Conlang made from interleaved vowel-words and consonant-words

[Axas mlö]

VAVNAC (verbs are vowels, nouns are consonants)
I'm making a new conlang.

Main idea and goals:
The main idea is that nouns are consonants, verbs are vowels. A simple sentence with one verb and one noun just puts the two together by alternating the phonemes, so that if 'mother' is m-m-t and 'sleep' is i-e-u then 'mother sleeps' is mimetu.

I'm content for this language to be a sort of a proof-of-principle rather than a well-fleshed out conlang. One aspect of that: the name of the conlang is just VAVNAC.

One example:
In VAVNAC (pronounced as in IPA)
Wefanə, nuhaʔo vyʔa takazə kəʔeʔo, tiqaʔi nohaʔiʔu viʔy fefubodəsəgə.
Separated into consonant-words and vowel-words, and then the gloss:

wfn\ea, nh\ua-o v\ya tkz\aa k\əeo, tq\iai nh\oa-iu v\iy ffb-dsg\euo.

that.way\INS, I\begin(v.i.)-CAUSATIVE THAT.VERB\PST journey\ACC PREVIOUS.NOUN\long, REL\along I\move-backwards THAT.VERB\PRF chair-king\into.

Translation: It was in this fashion that I began the long journey by which I have backed into the throne.

Notes on the example:
Each word is made by smooshing together a consonant-word and a vowel-word. I've separated them out so you can see the parts better, with a backslash in between. Backslash is not a terrible stretch from https://www.eva.mpg.de/lingua/resources ... -rules.php , rule 4D. That page is also where I'm getting the all-caps glosses when they're abbreviations; and v.t. means the gloss is using the transitive meaning of the English verb, v.i. for intransitive.

When there are more consonants than vowels, the sequence of vowels is padded with ə, which we can consider the neutral vowel - enough ə's are added to match the length of the sequence of consonants. Similarly, the consonants are padded with ʔ.

PREVIOUS.NOUN refers to the immediately previous noun. THAT.VERB refers to a nearby verb, generally the main verb; it fills a noun slot and can be thought of as meaning the action, process, state, or fact of the relevant verb. Probably there's a better name for such a thing? Or a reason organizing it like this isn't a good idea?

More details:
Verbs are vowels, as are adjectives, adverbs, adpositions and things like ACC, and some miscellaneous particles. Nouns are consonants, as are pronouns and some miscellaneous particles.

My idea is inspired a little bit by triconsonantal roots but mainly by a wish to have two streams of information that are somehow separate.

I'm obviously not looking to make something naturalistic, but I don't want it to be vastly more non-naturalistic than is required by the main idea. Or rather, I'm totally fine with it being too regular, I'm resigned to it being too English-like, but I don't want it to be bizarre in ways that no human languages are — except for what follows from the premise. And more positively, I want things to fit well with that premise.

I have more worked out, but I thought maybe I should start with posting just a little.

Questions, feedback sought, etc:
Does the main idea already exist, in a natlang or conlang?

Are there things about this conlang - apart from the main premise, and necessary consequences - that are non-naturalistic in the sense of being implausibly different from what's attested in natlangs?

Do you have aesthetic suggestions? Is the main idea sufficiently different from triconsonantal roots to be interesting?

Does it seem likely it's possible to make a whole language this way, or likely that it will break somewhere?

I'm generally interested in any constructive criticism.

[Raphael]

VAVNAC (verbs are vowels, nouns are consonants)
I'm making a new conlang.

Main idea and goals:
The main idea is that nouns are consonants, verbs are vowels. A simple sentence with one verb and one noun just puts the two together by alternating the phonemes, so that if 'mother' is m-m-t and 'sleep' is i-e-u then 'mother sleeps' is mimetu.

Wow, that's one of the most innovative conlanging ideas I've ever heard!

[Man in Space]

Was Iljena an influence?

[Axas mlö]

Raphael: Thanks!

Man in Space: No, I'd never heard of Iljena. Thanks for the link! Now my question of whether another conlang has done something like this is answered : ) I haven't read very much of the Iljena page yet, but already I'm seeing how that pattern (and the Semitic pattern), with some vowel slots empty, leads to a different (better?) sound than the staccato sound of VAVNAC.

Everyone: Is there a simple way to quote two people in one reply?

[Travis B.]

Everyone: Is there a simple way to quote two people in one reply?

When writing a comment there is a quote button that in the topic review below the text box that you can use to create a new quote automatically in the comment you are writing.

[Raphael]

Everyone: Is there a simple way to quote two people in one reply?

When writing a comment there is a quote button that in the topic review below the text box that you can use to create a new quote automatically in the comment you are writing.

Really? I didn't know that. Thank you for telling me!

[alice]

Do you have aesthetic suggestions? Is the main idea sufficiently different from triconsonantal roots to be interesting?

Does it seem likely it's possible to make a whole language this way, or likely that it will break somewhere?

It's a very interesting premise, and I'd certainly suggest seeing how far you can take it. Two things come to mind: do you really intend a strict CV syllable structure, and what do you do about derivations where you have a noun derived from a verb or vice versa? Are there sets of vowel-consonant correspondences which come into play here?

[Emily]

VAVNAC (verbs are vowels, nouns are consonants)
I'm making a new conlang.

Main idea and goals:
The main idea is that nouns are consonants, verbs are vowels. A simple sentence with one verb and one noun just puts the two together by alternating the phonemes, so that if 'mother' is m-m-t and 'sleep' is i-e-u then 'mother sleeps' is mimetu.

I'm content for this language to be a sort of a proof-of-principle rather than a well-fleshed out conlang. One aspect of that: the name of the conlang is just VAVNAC.

One example:
In VAVNAC (pronounced as in IPA)
Wefanə, nuhaʔo vyʔa takazə kəʔeʔo, tiqaʔi nohaʔiʔu viʔy fefubodəsəgə.
Separated into consonant-words and vowel-words, and then the gloss:

wfn\ea, nh\ua-o v\ya tkz\aa k\əeo, tq\iai nh\oa-iu v\iy ffb-dsg\euo.

that.way\INS, I\begin(v.i.)-CAUSATIVE THAT.VERB\PST journey\ACC PREVIOUS.NOUN\long, REL\along I\move-backwards THAT.VERB\PRF chair-king\into.

Translation: It was in this fashion that I began the long journey by which I have backed into the throne.

Notes on the example:
Each word is made by smooshing together a consonant-word and a vowel-word. I've separated them out so you can see the parts better, with a backslash in between. Backslash is not a terrible stretch from https://www.eva.mpg.de/lingua/resources ... -rules.php , rule 4D. That page is also where I'm getting the all-caps glosses when they're abbreviations; and v.t. means the gloss is using the transitive meaning of the English verb, v.i. for intransitive.

When there are more consonants than vowels, the sequence of vowels is padded with ə, which we can consider the neutral vowel - enough ə's are added to match the length of the sequence of consonants. Similarly, the consonants are padded with ʔ.

PREVIOUS.NOUN refers to the immediately previous noun. THAT.VERB refers to a nearby verb, generally the main verb; it fills a noun slot and can be thought of as meaning the action, process, state, or fact of the relevant verb. Probably there's a better name for such a thing? Or a reason organizing it like this isn't a good idea?

More details:
Verbs are vowels, as are adjectives, adverbs, adpositions and things like ACC, and some miscellaneous particles. Nouns are consonants, as are pronouns and some miscellaneous particles.

My idea is inspired a little bit by triconsonantal roots but mainly by a wish to have two streams of information that are somehow separate.

I'm obviously not looking to make something naturalistic, but I don't want it to be vastly more non-naturalistic than is required by the main idea. Or rather, I'm totally fine with it being too regular, I'm resigned to it being too English-like, but I don't want it to be bizarre in ways that no human languages are — except for what follows from the premise. And more positively, I want things to fit well with that premise.

I have more worked out, but I thought maybe I should start with posting just a little.

i like this idea a lot!

Does the main idea already exist, in a natlang or conlang? Are there things about this conlang - apart from the main premise, and necessary consequences - that are non-naturalistic in the sense of being implausibly different from what's attested in natlangs? Is the main idea sufficiently different from triconsonantal roots to be interesting?

i've certainly never heard of anything like this. the comparison to the semitic root system is obvious, but there the changes (which can involve added or deleted vowels and consonants) are used for inflection and derivation, whereas in your process the morphology is an intrinsic part of the syntax!

as far as naturalistic, the basic concept doesn't seem like something that would have evolved naturally. but that's fine!

Does it seem likely it's possible to make a whole language this way, or likely that it will break somewhere?

I'm generally interested in any constructive criticism.

it looks like you've already mostly solved my initial questions, which would be "what about adjectives" and "what about multi-argument verbs/prepositional phrases/etc.", so the only other real stumbling block i see is that there are of course thousands upon thousands of nouns and verbs in any given language, so if this were fleshed out into a full conlang it would need to have a high phoneme inventory or very long words or probably both

[Axas mlö]

When writing a comment there is a quote button that in the topic review below the text box that you can use to create a new quote automatically in the comment you are writing.

I see the button now, thanks. But it doesn't say "Travis B. wrote" with the fancy links and stuff - the stuff that appears if I use the quote button next to your comment. And the fancy links and stuff doesn't seem like something it's easy for me to make.

alice wrote:

It's a very interesting premise, and I'd certainly suggest seeing how far you can take it.

Thanks!

Two things come to mind: do you really intend a strict CV syllable structure,

...yes? I haven't posted the phonology yet, but yes syllables are strictly CV. If you want to suggest I might prefer to do something different than strict CV, I'd be curious to hear it.

I was also thinking earlier that there could be other dialects of this conlang where a noun can be tr-sk- which combines with a verb ai-o-eu to make traiskoʔeu.

and what do you do about derivations where you have a noun derived from a verb or vice versa? Are there sets of vowel-consonant correspondences which come into play here?

Yes, that's a big potential problem! I have notes about derivational morphology which I'll try to tidy up just enough to post. Basically, I'm thinking I have to live without those derivations, and just say things the long way. 'It's a stud finder' -> it\be tool\be REL\find stud\ACC. And where another language might have a noun derived from a verb or adj which is derived from a noun (circle -> circular -> circularity), or rather for common cases of that, have a single noun->noun suffix that does the whole job.

I did think about encoding each vowel as a consonant and each consonant as a pair of vowels. (There are 7 vowels and 21 consonants.) But it seemed like a really big increase in non-naturalism (unnaturalism?) for just one corner of the grammar. If it were only /i/ <--> /j/, /u/ <--> /w/, a correspondence would be fine, but what do I do with consonants like /tdnsz/? I don't have a better idea than Do Without Them. Edit: I meant do without them as in do without derivations of nouns from verbs or vice versa. Though do without /tdnsz/ could be interesting too...

I see that there's at least one more post but I have to go do other things now and I'll reply later.

[malloc]

Cool idea. This is precisely the kind of weird and fantastical thinking I enjoy in conlangs.

...yes? I haven't posted the phonology yet, but yes syllables are strictly CV. If you want to suggest I might prefer to do something different than strict CV, I'd be curious to hear it.

It would certainly help for allowing more possible morphemes, I suppose. If words of a given class can only have consonants or vowels, but not both, that greatly limits the set of possible roots after all.

[Axas mlö]

i like this idea a lot!

Thanks!

as far as naturalistic, the basic concept doesn't seem like something that would have evolved naturally. but that's fine!

Makes sense. I'm going for a theme of one drastic departure from being naturalistic, then try to see what follows.

Does it seem likely it's possible to make a whole language this way, or likely that it will break somewhere?

I'm generally interested in any constructive criticism.

it looks like you've already mostly solved my initial questions, which would be "what about adjectives" and "what about multi-argument verbs/prepositional phrases/etc.",

Thanks! Also interesting is the language iljena that Man in Space linked to above, that approaches the problem by just dispensing with multi-argument verbs.

so the only other real stumbling block i see is that there are of course thousands upon thousands of nouns and verbs in any given language, so if this were fleshed out into a full conlang it would need to have a high phoneme inventory or very long words or probably both

Ahh, I didn't think of that. And it's rather unlikely I'll eventually create thousands of words and find out directly if that's a problem.

I'm not sure how many words there usually are in a natural language - I saw a wide range of numbers on the Internet. Just to take a stab at the problem, let's say 10 000 nouns. There are 21 consonants in VAVNAC, and 21^3 is almost enough, which would be consonant-words of 3 consonants. But it would be awkward to have the space of possible consonant-words filled up. 21^4 is about 190 000, which sounds like enough words, and words that are not too long. Am I going about this in a reasonable way?

If however we want 10 000 verbs and adjectives and such, that's harder. There are 7 vowels in VAVNAC. Vowel-words of 5 vowels gives almost 17 000 possibilities, which seems like barely enough; words of 6 vowels gives over 110 000, which seems like plenty, but those are long words.

...is this why alice wanted to know if I really wanted to do CV?

What's a likely number of nouns and verbs-or-adjectives for a natlang that's not one of those with a very limited set of verbs?

[malloc]

If the language only allows CV syllables, then a large phoneme inventory would certainly make sense. More phonemes would allow more possible roots, compensating for the restricted syllable shapes. Conversely the restrictions on syllables would avoid the problem of complicated phonemes clashing in awkward ways since any two consonants have vowels between them and vice versa.

[Richard W]

I'm not sure how many words there usually are in a natural language - I saw a wide range of numbers on the Internet. Just to take a stab at the problem, let's say 10 000 nouns. There are 21 consonants in VAVNAC, and 21^3 is almost enough, which would be consonant-words of 3 consonants. But it would be awkward to have the space of possible consonant-words filled up. 21^4 is about 190 000, which sounds like enough words, and words that are not too long. Am I going about this in a reasonable way?

If however we want 10 000 verbs and adjectives and such, that's harder. There are 7 vowels in VAVNAC. Vowel-words of 5 vowels gives almost 17 000 possibilities, which seems like barely enough; words of 6 vowels gives over 110 000, which seems like plenty, but those are long words.

You also need to worry about redundancy, or error detection/correction.

One possibility is to reduce the number of morphemes, by making significant use of compounding.

[Axas mlö]

malloc wrote two posts, recommending a large phoneme inventory, and saying:

Conversely the restrictions on syllables would avoid the problem of complicated phonemes clashing in awkward ways since any two consonants have vowels between them and vice versa.

Ah, that makes sense! Thank you. Alright, I will increase the number of phonemes, especially vowels.

BTW malloc, I like your username. .... OK I just found the ZBB Census thread; I agree, it's a great name for a villain.

Richard W wrote:

You also need to worry about redundancy, or error detection/correction.

One possibility is to reduce the number of morphemes, by making significant use of compounding.

Yes, and thank you. I was worrying about redundancy and such, but I see I didn't say so. I was thinking that 10 000 ish words in a space of possibilities allowing say 100 000 words was enough redundancy. But I don't know that, I made those numbers up.

[malloc]

Also, have you considered introducing tone? You could greatly increase the possible vowel distinctions that way or you could even introduce a third dimension alongside consonants and vowels. Perhaps tone could indicate something distinct from verbs or nouns, like case marking or other inflectional categories.

[Travis B.]

Also, have you considered introducing tone? You could greatly increase the possible vowel distinctions that way or you could even introduce a third dimension alongside consonants and vowels. Perhaps tone could indicate something distinct from verbs or nouns, like case marking or other inflectional categories.

One should remember that in many languages, especially African ones, tone is used to express grammatical distinctions independent of lexical ones. This definitely could be another dimension you could add to your language.

[Axas mlö]

Also, have you considered introducing tone? You could greatly increase the possible vowel distinctions that way or you could even introduce a third dimension alongside consonants and vowels. Perhaps tone could indicate something distinct from verbs or nouns, like case marking or other inflectional categories.

One should remember that in many languages, especially African ones, tone is used to express grammatical distinctions independent of lexical ones. This definitely could be another dimension you could add to your language.

I did wonder about using tone for a third stream of information. I didn't think about it much, and I think that for the current project, I'll stick with just two streams. Though I might still use tone like you said, malloc, to increase the effective number of vowels - I'm still working on the new and improved phonology.

Even if I'm not going to do it (now), using tone for grammatical distinctions is a fun idea. If it were only for basically verbal inflectional categories, or basically nominal ones, that would be less fun than your suggestion to mark case which relates the two. (Or I guess we could mark the verb's TAM on the first syllable's tone and the noun's number on the second, or something.) Writing a word as consonants\vowels\tones which represent nouns\verbs\case, we could say I\give\NOM book\give\ACC brother\give\DAT house\give\LOC birthday\give\REASON 'I gave my brother a book for his birthday when we were at home', and all those words would have the same sequence of vowels.

[rotting bones]

I like it. Excited to see more. I would also appreciate details on who speaks it.

[Axas mlö]

I like it. Excited to see more. I would also appreciate details on who speaks it.

Thanks! Here's some current progress...

Approach to conlanging
I've been working on a revised phoneme inventory. I keep feeling it isn't good enough. I found this thread:

That said, the realistic approach, or the Realistic School (I love the idea of belonging to an artistic school!) suits me very well, but I don't think there's anything wrong with other approaches. And I'd love to see other schools put in practice!

Oh indeed! I'd like to see the other schools more in practice too! And I'm of the Irrealist School.

and that was helpful. So I declare that I'm currently working in the One Unusual Thing artistic school, where we (I) make conlangs with one intentionally unusual thing - not naturalistic, or typologically unusual; hopefully not merely unusual from a Eurocentric point of view - and work out consequences.

Phoneme inventory (and a bit of other phonology)
My goal is to have an basically-unremarkable phonology which supports the main idea of interleaving consonants and vowels - so, strictly CV. My other goal is to have more phonemes, especially more vowels, as discussed in some earlier posts.

(Earlier I was using the 26 IPA symbols /a/ through /z/, plus a couple more I liked. Minor question: Is that a common technique for a conlanger whose main interest is elsewhere?)

Here's the current draft for the inventory.

Code: Select all

/p b  t d  tʃ dʒ  k g  q ɢ   ʔ
  m    n     ɲ     ŋ
       ɾ
 f v  s z   ʃ  ʒ  x ɣ  χ ʁ   h
w ɥ         j
       l

i y ɨ u
e ø   o 
ɛ œ ə ɔ
a     ɑ/

My main question for this post: Is this in fact an unremarkable inventory?

(Would you rather I format that list some other way?)

I decided I didn't like vowel length aesthetically. I may still introduce tones and count them as part of vowel-information.

Another question: I like nasalization - but wouldn't that mean, if I'm being naturalistic, that I'd need more rules about how the nasalized vowels interact with surrounding consonants? I'm not quite sure where to find that information.

And I'm also hoping to declare that syllables are strictly CV with no further restrictions; and that whatever other phonological details, assimilation, allophony, etc there might be are nothing that stops me from just writing out my CV syllables.

The first syllable of a word is stressed, which feels right given how padding with neutral sounds works.

Conworld or conculture
I was going to say "none", and it's still not my main interest, but you've already seen I used a sentence from a novel (Book of the New Sun) as my example sentence, and the other sentences in my notes are from other fantasy and science fiction books. So I'm going to say that whatever society speaks this language, they have works of fiction overlapping ours: journeymen in the torturer's guild who back into the throne, rings to rule them all, and so on. If an example sentence is 'March the twenty-third was the Feast of St. Turibius', then St. Turibius, saints, and of course March, are concepts they're comfortable with in fiction, and may or may not have in their actual world.

And I suspect that will make their actual world resemble ours.

[alice]

There's nothing wrong with your inventory, it's just a bit large, but in the circumstances that's probably OK. You might want to tihnk about historical development to justify some of it, though, especially the vowels and minor details like the absence of /w/ and the uvular nasal. Otherwise, just go ahead with it; what you do with these phonemes is going to be much more interesting anyway.

(Just for fun: I'm not the only conlanger who has contemplated using all and only the IPA phonemes corresponding to the 26 letters of the lowercase roman alphabet; but I remember suggesting this on this board many moons ago and someone pointed out that you could also the 26 uppercase letters instead just as well.)