The Great Thread on Non-Concatenative Conlanging

[MrKrov]

To answer the OP, nope. Never.

[mbrsart]

I'm currently working on a triconsonantal lang with a Slavic-inspired sound system, written with the Russian alphabet. It's going to have a lot of stuff inspired by Ancient Hebrew.

[Creyeditor]

Sounds like it is set in the real world. Is there a conhistory?

---

In other news, a while ago I started a thread on Ook, which is basically non-concatenative all the way,. but wearing a concatenative cloak, or the other way 'round.

[eldin raigmore]

Originally, I decided to go for a system of biconsonantal rather than triconsonantal roots. This setup prevents content words from becoming too long and ensures the ready availability of minimal pairs. However, at present, I'm wondering if I should add some triconsonantal roots on top of the biconsonantal ones. It seems best to avoid having a situation where a large number of different lexemes are only minimally distinct from one another and where more primary concepts like 'father' and more secondary ones like 'to be extravagant' are all of the same or very similar length.

If you're the same Trebor -- Hi! Nice to hear from you again.

---

How big do you want your basic vocabulary to be? (that is, how many words?)
If you have 20 consonants and 5 vowels, and you allow up to 3 consonants and up to 2 vowels per word, and don't allow vowel-clusters and allow at most one consonant-cluster, and limit consonant-clusters to 2 consonants, you can have more than a million words.
CVCVC 200,000
CCVCV 200,000
VCCVC 200,000
CVCCV 200,000
VCVCC 200,000
That's five times 200,000; that is, a million.
You could also add in the 50,000 VCVCV words.

---

How many word-roots do you want?
If you want a million, and you have 32 consonants, you could have 1,048,576 (220 = 324) C1-C2-C3-C4 roots.*
IMO that's excessive.
But if you're using 3Cons roots (C1-C2-C3), you need:
86 consonants to get 625,000 roots:
61 consonants to get 225,000 roots:
37 consonants to get 50,000 roots:
32 consonants to get 32,768 roots.

But you'd need only 18 consonants to get 5,000 roots, and only 15 consonants to get 3,000 roots.

I would suggest that past the oldest and/or most-frequently-used few myriad roots, your conlang have some 4Cons roots, or maybe even some 5Cons roots if it has fewer than 32 consonants.

For instance, suppose your 'lang has just 16 consonants.
You could have
16^5 = 1,048,576 5Cons roots;
16^4 = 65,536 4Cons roots;
16^3 = 4,096 3Cons roots;
16^2 = 256 2Cons roots.

Most simple conversations could be handled using just the 2- and 3-consonant roots. Most single-field-of-expertise in-depth conversations could still get by using those plus 4-consonant roots. Your 'lang's version of the OED could fill up with those plus the 5-consonant roots.

*

Spoiler:

With 32 consonants you could have:
1,048,576 = 220 = 324 C1-C2-C3-C4 roots;
32,768 = 215 = 323 C1-C2-C3 roots;
1,024 = 210 = 322 C1-C2 roots.

---

Do two different CVCV words (or morphemes) have to be semantically related if they have the same consonants but different vowels?
With 18 consonants and 4 vowels you can have 5,184 different CVCV sequences.

OTOH with 18 consonants you can have 5,832 different C-C-C sequences.

---

Was any of that helpful?
What was helpful and what wasn't?
How helpful was it?

[HoskhMatriarch]

I'm trying to make every single regular noun and verb in my conlang have some form of apophony (the irregular ones being the result of suppletion), mostly ablaut but there will likely be a little bit of consonant gradation too. It doesn't seem likely to make the function of every new word clear with nothing but apophony, so there will almost certainly be some different fusional morphemes that occur along with it for at least some features of some words.

[atman]

There are a few irregular verbs in Atlántika where the forces of (simulated) linguistic change really conspired to create opaque, weird forms. Let's see a couple of examples:

Present:
labano [lə'bano] "I grab"
Here lab- is the synchronic root; -an- is an old present-forming suffix of unclear meaning, and -o is the usual 1SG thematic ending.

Preterite:
ilaxa [i'laʃɐ] "I grabbed"
This verb originally made a root aorist; in Atlántika it was remade as a sigmatic one (élapsa > eláfsa > iláxsa and so on). The i- at the beginning is the reflex of a past-forming prefix called the "augment", typical of Graeco-Aryan languages in general (Skt. a-, Cl.Arm. e-)

Perfect:
lilbèka [liɫ'bɛXɐ] "I have grabbed"
A most conservative feature of Atlántika is the continued use of the ancient IE reduplicated perfects. The main reason this is so seems to be that the creator of Atlántika likes them so much . Here, the -a- from the lab- root was deleted due to its medial unstressed position (if you're a medial unstressed vowel in Atlántika your days are numbered): lelabèka > lilə'bèxa > lilbèXa...

But labano isn't the worst verb around. That honour goes to (thunderbolts and lightning) xtilho, whose principal parts are written below (Note that Atlántika verbs only have 4 principal parts, Ancient Greek required six, some Indian language at least ten ):

xtilho ['ʃtiʎo] "I send"
xtilhèso [ʃti'ʎɛso] "I will send"
ixtèla [i'ʃtɛlɐ] "I sent"
sixtlòka [si'ʃtɫɔXɐ] "I have sent"

Besides the unproblematic future form, the preterite and perfect are truly something else. The root vowel changes each time, the reduplication in the perfect is obscured by the s > ʃ change before stops, and incredibly the former (double) lambda is pronounced as a palatal in the present, as a plain alveolar in the preterite, and as a velarized alveolar in the perfect!

[mbrsart]

I now have a few examples from Velyesh (Вэлеш), my experimental triconsonantal lang with a slavic sound system (heavily inspired by Biblical Hebrew).
гус - lip
гис - to kiss

Then there are a couple of different ways to describe the act of kissing, depending on how poetic you want to get.

Ски́зной По́л гугу́с сафу Ры́нне.
[ˈskiːznɔj ˈpɔl guˈguːz safu ʀɨɲːɲɛ]
с-кис-зной Пол гу~гус са-фу Рынне
3CS-give.PRF-3FS Paul PL~lip 3FS-to Rinnye
"Paul kissed Rinnye", lit. "Paul gave them the lips to her Rinnye." [Highly poetic]

Гыгзно́й Ры́нне.
[gɨgznɔj ˈʀɨɲːɲɛ]
Ø-гы~гис-зной Рынне
1CS-IPFV~kiss-3FS Rinnye
"I will kiss Rinnye."

P.S. Head marking is FUN