Conlang Random Thread

[missals]

Now, I’ve been thinking about my #1 a bit more, and I realised that the systems in Spanish, Ancient Greek etc. weren’t quite what I had in mind. In those languages, the article agrees with the noun in number, but number is primarily marked as an affix on the noun. But in the system I was thinking about, number is marked only on the article. So e.g. the systems I’ve been given do something like ‘DEF.PLU person-PLU’ for the people, but I was thinking about doing ‘DEF.PLU person’, where the only marker of plurality is the article. Does this system sound plausible as well?

I'm about to be a little snippy here, but you're definitely overthinking this. This is, more or less, exactly what French does, and it's exactly what dozens of other minor Romance varieties do. And in fact it's what dozens upon dozens of languages, if not hundreds, do all over the world. Just offhand I recall that Maori (and probably every other Polynesian language) does this, having the definite articles te (singular) and ngā (plural), but no number marking on nouns themselves. It's totally normal for common nominal categories to be marked on an agreeing element but not the noun itself. German approaches such a system, relying heavily on the articles to mark case/number.

[bradrn]

Now, I’ve been thinking about my #1 a bit more, and I realised that the systems in Spanish, Ancient Greek etc. weren’t quite what I had in mind. In those languages, the article agrees with the noun in number, but number is primarily marked as an affix on the noun. But in the system I was thinking about, number is marked only on the article. So e.g. the systems I’ve been given do something like ‘DEF.PLU person-PLU’ for the people, but I was thinking about doing ‘DEF.PLU person’, where the only marker of plurality is the article. Does this system sound plausible as well?

I'm about to be a little snippy here, but you're definitely overthinking this.

For me, that’s hardly unusual.

This is, more or less, exactly what French does, and it's exactly what dozens of other minor Romance varieties do. And in fact it's what dozens upon dozens of languages, if not hundreds, do all over the world. Just offhand I recall that Maori (and probably every other Polynesian language) does this, having the definite articles te (singular) and ngā (plural), but no number marking on nouns themselves. It's totally normal for common nominal categories to be marked on an agreeing element but not the noun itself. German approaches such a system, relying heavily on the articles to mark case/number.

That’s really helpful — thank you!

[jal]

German approaches such a system, relying heavily on the articles to mark case/number.

Though I totally agree with your other observations, I think this one is pretty false. German plurals are always marked, often even in good Germanic fashion using ablaut/umlaut, and there's still some case marking on the nouns as well.


JAL

[Esneirra973]

What orthography could I use for this vowel system? I was thinking of using tildes to indicate nasalization, but I am not sure how I could write /ɨ/ and /ɨ̃/. I was thinking of using <v> and <ṽ> respectively, but I don't think any language uses those letters for those sounds.

/i ɨ u e o a/
/ĩ ɨ̃ ũ ã/

[Pabappa]

Why not y? Its in common use elsewhere. Unless you need that for the IPA /j/.

a e i o u y

ã ĩ ũ ỹ

Sure, the last one looks a bit weird in some fonts, but ṽ probably does too.

Ive also used ə for the IPA /ɨ/ but only when that vowel is cognate to a true IPA /ə/ in related languages. Helps keep the etymologies together.

[Esneirra973]

Why not y? Its in common use elsewhere. Unless you need that for the IPA /j/.

a e i o u y

ã ĩ ũ ỹ

Sure, the last one looks a bit weird in some fonts, but ṽ probably does too.

Ive also used ə for the IPA /ɨ/ but only when that vowel is cognate to a true IPA /ə/ in related languages. Helps keep the etymologies together.

I would, but I'm already using <y>. Also, I already decided on using this (got it from a different forum):

i ı u e o a
į ı̨ ų ą

[Pabappa]

Ive been working on Icecap Moonshine lately, a very high-effort and low-output conlang that I compare to adult coloring books ... it requires so much of my mental acuity that I cant focus on other things and therefore it helps me push away stressful thoughts. It is really very unimpressive .... the sound changes go wild, so basically I can get away with just about anything, such as /hə ŋŭku čĭ gə/ > kňáš where not even one phoneme matches. It takes a lot of effort to coin each new word, but the result is hardly distinguishable from what I would have if I literally just made every word up with no attention to etymology.

[bradrn]

What orthography could I use for this vowel system? I was thinking of using tildes to indicate nasalization, but I am not sure how I could write /ɨ/ and /ɨ̃/. I was thinking of using <v> and <ṽ> respectively, but I don't think any language uses those letters for those sounds.

/i ɨ u e o a/
/ĩ ɨ̃ ũ ã/

I know you’ve already decided, but I’d probably do something like:

/i ɨ u e o a/ ⟨i ɨ u e o a⟩
/ĩ ɨ̃ ũ ã/ ⟨ĩ ẽ ũ ã⟩

[Imralu]

German approaches such a system, relying heavily on the articles to mark case/number.

Though I totally agree with your other observations, I think this one is pretty false. German plurals are always marked, often even in good Germanic fashion using ablaut/umlaut, and there's still some case marking on the nouns as well.

The majority of masculine and neuter nouns that end in -er, -el or -en are unmarked (except of course for the -s in sg. genitive and the -n for pl. dative if applicable).

das Messer - the knife
die Messer - the knives

der Fahrer - the driver
die Fahrer - the drivers

Just offhand I recall that Maori (and probably every other Polynesian language) does this, having the definite articles te (singular) and ngā (plural), but no number marking on nouns themselves.

There's a small group of common nouns that do mark singular and plural.

te tangata - the person
ngā tāngata - the people

te wahine - the woman
ngā wāhine - the women

te tamaiti - the child
ngā tamariki - the children

There are a few more, mostly words for family members.

[bradrn]

Just offhand I recall that Maori (and probably every other Polynesian language) does this, having the definite articles te (singular) and ngā (plural), but no number marking on nouns themselves.

There's a small group of common nouns that do mark singular and plural.

te tangata - the person
ngā tāngata - the people

te wahine - the woman
ngā wāhine - the women

te tamaiti - the child
ngā tamariki - the children

There are a few more, mostly words for family members.

Interesting! I might do that as well.

But I can’t help but wonder: why is it family members specifically which get inflected for number? I cannot understand why a language would want to do that. As EastOfEden said (about Sakao, another Austronesian language):

- Common nouns are not inflected for number, but for some reason kinship terms are, because apparently it's really important to know if I'm talking about my aunt (/ðjœɣ/) or my aunts (/rðjœɣ/).

[Imralu]

Interesting! I might do that as well.

But I can’t help but wonder: why is it family members specifically which get inflected for number? I cannot understand why a language would want to do that.

It might just be because they're common words, and then it's regularised a bit by analogy. The word whaea 'mother', 'aunt' has the plural form whāea which is used by some tribes, but not everyone, which suggests to me that analogy with other kinship terms has caused the change.

[bradrn]

Interesting! I might do that as well.

But I can’t help but wonder: why is it family members specifically which get inflected for number? I cannot understand why a language would want to do that.

It might just be because they're common words, and then it's regularised a bit by analogy. The word whaea 'mother', 'aunt' has the plural form whāea which is used by some tribes, but not everyone, which suggests to me that analogy with other kinship terms has caused the change.

Good point — analogy can do strange things (cf. triconsonantal systems). And if you start with only a handful of common words which mark number, and these happen to include some kinship words, the forces of analogy are certainly powerful enough to extend that to all kinship terms.

[Kuchigakatai]

It might just be because they're common words, and then it's regularised a bit by analogy. The word whaea 'mother', 'aunt' has the plural form whāea which is used by some tribes, but not everyone, which suggests to me that analogy with other kinship terms has caused the change.

Good point — analogy can do strange things (cf. triconsonantal systems). And if you start with only a handful of common words which mark number, and these happen to include some kinship words, the forces of analogy are certainly powerful enough to extend that to all kinship terms.

Or maybe, it's just because they're human nouns, specifically the most important human nouns. Recall the number hierarchy at the end of WALS chapter 14:
speaker > addressee > 3rd person > kinship terms > other humans > “higher” animals > “lower” animals > discrete inanimates > nondiscrete inanimates

Further examples:

- Mandarin has a plural suffix that only attaches to 1) the personal pronouns (basically obligatorily), and 2) common human nouns (optionally) as long as they're definite, i.e. specific in context (e.g. "the people (that will see this, who were involved...)", "the managers", "the moderators").

- Modern Standard Arabic adjectives agree in gender with their noun if the noun is rational (human or divine). If it is non-rational, then a further form is used (identical with the feminine singular) which does not distinguish gender.

fan:a:nu:na dʒududun ("male.artist.PL new.PL.MASC") 'new male artists'
fan:a:na:tun dʒadi:da:tun ("female.artist.PL new.PL.FEM") 'new female artists'

suqu:f dʒadi:datun ("ceiling.PL new.PL.INAN") 'new ceilings' (saqfun 'ceiling' is a masculine noun)
ʔardˤi:ja:tun dʒadi:datun ("floor.PL new.PL.INAN") 'new floors'
ʔardˤi:jatun dʒadi:datun ("floor.SG new.SG.FEM") 'a new floor'
fan:a:natun dʒadi:datun ("female.artist.SG new.SG.FEM") 'a new female artist'

- Mwotlap uses articles to make number distinctions (dual, trial, plural) in human nouns, but nouns lower than humans in the hierarchy are ambiguous unless followed by a numeral (two, four, seven...).

wɛlan 'a/the chief', jɔɣɛ wɛlan '(the) two chiefs', tɪlɣɛ wɛlan '(the) three chiefs', iɣɛ wɛlan '(the) 4+ chiefs'
na-pnō 'a/the village/villages' (and then vʊnʊ vʊjʊ 'two villages', vʊnʊ vɪtɪl 'three villages'...)

[bradrn]

It might just be because they're common words, and then it's regularised a bit by analogy. The word whaea 'mother', 'aunt' has the plural form whāea which is used by some tribes, but not everyone, which suggests to me that analogy with other kinship terms has caused the change.

Good point — analogy can do strange things (cf. triconsonantal systems). And if you start with only a handful of common words which mark number, and these happen to include some kinship words, the forces of analogy are certainly powerful enough to extend that to all kinship terms.

Or maybe, it's just because they're human nouns, specifically the most important human nouns. Recall the number hierarchy at the end of WALS chapter 14:
speaker > addressee > 3rd person > kinship terms > other humans > “higher” animals > “lower” animals > discrete inanimates > nondiscrete inanimates

I don’t believe it — I wrote a whole post about it, but I somehow still managed to forget about the animacy hierarchy‽ I feel stupid now.

[Knit Tie]

Are there sny languages with a voicing distinction where the voiced series is less marked and/or more numerous?

[Pabappa]

Bantu languages at least since proto-Bantu have typically featured voiced stops that are more common than their respective voiceless stops. On top of that, they also often have voiced prenasalized consonants. Its worth note that there are no approximant consonants or fricatives in proto-Bantu, and that some scholars project the language back to an earlier stage where there were no nasals (but there might hhave been sonorants), which could have led to this unusual setup.

Some early IE languages might count, and if you call /bh dh gh/ voiced sounds, PIE itself might count. I suspect the most common stop series in PIE was the voiceless one, but even so, its possible that the voiced and voiced aspirated series together would still outnumber t he voiceless series. Its worth noting though that some languages, such as Scottish Gaelic, have undergone a devoicing shift such that what they write as b d g are phonetically /p t k/ and the corresponding voiceless symbols now stand for voiceless aspirated stops.

[Kuchigakatai]

Persian seems to use voiced consonants pretty heavily statistically speaking, even though it has /p t tʃ k f s ʃ x/ with their voiced counterparts. I would expect that if someone took a look at how often voiceless vs. voiced obstruents appear in English, Spanish or Russian vs. Persian, Persian would turn out to use voiced obstruents a lot more often.

[mèþru]

I'm doing a collaborative conworld for a kind of Particracy/Nation States game I want to make public on Discord. If anyone wants to have a role in creating that world, PM me and I might invite you to the server. If I reject you it's only because I'm trying to keep it small and safe to start with, it's not that you are an awful person or that I hate you.

[Qwynegold]

Has anyone ever posted a sentence from their conlang into Google translate and see what language it guesses it to be?

Yeah, I do that sometimes. Just now I tried entering this into it:

Kutsonga myosyak'yattong. Rūjungodu kurwa k'ūding sūnnat'tūk. Yazokkak' rallūllado kyot'sōk'yang. Tille tashp'okto sungkittige kūjungobak'p'o. K'ūju pofōbak', huga pot'a zhok'p'ūp'so kūjot udittong, mōllussat kangakka atk'o.

And it thought it was Filipino. :S Is it because of all the <ng>? Because nothing else here looks Filipino.

EDIT: Whoops! I should check the date of posts before replying to them.

[bradrn]

I am trying to work out some details of the phonology of my current conlang, but I am running into some problems. Basically, I want my conlang to have two particular features:

1. I want it to have a minimal word size: all phonological words must be at least 2 moras. But I would like to allow monosyllable CVC words, so I must treat a coda consonant as being +1μ.
2. I want stress to be mostly on the first syllable. In particular, CVCVC words are very common, so I particularly want those to be stressed on the first syllable. But I want CVCVː words, ending with a long vowel, to be stressed on the second syllable, so I need some sort of weight-sensitive stress. I can get this particular stress pattern by not counting coda consonants for syllabic weight.

You see the problem? If I count coda consonants as giving weight, then I get to have monosyllabic CVC words, but CVCVC words get stressed on the second syllable rather than the first. (It’s probably worth noting here that all syllables are CV(C), with an obligatory onset, so I can’t just ‘cheat’ by syllabifying it as CVC.VC.) On the other hand, if I don’t count coda consonants for weight, then I get my CV́CVC words, but I can’t have monosyllabic CVC words anymore because they are now only 1μ rather than 2μ.

At the moment, I can see two possible resolutions:

• As suggested by WALS, there are a few languages which treat only some coda consonants as having weight. That would let me have monosyllabic CVC′ when C′ is a heavy consonant, but also CV́CVC′ when C′ is a light consonant. My problem with this is that WALS is regrettably vague about which criteria I may use to determine coda consonant weight, and I can’t find anywhere else which mentions this, so I have practically no idea how I would go about doing this. (e.g. can I say that only obstruents make a syllable heavy? How about only voiced consonants?)
• Alternatively, I can just make my stress pattern weight-insensitive and always stress the first syllable, while retaining my minimal word size constraint. This gets me most of the way to my goal, but also means that any CVCVː words (which admittedly are a minority) will get stressed on the first syllable rather than the second. Besides, I’m not even sure whether it’s plausible to have a mora-based minimal word size constraint, but then not use moras or weight-sensitivity anywhere else in the language.

So at this point, I think I’m a bit stumped… Does anyone else have any ideas?