Conlang Random Thread

[akam chinjir]

Surely in the IPA ʝ˕ just is j.

Is the idea that you've got i~y and j~ɥ, and are wondering if you could write them both as ɩ? I don't see why not.

[bradrn]

Surely in the IPA ʝ˕ just is j.

I don’t believe so. [j] is specifically unrounded; [ʝ˕] is unspecified for roundness. From Wikipedia:

Some languages […] have a palatal approximant that is unspecified for rounding, and therefore cannot be considered the semivocalic equivalent of either [i] or its rounded counterpart [y] (which would normally correspond to [ɥ]). An example of such language is Spanish, which distinguishes two palatal approximants: an approximant semivowel [j], which is always unrounded, and an unspecified for rounding approximant consonant [ʝ̞]. […] "[T]he IPA shows a lack of precision in the treatment it gives to approximants, if we take into account our understanding of the phonetics of Spanish. [ʝ̞] and [j] are two different segments, but they have to be labelled as voiced palatal approximant consonants. I think that the former is a real consonant, whereas the latter is a semi-consonant, as it has traditionally been called in Spanish, or a semi-vowel, if preferred. The IPA, though, classifies it as a consonant."

[akam chinjir]

Interesting. The fact that they're in complementary distribution, at least in Spanish, maybe makes it a bit complicated, but the argument looks pretty good to me.

[TurkeySloth]

Surely in the IPA ʝ˕ just is j.

Is the idea that you've got i~y and j~ɥ, and are wondering if you could write them both as ɩ? I don't see why not.

At one time, the language language family had /i, j/, written <i>, and /y/, written <y>. I split /i/ and /j/ into <i> and <j> orthographically because I didn't want semivowels in the language family at all, meaning there was a short time the family had /j, i, y/ as <j, i, y>. Then, I shifted [j → ʝ˕], giving the family /ʝ˕, i, y/ as <j, i, y>. Lastly—for now, I merged {i, y → ɩ}, potentially creating a semivowel because both sounds have unspecified roundness. Currently, the family has /ʝ˕, ɩ/ as <i, y> Thus, my question about using the same orthography for the /ʝ˕, ɩ/ phonemic transcription doublet. Though, please be careful when you separate phonemes or phones with a tilde because it, actually, symbolizes free variation, which was never true of the family's /i, y/ phonemic pair.

[akam chinjir]

Though, please be careful when you separate phonemes or phones with a tilde because it, actually, symbolizes free variation, which was never true of the family's /i, y/ phonemic pair.

The tilde is also regularly used in cases of conditioned alternation.

[TurkeySloth]

Though, please be careful when you separate phonemes or phones with a tilde because it, actually, symbolizes free variation, which was never true of the family's /i, y/ phonemic pair.

The tilde is also regularly used in cases of conditioned alternation.

Okay. But, the language's /i, y/ never varied with each other. On top of that, the family never had /ɥ/ or [ɥ].

[Pabappa]

Ive noticed that diachronics often "get away from me". e.g. a language like Oyster, which I originally had intended to be extremely .... almost impossibly ... conservative, is now just about average and maybe even has more changes than its sister languages. Likewise Icecap Moonshine once had over 200 individual sound changes in its list and is now just down to 52 because I pruned out all the changes that were made irrelevant by other changes. The grand champion right now is Tarise, because while a list of 41 sound changes might not look like a lot, it's only a list of the changes of consonant clusters in the onset ... that is, i havent even gotten to single-consonant onsets, coda consonants, conditioning environments, etc .... and I dont have a single sound change involving vowels up there yet, even though the language is supposed to go from six vowels to three.

[masako]

Got a little bored, so I adapted Hangul for Amal. This was only partly inspired by the recent celebration of Hangul Day.



Based on this chart, you can see that the CVC roots fit very nicely into the Hangul system. The glottal stop being represented by the “ieung” when intervocalic. The “ssangsiot” is used for /ʃ/. The distinction between /r/ and /l/ must be realized contextually, as well as /i/ vs /ji/. Other than these notes, it is a fairly straightforward system.





This will not be used except in notes and practice...just for fun.

[KathTheDragon]

Why do you divide "amanti" as "am-an-ti", "buqura" as "bu-qur-a", and "shabanla me ede" as "shab-an-la ma ed-e", with inconsistent assignment of onsets and codas?

[linguistcat]

This question might not be applicable to so many conlangers, but I'd like to see any answers I might get. When deriving a conlang from a proto-language reconstruction (like for a bogolang or similar), do you tend to choose the reconstruction that is easiest to use/work with, or the one that is considered likely the most accurate? Assuming the two are not the same.

I'm needing to derive some things from both Early Middle Chinese and Old Japanese, and while I'm likely to go the easier route for both, I'd like to see how other people approach projects like this.

[Pabappa]

The former, but the only time I tried doing this I gave up very quickly because it was PIE and even the easiest reconstruction of PIE has problems all over the place. I wish you luck no matter which path you choose.

[Moose-tache]

This question might not be applicable to so many conlangers, but I'd like to see any answers I might get. When deriving a conlang from a proto-language reconstruction (like for a bogolang or similar), do you tend to choose the reconstruction that is easiest to use/work with, or the one that is considered likely the most accurate? Assuming the two are not the same.

I'm needing to derive some things from both Early Middle Chinese and Old Japanese, and while I'm likely to go the easier route for both, I'd like to see how other people approach projects like this.

A great question! In most cases I think there won't be much difference because both versions are based on reflexes, which is what you're trying to make yourself. Glottalic vs. non-glottalic PIE, for example, makes no difference since you're not going to have glottals in the finished product unless you're a crazy person.

As far as EMC and OJ are concerned, I assume you're talking about the exact realizations of the lost vowels of Japanese and the controversial rimes of Chinese? In many cases I think it won't matter. For example, Korean borrowed some i as a. That tells us nothing about Chinese, but a lot about Korean, because they thought the preceding velarization was more important than vowel height. Similarly Korean words borrowed from OJ don't fret about whether schwa was actually a diphthong or not. The phonology of your target language is going to determine how words are borrowed, and will help to turn some of those phonetic controversies into simple solutions.

[linguistcat]

... As far as EMC and OJ are concerned, I assume you're talking about the exact realizations of the lost vowels of Japanese and the controversial rimes of Chinese? In many cases I think it won't matter. For example, Korean borrowed some i as a. That tells us nothing about Chinese, but a lot about Korean, because they thought the preceding velarization was more important than vowel height. Similarly Korean words borrowed from OJ don't fret about whether schwa was actually a diphthong or not. The phonology of your target language is going to determine how words are borrowed, and will help to turn some of those phonetic controversies into simple solutions.

Good point. I think there could be some interesting details that could come out of a certain OJ vowel reconstruction, but I also like that reconstruction personally so I'm already biased. For EMC, I'll just go with the easiest one since I'm a lot less familiar.

[Xwtek]

Is this syntax realistic:

1. If the subject and the object is both core pronoun, the compound pronoun is used.
2. If only one of them is a core pronoun, then only core pronoun is required. The rest of the subject/object is understood for third person. (Alternatively, the full version of 3rd person pronoun is required)
3. If both of the subject and the object is third person, the 3rd person pronoun is required if the subject is omitted or right clefted.

So that third-person object never appears. My target is to make these pronoun an affix of verb that simultaneously codes voice/inverse (My language is neither a pure direct/inverse language or a pure symmetric voice.)

[Raholeun]

This question might not be applicable to so many conlangers, but I'd like to see any answers I might get. When deriving a conlang from a proto-language reconstruction (like for a bogolang or similar), do you tend to choose the reconstruction that is easiest to use/work with, or the one that is considered likely the most accurate? Assuming the two are not the same.

What worked best for me, was to combine several RL linguistic reconstructions at the same time and then deviating from that when it really suited my language. Explaining those deviations and giving them some background story in a faux academic style is a big part of the fun for me.

[Kuchigakatai]

Is this syntax realistic:

1. If the subject and the object is both core pronoun, the compound pronoun is used.
2. If only one of them is a core pronoun, then only core pronoun is required. The rest of the subject/object is understood for third person. (Alternatively, the full version of 3rd person pronoun is required)
3. If both of the subject and the object is third person, the 3rd person pronoun is required if the subject is omitted or right clefted.

So that third-person object never appears. My target is to make these pronoun an affix of verb that simultaneously codes voice/inverse (My language is neither a pure direct/inverse language or a pure symmetric voice.)

I have no objections to your syntax. In fact it reminds me of Mandarin, where there's a tendency to omit 3rd person inanimate objects. There is a pronoun for such objects, 它 tā, plural 它們 tāmen, but it is omitted a lot of the time. Much to the shock of us Westerners, it's perfectly fine to say things like:

我不要。
wǒ bú yào
1SG not want
'I don't want it.'

很喜歡嗎？
hěn xǐhuan ma
very like Q
'Do you like it a lot?'

不要銷毀。
bú yào xiāohuǐ
not want destroy
'Don't destroy it.' ("I don't want you to destroy it", said in the context of talking about some criminal evidence.)

You can add 它 tā, but it's a perfectly idiomatic option to omit it.

[TurkeySloth]

My main language's [l] vocalized to [u] in certain situations, resulting in the following tentative vowels: [ɤ, ʌ̞, ɯ, uː, ɑu̯, ou̯] <el, æl, il, ul, al, ol>. The long vowel and diphthongs are solid. Are [ʌ̞, ɤ, ɯ] plausible outcomes for [æl, el, il]? If not, what are plausible outcomes?

[Travis B.]

我不要。
wǒ bú yào
1SG not want
'I don't want it.'

I somehow am reminded of the "DO NOT WANT" meme from back when.

[masako]

Why do you divide "amanti" as "am-an-ti", "buqura" as "bu-qur-a", and "shabanla me ede" as "shab-an-la ma ed-e", with inconsistent assignment of onsets and codas?

It's not inconsistent, it's based on the root.

qam /ʔam/ - love; adoration; romance >> amek - to love (tv)
qan /ʔan/ - 1st person singular >> ana / -an - I; me (pro)
tiq /tiʔ/ - 2s >> ti / -ti/-e - you (pro)

buq /buʔ/ - this [proximal] >> bu- / be- (based on root vowel) - this (by me) (det)

shab /ʃab/ - knowledge; knowing; understanding >> shabra - to know (v)

...etc

[KathTheDragon]

Fair enough