Which language? English? In mathematical proofs, one often finds 'but' where 'and' would be more natural; the use of 'but' reflects the use of sed 'but' in the Latin forms of proofs.
Conlang Random Thread
Re: Conlang Random Thread
Re: Conlang Random Thread
Never heard of this use in English and it sounds either clumsy or rigidly formulaic/fixed. Or a shortened form of "but also"
But I'm thinking of autoantonymous but/and in normal, non mathematical, usage. And I think I saw it in a Mesopotamian (i.e. Hurro-Urartian) or Semitic language.
- Glass Half Baked
- Posts: 104
- Joined: Fri Mar 06, 2020 6:16 am
Re: Conlang Random Thread
Perhaps we could add to the heap the practice of US legal English using "whereas" to mean "here begins a sentence."
Re: Conlang Random Thread
I had the impression that in US legal English, "whereas" means basically "because".Glass Half Baked wrote: ↑Sun May 26, 2024 6:22 pm Perhaps we could add to the heap the practice of US legal English using "whereas" to mean "here begins a sentence."
- Glass Half Baked
- Posts: 104
- Joined: Fri Mar 06, 2020 6:16 am
Re: Conlang Random Thread
Superficially, sure. But it is used to introduce any information considered pertinent, whether there is a clear causal relationship or not. Basically, law makers use it at the beginning of a law as bullet points, to list their motivations for making the law or beliefs about the law. The semantic content of the conjunction is heavily bleached.
Re: Conlang Random Thread
Reexploring having object suffixes in the verb for probably the hundredth time.
Always they look good in some combinations and ugly in others. And it's hard to make them look good while following my constraints, such as always having penultimate stress on the verb.
I wonder if I can have it that the jussive/desirative form can be unable to take object suffixes even if they're transitive.
Subject:
Object:
Example with 3rd person subjects:
They look ok when presented like this, but I'm still not sure. Then you also have problems with words having too many geminate consinants:
irrabbattánnū iddā́na
[ʔir.ˌrɑb.bɑt.ˈtɑ́n.nuː ʔid.ˈdɑ̂ː.nɑ]
id-0-rabbad-tan-ū idd-āna
REL-REAL\guard\CAUS-3pl>2sg DEM.DIST-INAN.ERG/INS.PL
"[they] who made you guard those things"
That is four series of geminated consonants! I don't know if that is naturalistic to have so many.
And this is why I want to disallow the jussive/desirative from taking object suffixes because it combined with the causative stem fucks up my placement of secondary stress:
irrabbadattánnū iddā́na
ʔir.ˌrɑb.bɑ.dɑt.ˈtɑ́n.nuː ʔid.ˈdɑ̂ː.nɑ]
id-rabbad-aC-tan-ū idd-āna
REL-guard\CAUSE-JUSS-3pl>2sg DEM.DIST-INAN.ERG/INS.SG
"[they] who must make you guard those things"
Always they look good in some combinations and ugly in others. And it's hard to make them look good while following my constraints, such as always having penultimate stress on the verb.
I wonder if I can have it that the jussive/desirative form can be unable to take object suffixes even if they're transitive.
Subject:
More: show
More: show
More: show
They look ok when presented like this, but I'm still not sure. Then you also have problems with words having too many geminate consinants:
irrabbattánnū iddā́na
[ʔir.ˌrɑb.bɑt.ˈtɑ́n.nuː ʔid.ˈdɑ̂ː.nɑ]
id-0-rabbad-tan-ū idd-āna
REL-REAL\guard\CAUS-3pl>2sg DEM.DIST-INAN.ERG/INS.PL
"[they] who made you guard those things"
That is four series of geminated consonants! I don't know if that is naturalistic to have so many.
And this is why I want to disallow the jussive/desirative from taking object suffixes because it combined with the causative stem fucks up my placement of secondary stress:
irrabbadattánnū iddā́na
ʔir.ˌrɑb.bɑ.dɑt.ˈtɑ́n.nuː ʔid.ˈdɑ̂ː.nɑ]
id-rabbad-aC-tan-ū idd-āna
REL-guard\CAUSE-JUSS-3pl>2sg DEM.DIST-INAN.ERG/INS.SG
"[they] who must make you guard those things"
Re: Conlang Random Thread
Same here more or less. One of the big challenges in my polysynthetic conlang is getting the inflectional suffixes to fit together elegantly. Likewise with stringent phonological constraints that limit one's options for designing affixes.
You could introduce rules degeminating consonants in particular contexts. Consider the phenomenon of consonant gradation in Finnish, which often involves geminate consonants shortening before closed syllables. It seems plausible that something like avoiding stress clash could lead to geminate consonants becoming singletons to avoid too many sequential heavy syllables. That said, I see nothing inherently unnaturalistic about sequences of geminates.They look ok when presented like this, but I'm still not sure. Then you also have problems with words having too many geminate consinants...That is four series of geminated consonants! I don't know if that is naturalistic to have so many.
Mureta ikan topaasenni.
Koomát terratomít juneeratu!
Shame on America | He/him
Koomát terratomít juneeratu!
Shame on America | He/him
Re: Conlang Random Thread
My Wēchizaŋkəŋ solves this particular problem by being intensely fusional, such that the phonetic rules give it a coherent aesthetic no matter what I do with individual affixes.
(On the other hand, Eŋes just embraces the ugliness, though it’s debatable whether it’s polysynthetic at all.)
Note that degemination is only one part of a larger system of consonant mutation (see Merrill for a succinct overview), a possibility which is always worth looking into for affix-dependent euphony — it’s a big part of what makes Wēchizaŋkəŋ so distinctive, alongside intervocalic lenition and extensive stress-dependent syncope.You could introduce rules degeminating consonants in particular contexts. Consider the phenomenon of consonant gradation in Finnish, which often involves geminate consonants shortening before closed syllables. It seems plausible that something like avoiding stress clash could lead to geminate consonants becoming singletons to avoid too many sequential heavy syllables. That said, I see nothing inherently unnaturalistic about sequences of geminates.They look ok when presented like this, but I'm still not sure. Then you also have problems with words having too many geminate consinants...That is four series of geminated consonants! I don't know if that is naturalistic to have so many.
Conlangs: Scratchpad | Texts | antilanguage
Software: See http://bradrn.com/projects.html
Other: Ergativity for Novices
(Why does phpBB not let me add >5 links here?)
Software: See http://bradrn.com/projects.html
Other: Ergativity for Novices
(Why does phpBB not let me add >5 links here?)
Re: Conlang Random Thread
What I have done with Fad'ami is first to give each morpheme a number of different forms depending on what precedes it phonetically, and secondly to primarily put things like geminates, aspirated consonants, and ejectives in roots, and avoid them in affixes (there are a couple notable ejectives in affixes, but that is about it).
Yaaludinuya siima d'at yiseka wohadetafa gaare.
Ennadinut'a gaare d'ate eetatadi siiman.
T'awraa t'awraa t'awraa t'awraa t'awraa t'awraa t'awraa.
Ennadinut'a gaare d'ate eetatadi siiman.
T'awraa t'awraa t'awraa t'awraa t'awraa t'awraa t'awraa.
Re: Conlang Random Thread
My person-marking suffixes are basically this (vertical row: 1st, 2nd, 3rd; horizontal column: sg and pl):
Affirmative Intransitive Subject:
-ni / -ti
-ma / -man
-ta / -tan
Negative Intransitive Subject:
-nisi / -tisi
-masi / -masin
-tasi / -tasin
Affirmative Transitive Subject:
-nīy- / -tīy-
-māy- / -mann-
-tāy- / -tann-
Negative Transitive Subject:
-nīs- / -tīs-
-mās- / -mass-
-tās- / -tass-
Transitive Object:
-in / -it
-ū / -un
-ā / -an
Problem is, this means that stress falls on the transitive subject suffixes, which ends up making my verbs sound clunky and unrhythmic when adding the mood prefixes and conjunctive clitics, this is more prominent in the causative jussive.
Here you have examples of verb forms with stress placement indicated. Acute for primary stress, grave for secondary stress.
Without object suffixes:
With object suffixes:
Exampmes 1a, 1b, and 2e to 2h feel janky to me. And 2a to 2d feel janky when the relativizing prefix iC- is added or the conjunctive clitics (e.g. ur= "and").
Ultimately I reason my issue has more to do with that I'm more comfortable with trochaic rhythmic feet and not iambic rhythmic feet.
https://wals.info/chapter/17
I don't know what I can do about this since consonant and vowel length have a lot of functional load. And I'm not sure if it's natural to have it that verbs with a conjunctive clitic attached cannot take object suffixes/clitics. Or to make it so jussives cannot take object suffixes/clitics.
A possible avenue is to shorten or lengthen the transitive subject suffixes depending on rhythm.
Affirmative Intransitive Subject:
-ni / -ti
-ma / -man
-ta / -tan
Negative Intransitive Subject:
-nisi / -tisi
-masi / -masin
-tasi / -tasin
Affirmative Transitive Subject:
-nīy- / -tīy-
-māy- / -mann-
-tāy- / -tann-
Negative Transitive Subject:
-nīs- / -tīs-
-mās- / -mass-
-tās- / -tass-
Transitive Object:
-in / -it
-ū / -un
-ā / -an
Problem is, this means that stress falls on the transitive subject suffixes, which ends up making my verbs sound clunky and unrhythmic when adding the mood prefixes and conjunctive clitics, this is more prominent in the causative jussive.
Here you have examples of verb forms with stress placement indicated. Acute for primary stress, grave for secondary stress.
Without object suffixes:
More: show
More: show
Ultimately I reason my issue has more to do with that I'm more comfortable with trochaic rhythmic feet and not iambic rhythmic feet.
https://wals.info/chapter/17
I don't know what I can do about this since consonant and vowel length have a lot of functional load. And I'm not sure if it's natural to have it that verbs with a conjunctive clitic attached cannot take object suffixes/clitics. Or to make it so jussives cannot take object suffixes/clitics.
A possible avenue is to shorten or lengthen the transitive subject suffixes depending on rhythm.
-
- Posts: 288
- Joined: Wed Jul 08, 2020 9:15 am
Re: Conlang Random Thread
Have you heard of paradigm uniformity effects? Sometimes inflected forms preserve the stress of the base form even if it contradicts the regular stress assigment found in other forms. This makes the overall system less regular but I think it could help solve the particular problem you have.
Re: Conlang Random Thread
I just posted this on Reddit so I may as well put it here too:
Is there a natlang precedence for the sound change /ɸ/ > /u/? Changes of velars like /x ɣ/ > /i/ seem well attested so why not a labial > rounded vowel, particularly in the environment VɸC? It seems like a good way of making new diphthongs.
Is there a natlang precedence for the sound change /ɸ/ > /u/? Changes of velars like /x ɣ/ > /i/ seem well attested so why not a labial > rounded vowel, particularly in the environment VɸC? It seems like a good way of making new diphthongs.
Unsuccessfully conlanging since 1999.
- Rounin Ryuuji
- Posts: 2994
- Joined: Wed Dec 23, 2020 6:47 pm
Re: Conlang Random Thread
I know that the second element in Ancient Greek diphthongs such as /au/ yield /av~af/ in the modern language, so I don't see why that couldn't happen in reverse. Note also Spanish bautismo.Jonlang wrote: ↑Tue Jun 04, 2024 2:17 pm I just posted this on Reddit so I may as well put it here too:
Is there a natlang precedence for the sound change /ɸ/ > /u/? Changes of velars like /x ɣ/ > /i/ seem well attested so why not a labial > rounded vowel, particularly in the environment VɸC? It seems like a good way of making new diphthongs.
Re: Conlang Random Thread
I wasn't aware of that. I've also been informed that it occurred in Japanese: oto-pito > otoɸito > otoɸto > otouto ‘little brother’.Rounin Ryuuji wrote: ↑Tue Jun 04, 2024 3:17 pmI know that the second element in Ancient Greek diphthongs such as /au/ yield /av~af/ in the modern language, so I don't see why that couldn't happen in reverse. Note also Spanish bautismo.Jonlang wrote: ↑Tue Jun 04, 2024 2:17 pm I just posted this on Reddit so I may as well put it here too:
Is there a natlang precedence for the sound change /ɸ/ > /u/? Changes of velars like /x ɣ/ > /i/ seem well attested so why not a labial > rounded vowel, particularly in the environment VɸC? It seems like a good way of making new diphthongs.
Unsuccessfully conlanging since 1999.
- Rounin Ryuuji
- Posts: 2994
- Joined: Wed Dec 23, 2020 6:47 pm
Re: Conlang Random Thread
I've never seen it given that way before, but it could have happened thus; I normally see an intermediary stage /otofuto/ (/pi/ > /fu/ is a common onbin in Kansai, I believe), but /otofto/ is certainly plausible.Jonlang wrote: ↑Tue Jun 04, 2024 3:27 pmI wasn't aware of that. I've also been informed that it occurred in Japanese: oto-pito > otoɸito > otoɸto > otouto ‘little brother’.Rounin Ryuuji wrote: ↑Tue Jun 04, 2024 3:17 pmI know that the second element in Ancient Greek diphthongs such as /au/ yield /av~af/ in the modern language, so I don't see why that couldn't happen in reverse. Note also Spanish bautismo.Jonlang wrote: ↑Tue Jun 04, 2024 2:17 pm I just posted this on Reddit so I may as well put it here too:
Is there a natlang precedence for the sound change /ɸ/ > /u/? Changes of velars like /x ɣ/ > /i/ seem well attested so why not a labial > rounded vowel, particularly in the environment VɸC? It seems like a good way of making new diphthongs.
Re: Conlang Random Thread
Middle Japanese /ɸ/ was rather unstable, and while it regularly became Modern Japanese [h] except before /i/ or /ɯ/, in many words it was simply lost. I have never seen it posited that Japanese ever had /ɸC/ clusters, though.Rounin Ryuuji wrote: ↑Tue Jun 04, 2024 4:01 pmI've never seen it given that way before, but it could have happened thus; I normally see an intermediary stage /otofuto/ (/pi/ > /fu/ is a common onbin in Kansai, I believe), but /otofto/ is certainly plausible.Jonlang wrote: ↑Tue Jun 04, 2024 3:27 pmI wasn't aware of that. I've also been informed that it occurred in Japanese: oto-pito > otoɸito > otoɸto > otouto ‘little brother’.Rounin Ryuuji wrote: ↑Tue Jun 04, 2024 3:17 pm
I know that the second element in Ancient Greek diphthongs such as /au/ yield /av~af/ in the modern language, so I don't see why that couldn't happen in reverse. Note also Spanish bautismo.
Yaaludinuya siima d'at yiseka wohadetafa gaare.
Ennadinut'a gaare d'ate eetatadi siiman.
T'awraa t'awraa t'awraa t'awraa t'awraa t'awraa t'awraa.
Ennadinut'a gaare d'ate eetatadi siiman.
T'awraa t'awraa t'awraa t'awraa t'awraa t'awraa t'awraa.
- Rounin Ryuuji
- Posts: 2994
- Joined: Wed Dec 23, 2020 6:47 pm
Re: Conlang Random Thread
I cound see /ɸC/ > /CC/ happening very quickly, though, since something like this probably lead to the -う (-u), -つ (-tsu), -る (-ru) Godan verbs having their te-form in って (-tte). This seems to have mostly only occurred in longer words, and there isn't quite a consistent pattern, the expected outcome of Old Japanese */pi pe/ intervocalically being [ɸi ɸe] > [wi we] > [ɥe ɥi] > [(j)e i]. I remember reading somewhere that the current te-forms are a mix of the forms that would've been expected in the historically prestigious Kyoto dialect and the Tokyo form that was taking its place in the 1800s, but I can't remember where right now.Travis B. wrote: ↑Tue Jun 04, 2024 4:41 pmMiddle Japanese /ɸ/ was rather unstable, and while it regularly became Modern Japanese [h] except before /i/ or /ɯ/, in many words it was simply lost. I have never seen it posited that Japanese ever had /ɸC/ clusters, though.Rounin Ryuuji wrote: ↑Tue Jun 04, 2024 4:01 pmI've never seen it given that way before, but it could have happened thus; I normally see an intermediary stage /otofuto/ (/pi/ > /fu/ is a common onbin in Kansai, I believe), but /otofto/ is certainly plausible.
- linguistcat
- Posts: 453
- Joined: Sun Jul 08, 2018 12:17 pm
- Location: Utah, USA
Re: Conlang Random Thread
I'd say the most plausible steps for pi > u in Japanese is the pi > ɸu > u route, since Japanese has basically always disallowed wu (and yi, but that's a different matter). And yeah as Ryuuji said, Eastern Old Japanese ( which Tokyo and similar dialects evolved from) and Western Japanese (Kyoto and the like) tended to do different things with certain consonants followed by high vowels.Rounin Ryuuji wrote: ↑Tue Jun 04, 2024 6:21 pmI cound see /ɸC/ > /CC/ happening very quickly, though, since something like this probably lead to the -う (-u), -つ (-tsu), -る (-ru) Godan verbs having their te-form in って (-tte). This seems to have mostly only occurred in longer words, and there isn't quite a consistent pattern, the expected outcome of Old Japanese */pi pe/ intervocalically being [ɸi ɸe] > [wi we] > [ɥe ɥi] > [(j)e i]. I remember reading somewhere that the current te-forms are a mix of the forms that would've been expected in the historically prestigious Kyoto dialect and the Tokyo form that was taking its place in the 1800s, but I can't remember where right now.Travis B. wrote: ↑Tue Jun 04, 2024 4:41 pmMiddle Japanese /ɸ/ was rather unstable, and while it regularly became Modern Japanese [h] except before /i/ or /ɯ/, in many words it was simply lost. I have never seen it posited that Japanese ever had /ɸC/ clusters, though.Rounin Ryuuji wrote: ↑Tue Jun 04, 2024 4:01 pm
I've never seen it given that way before, but it could have happened thus; I normally see an intermediary stage /otofuto/ (/pi/ > /fu/ is a common onbin in Kansai, I believe), but /otofto/ is certainly plausible.
A cat and a linguist.
Re: Conlang Random Thread
So I noticed that Hurrian has a fair number of clitics with conjunction-like behaviour, like =an "and", =mān "but" and =nīn "truly" among others.
I want Vrkhazhian to have these, so I did come up with a good many. Though I wonder how naturalistic it would be to have the majority of them be of the CVCV shape, like these:
I want Vrkhazhian to have these, so I did come up with a good many. Though I wonder how naturalistic it would be to have the majority of them be of the CVCV shape, like these:
More: show