:31 am Saying my /ji/ over and over, I think there's some sort of difference of quality between the glide and the vowel. I don't detect any frication, lateralisation, or anything else that would make it more consonant-y.

I’m a big proponent of the ‘[j] and [i] are the same thing’ position, but I have to admit that I tried this exercise too and found the same thing. I even tried making a spectrogram, but at first glance that showed no obvious difference. I do wonder if I’m imagining a difference between them where there isn’t one.

The /j/ should be shorter in duration than /i/. There might even be slightly more narrower constriction.

What are some conceivable ways to end up with a T/D/Dʱ series?

I found ⁿD > Tʱ in e.g. some Bantu languages.

I also found DD > Dʰ in Kiput .

Both of these, though, don't seem to be technically [+voiced +breathy] (one is [-voiced +breathy], the other [+voiced +aspirated], and the 2nd (in the real word at least) only functions word-internally.

Any other examples (real world or otherwise) or colorable arguments? Any thoughts on the first example just not devoicing? I'm aiming to have the modern language involve stress/pitch rather than tone.

Welcome to the ZBB, Oxygenman!

As for your question, my personal theory about the PIE stop system (which is, according to the standard reconstruction, just like what you have in mind) may give you inspiration. This goes as follows:

1. Pre-PIE had just two types of stops: voiceless and voiced. The voiceless stops were aspirated: Tʰ D.
2. Some morphemes carried a prosodic feature which caused any aspirated stops in them to voice: Tʰ D Dʱ.
3. In most branches, the (voiceless) aspirated stops lost their aspiration.

I don't know whether this is what actually happened, but at least it seems to work to me.

Oxygenman wrote: ↑Fri Sep 23, 2022 6:45 am What are some conceivable ways to end up with a T/D/Dʱ series?

If you want something a posteriori, you could create an Indo-European (or a Para-Indo-European) language with an extremely conservative consonant inventory.

Oxygenman wrote: ↑Fri Sep 23, 2022 6:45 am What are some conceivable ways to end up with a T/D/Dʱ series?

I found ⁿD > Tʱ in e.g. some Bantu languages.

I also found DD > Dʰ in Kiput .

Both of these, though, don't seem to be technically [+voiced +breathy] (one is [-voiced +breathy], the other [+voiced +aspirated], and the 2nd (in the real word at least) only functions word-internally.

Any other examples (real world or otherwise) or colorable arguments? Any thoughts on the first example just not devoicing? I'm aiming to have the modern language involve stress/pitch rather than tone.

Low tone/accent creates breathy-voiced vowels whose quality transfers to voiced stops. And there's certainly no reason why [-voiced +breathy] couldn't spontaneously become [+voiced +breathy] either. [+voiced +aspirated] is also likely to become [+voiced +breathy] or [-voiced +aspirated] spontaneously.

I wonder whether Oxygenman is just researching the prehistory of PIE, masquerading as a conlanger to avoid heated debates

Is morpheme-internal vowel harmony a thing?

Yes. These are found in languages if West Africa and Oceania, IINM. They often have slightly different properties and look less regular/natural though. I would need to check my notes to give you more details, though.

It is that I have low-vowel harmony between <e> and <a>, the degree of spread of the harmony is hard to determine (e.g. usually stopped by consonant clusters, but sometimes not, like the below example) but the direction is usually regressive and affixes cannot trigger the harmony.

So basically you can have kemē-na "stand-1cs" (not kemē-ne or kamā-na) and ne-kmē-na "FUT-stand-1cs" (not na-kmē-na or ne-kmē-ne). But I also have relative subject verb markers that are basically composed like this -n<ēz>a "1cs<REL>" and -t<ēz>as "3ns<REL>" so I'm wondering if the morphemes can harmonize to -nēze and -tēzes

So I'd have a verb form kā-nēze "I who am" and it doesn't become kā-nāza or kē-nēze

Ahzoh wrote: ↑Sat Sep 24, 2022 10:39 am Is morpheme-internal vowel harmony a thing?

Old Japanese seems to have had something like this, on a small level. The vowel denoted as o₂ was not found in the same morphemes as the vowels a, o₁ or u, according to Arisaka, but could co-occur with itself, i₁, i₂, e₁ and e₂. This may have been a remnant of a farther reaching vowel harmony in Proto Japonic or may have had to do with the phonological changes that occurred to produce the Old Japanese vowel system.

Ahzoh wrote: ↑Sat Sep 24, 2022 11:37 am [...] But I also have relative subject verb markers that are basically composed like this -n<ēz>a "1cs<REL>" and -t<ēz>as "3ns<REL>" so I'm wondering if the morphemes can harmonize to -nēze and -tēzes

So I'd have a verb form kā-nēze "I who am" and it doesn't become kā-nāza or kē-nēze

In some Austronesian languages, infixes are subject to phonological regularities that otherwise only apply inside morphemes only. The Austronesian one is about two labial consonants in a root, but I think your example looks perfectly natural, too. A process like sporadic low vowel harmony is exactly what I would expect morpheme-internally.

Nitpick, but what does 'morpheme-internal vowel harmony' even mean? If a sound change is stopped by morpheme boundaries, it'll soon cease to be productive and older morpheme forms will be replaced by newer ones. There's no ground for synchronic variation here.
ne-kmē-na "FUT-stand-1cs" contains three morphemes and harmony isn't restricted to just one of them. Infixes are also separate morphemes.

Zju wrote: ↑Sun Sep 25, 2022 4:50 pm Nitpick, but what does 'morpheme-internal vowel harmony' even mean? If a sound change is stopped by morpheme boundaries, it'll soon cease to be productive and older morpheme forms will be replaced by newer ones. There's no ground for synchronic variation here.
ne-kmē-na "FUT-stand-1cs" contains three morphemes and harmony isn't restricted to just one of them. Infixes are also separate morphemes.

From Ahzoh’s posts, he seems to be referring to infixes which are phonologically placed within other morphemes.

Zju wrote: ↑Sun Sep 25, 2022 4:50 pm Nitpick, but what does 'morpheme-internal vowel harmony' even mean? If a sound change is stopped by morpheme boundaries, it'll soon cease to be productive and older morpheme forms will be replaced by newer ones. There's no ground for synchronic variation here.
ne-kmē-na "FUT-stand-1cs" contains three morphemes and harmony isn't restricted to just one of them. Infixes are also separate morphemes.

It means that if you have a morpheme *ʕalar it becomes ʕælær and not ʕælɑr and *kamaʕ becomes kæmæʕ and not kɑmæʕ. It is akin to simple vowel mutation but theoretically able to spread through a morpheme that is greater than two syllables.

I also conceived of suffixes like -nēze as a superunit even though they are really two morphemes -na "1cs" and <ēz> "REL". So I want -n<ēz>a to become -nēze and I don't want it to mutate the root.

But I reread the wiki on vowel harmony and remembered that "stem-controlled vowel harmony" is a thing and decently describes how my words harmonize the low vowels.

That sounds more like a sound change than vowel harmony, which is a synchronic process.

So I want -n<ēz>a to become -nēze and I don't want it to mutate the root.

Affixes do develop irregularly, so you don't even need to make up a sound change to explain that. You can have it to be a just one off -n<ēz>a → -nēze for that affix alone.

Zju wrote: ↑Mon Sep 26, 2022 12:30 pm That sounds more like a sound change than vowel harmony, which is a synchronic process.

So I want -n<ēz>a to become -nēze and I don't want it to mutate the root.

Affixes do develop irregularly, so you don't even need to make up a sound change to explain that. You can have it to be a just one off -n<ēz>a → -nēze for that affix alone.

That all depends upon whether it is productive, i.e. whether it affects new coinages, loans, and repurposing words from one word class as members of another word class.

I've been left with the impression that this is a part of verbal morphology and that there are no conjugations.

How common is phonemic /d͡z/ without phonemic /z/?

StrangerCoug wrote: ↑Sat Jan 14, 2023 7:06 pm How common is phonemic /d͡z/ without phonemic /z/?

Latin, depending on the pronunciation.

Man in Space wrote: ↑Sat Jan 14, 2023 10:12 pm

StrangerCoug wrote: ↑Sat Jan 14, 2023 7:06 pm How common is phonemic /d͡z/ without phonemic /z/?

Latin, depending on the pronunciation.

IIRC Old Latin had [z] but it became /r/.