Conlang made from interleaved vowel-words and consonant-words

[Travis B.]

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.

Mind you, lack of /w/ is common if you have /v/ (e.g. many Romance, Germanic, Slavic, and Turkic languages), and many, many languages with uvular stops lack uvular nasals.

[Axas mlö]

General reply: thank you both, I appreciate it!

You might want to tihnk about historical development to justify some of it, though

Ah, OK. Historical development hasn't been my main interest and I only slightly understand it. If I do go that route, how would you recommend dipping my toe in the water? Say, one slightly-earlier stage and one sound change?

the absence of /w/

Hmm, both of you said absence of /w/, when I intended to put it in the inventory, so I suppose I made the chart confusingly.... What happened was that once I had that and /ɥ/ I said "how to I format the chart with co-articulations? OK I'll just stuff them in this corner." But anyway there's definitely a /w/.

(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.)

Ah, that's great.

[rotting bones]

Good work on the phonology. I'm curious to see more details about the morphology. Maybe I missed this, but will there always be /ʔ/ and /ə/ at the ends of compounds and words with affixes, or do the noun and verb patterns fit the specific other they refer to? For example, could fefubodəsəgə have been fəfəbədesugo if the "into" had been relevant to "king" instead of "chair"?

[Neonnaut]

Was Iljena an influence?

I'd never heard of Iljena either, nor that newsletter.

[Axas mlö]

Good work on the phonology. I'm curious to see more details about the morphology. Maybe I missed this, but will there always be /ʔ/ and /ə/ at the ends of compounds and words with affixes, or do the noun and verb patterns fit the specific other they refer to? For example, could fefubodəsəgə have been fəfəbədesugo if the "into" had been relevant to "king" instead of "chair"?

The way I'm thinking of it is that there is a sequence of exactly the same number of vowel-words as consonant-words (some of them can be null), and a compound like ffbdsg 'chair-king' is exactly one consonant-word, the same as if it were unanalysable like English 'throne'. So it matches up with exactly one vowel-word, euo 'into'. The match is always at the beginning, so fefubo..., and the rest of the consonants are made into syllables with ə's, fefubodəsəgə. If somehow 'into' matched more with 'king', there would be no way to express that with the compound 'chair-king'.

Roots are often lengths other than three letters, btw.

But your question makes me wonder if I'm really meeting my conlang design goals. I think there are going to be lots of morphemes filled in with nulls (ʔ and ə), which is less pretty than matching up non-null vowels and consonants. So an answer to your question is there will quite often be ʔ's or ə's at the end of compounds and words with affixes.

[rotting bones]

But your question makes me wonder if I'm really meeting my conlang design goals. I think there are going to be lots of morphemes filled in with nulls (ʔ and ə), which is less pretty than matching up non-null vowels and consonants. So an answer to your question is there will quite often be ʔ's or ə's at the end of compounds and words with affixes.

Thanks. This is not necessarily a problem. Having words often end with certain sounds gives character to a language, right? You can add more character by having things like consonant clusters that don't get vowels between them, consonants that cluster instead of getting null vowels between them, etc. All of this will increase the difficulty of the language.

[Travis B.]

I think it is a good example to have consonant sequences that don't receive schwas between them, and consonant clusters in roots that are not broken up by vowels, to add more variety to the resulting words.

[Richard W]

And you can have clusters that behave like single consonants. /st/ is a good example in English.

[rotting bones]

Diphthongs too.

[Travis B.]

Diphthongs too.

Yes, good idea.

[Axas mlö]

But your question makes me wonder if I'm really meeting my conlang design goals. I think there are going to be lots of morphemes filled in with nulls (ʔ and ə), which is less pretty than matching up non-null vowels and consonants. So an answer to your question is there will quite often be ʔ's or ə's at the end of compounds and words with affixes.

Thanks. This is not necessarily a problem. Having words often end with certain sounds gives character to a language, right?

That's true. Thanks.

You can add more character by having things like consonant clusters that don't get vowels between them, consonants that cluster instead of getting null vowels between them, etc. All of this will increase the difficulty of the language.

I could do that, but I'm realizing my use of "pretty" wasn't clear, and my sense of what's pretty may be very different than other people's. I mainly meant a more abstract kind of pretty, that the structure and organization and logic are pretty, that the pattern is pleasing when contemplated. So, words paired up without too much leftover space. I'm less concerned with it being pretty in the way that Elen síla lúmenn’ omentielvo is supposedly a prettier sequence of phonemes than Uglúk u bagronk sha pushdug Saruman-glob búbhosh skai.

Travis and Richard, this applies to your comments too - not the direction I wanted to go, but the suggestions are nonetheless appreciated.

[Axas mlö]

The general pattern of a sentence or clause

Reminder: word1\word2 means word1 is made of consonants only and word2 vowels only, and they are interleaved to produce what's actually said.

The general pattern of a sentence or clause is

subject\main.verb object\ACC

where the object\ACC is optional; followed by zero or more additional

noun\verb

pairs, some of which may be interpreted as

noun\adposition

and some of which may have THAT.VERB in the noun slot, in which case they may be interpreted as

THAT.VERB\adverb.

Design decisions
The plan of having as many noun\verb pairs as you like is influenced by the pattern in conlangs I've seen with no transitive verbs (Mark Rosenfelder's Bhögetan and Imralu's Balog), and by an idea I read years ago which, tracking it down now, seems to be called "semantic frames" https://en.wikipedia.org/wiki/FrameNet .

Here's an example:
Vyʔa dysugɛqɨɲə døsɔguqəɲə.

v\ya

THAT.VERB\PST

dsgqɲ\yuɛɨ

aristocrat\woke.up

dsgqɲ\øɔu

aristocrat\be.angry

'The aristocrat woke up angry'

Generated with Gloss My Gloss; I'm still not sure I'm using it properly given I have extra information to include. Also, dsgqɲ 'aristocrat' is really dsg-qɲ 'monarch-lesser.thing' and yuɛɨ 'wake up' is really yu-ɛɨ 'be.awake-INCHOATIVE'. But including that made the gloss look even weirder.

The point is, the last noun\verb is extra information and it should be clear how it fits into the sentence.

I can design many sentences to be noun1\verb1 noun2\verb2 instead of noun1\verb1 noun2\ACC, and it doesn't feel awkwardly long (to me) - the reason is that aa 'ACC' is only two vowels, and typical nouns are at least two consonants. So noun\ACC will often be as long as the noun, and often no shorter than noun\verb for a simple verb. I suppose I could make all two-argument sentences work with some verb2, like in monovalent conlangs, but I don't feel like doing that. I'll keep ACC for noun2's that are really patients, and some other cases, but I'll look for opportunities to use a second verb. I'd be interested in hearing about cases where I missed an option that might be worth considering.

Specific constructions

Subjects: So far this pretty much corresponds with English. For transitive VAVNAC verbs, subjects are often agents or experiencers. For intransitive, they can be all sorts of things; and I want a diversity of verbs, to use in several roles in the sentence.

ACC: Nouns appearing with ACC are patients, and anything that feels like an object if I can't find another role for it. Realistically, "feels like an object" means nouns that would be an object in English.

The same verb twice: For mutual or symmetric expressions. Just repeat the main verb for the second noun.

Nɑhɔ dɑsɔgə.

nh\ɑɔ

I\be

dsg\ɑɔ.

monarch\be.

'I am the king' (or queen, autarch, etc)

Poqukə. ʔuʔi tanəʔa kapəʔa kaʔuʔo.

pqk\ou.

wall\exist.

0\ui

0\not

tn\aəa

it\resemble

kp\aəa

thing\resemble

k\auo.

PREVIOUS.NOUN\big.

'There was a wall. It did not look important.' ('It did not resemble an important thing.') 'Resemble' is repeated.

The same pattern of repeating the verb applies for 'meet', 'cooperate', etc. It's also to make something reciprocal that isn't by default:

Nɨhø xɨŋø.
nh\ɨø xŋ\ɨø.
I\hate you.SG\hate.
'You and I hate each other.'
(Maybe I'll skip any fancy formatting for glosses sometimes.)

The above sentence is ambiguous - it could also mean 'You and I both hate (someone).'

Verbs that come in pairs:

• 'be above' and 'be below'
  • Tʃɔɥɑto winyɾɨ.
    ʧɥt\ɔɑo wnɾ\iyɨ
    Mouse\be.below table\be.above. 'The mouse is under the table.'
  • Tʃiɥytɨ wɔnɑɾo.
    ʧɥt\iyɨ wnɾ\ɔɑo
    Mouse\be.above table\be.below. 'The mouse is on or over the table.'
• iœ 'be inside, be contained' and oɑ 'be outside, be a container, contain'
• some other words for location, e.g. 'be to the left' and 'be to the right', come in pairs, although some, like 'be to the side', follow the same-verb-twice pattern.

Stimulus and related concepts:

Nehaʔɑ ɲutyŋə.
nh\eaɑ ɲtŋ\uy
I\see Moon\STIMULUS. 'I see the Moon.'

The vowel-word uy, glossed as STIMULUS, is a verb, loosely translated as 'be the subject matter', and it applies to more abstract situations too:

Nehoʔœ ɲutyŋə.
nh\eoœ ɲtŋ\uy
I\think Moon\STIMULUS. 'I think about the Moon.'

The semantic role for 'think' might be called "theme", or one kind of theme, rather than "stimulus"; but online materials on the semantic roles are confusing to me and possibly not consistent with each other. In any case, the VAVNAC word applies to the subject matter of physical senses, thoughts, emotions, etc.

Many of these verbs can be interpreted as adpositions: Certainly uy 'STIMULUS' or 'be the subject matter' is functioning in an adposition-like way. So is ea 'INSTRUMENTAL' or 'be the tool or means' (from the first post in this thread). The location words mentioned above also fit:

Tʃiɥetu wɔnɑɾo.
ʧɥt\ieu wnɾ\ɔɑo
Mouse\sleep table\be.above.
'The mouse is sleeping under the table.'

(I'm not sure if the above "adpositions" might really be cases, but I'm choosing to analyze them as adpositions. In particular, adpositions are a kind of verb here, though not sharply distinguished from other verbs. Therefore, ACC is not an adposition, since it doesn't make sense for it to be a verb.)

But interpreting them as adpositions results in confusing glosses: We can gloss ɔɑo with an English verb phrase 'be above' in Mouse\sleep table\be.above. But viewing it as an adposition, it corresponds to 'under' in 'The mouse is sleeping under the table' and we could gloss it from that POV as Mouse\sleep table\under. Which is what I was doing before I realized adpositions can be seen as just verbs. And the two glosses are basically opposites.... I'm planning to use the verb approach from now on, as in Mouse\sleep table\be.above.

More about motion: We can have

Tʃoɥatɑʔɨ wɔnɑɾo.
ʧɥt\oaɑɨ wnɾ\ɔɑo
Mouse\run table\be.above.

I'm thinking that will mean 'The mouse runs while under the table' and we'll need a different sentence for 'The mouse runs, ending under the table' and for 'The mouse runs, starting under the table'. I'm leaning towards using kz 'endpoint' to make some longer construction, probably -

Tʃoɥatɑʔɨ, kɔzɑʔo wɔnɑɾo.
Mouse\run, endpoint\be.below table\be.above

and similarly for tk 'starting point'. I'm pretty sure I don't want a single word, like English 'into' or 'onto', for combining spatial relation and destination. Of course, that's what I did in the first post in this thread - for motion ending in the throne, I used euo 'into'. Incidentally, that particular sentence would be fine with a word for 'to' rather than 'into', but more broadly, it's feeling awkward to have a word that's going to have to be a verb meaning 'X is the container of the endpoint of a motion'. I'd be interested in suggestions on this topic (and in general), though I might end up just doing what I already said.

[sasasha]

I'd be interested in suggestions on this topic (and in general), though I might end up just doing what I already said.

Interesting project.

I think your solution to the motion example is fine, although it adds to your sentences having a fairly homogenous structure, which you may or may not like. A few suggestions:

1. Many languages (including English) permit ambiguity in these kinds of situations, and rely on context to disambiguate. “The mouse ran under the table” could be applied to all three of your directional cases. Perhaps you don’t need the level of specificity you are seeking (at least, most of the time)?

2. You could focus on the result: Mouse\run table\be.above exposure\be, Mouse\run table\be.above security\be

3. You could work in tense or aspect: Mouse\run table\was.above, Mouse\run table\will.be.above

4. You could create a class of adverbials that mean ‘out’, ‘in’ etc. I would suggest not being too frightened to create a special word class, even if you have a strong feeling about trying to express things through intransitive verbs: Lojban would be much more monotonous without the attitudinals, for instance.

[Imralu]

I made one like this back in 2017, Letaale, inspired by both Iljena and Abakwi and with my usual thing of not having a noun-verb distinction, so all roots have a noun form (triconsonantal) and a verb form (trivocalic) with the noun form being equivalent to the agent noun of the verb form and the verb form being equivalent to the noun form plus "be". They are alternated through a series of correspondences between the consonants and vowels.