tones for a protolang

[linguistcat]

I'm not great with tones so I've mostly ripped of Mandarin with a couple tweaks for this. I do have a couple specific questions though.

/a e o i u/ <a e o i u>
/a_M a:_M a_R a:_FR a_F/ <a ā á ǎ à>

Right now I have plain vowels and vowels written with macrons as having the same tone value (mid, no contour) and the difference being vowel length, but I'm considering changing one or both. I'm not thinking about earlier stages since this is already a protolang, and I'm not worried about exact tones. It's just important that we have a short tone that stays flat, one that goes up and one that goes down. Likewise, the long vowels either stay flat, or have a falling-rising contour. (If I have to this could be explained by rising or falling long vowels shortening before this stage of the language. Or flattening out.)

Q1: Would I need to explain away not having long vowels with a rising-falling contour or is this fine?
Q2: I feel like with my explanation of why there's no rising or falling long vowels, the two mid tones would also merge if they weren't already different tones. Should I change one or both just to explain why they haven't? Or would that not be strange in this circumstance?

[zompist]

Losing length for the contour tones seems a little weird, probably because in some languages, like Ancient Greek and Vedic Sanskrit, contour tones appeared (in certain cases) when vowels adjoined. This strikes me as a case where you're running into limitations of romanization; maybe consider indicating length another way.

It also strikes me that you've copied Mandarin maybe too faithfully... maybe leave out one tone. Mandarin's falling-rising tone is pretty rare among Chinese dialects. Or consider Cantonese, which has low-rising and high-rising tones (24, 35) as well as low-falling and high-falling (21, 53).

[Travis B.]

One thing I would do is simplify the tones themselves, but combine them with phonation, i.e. add glottalization or creaky voice to one register (and if you glottalize it, I would suggest making the vowel short) and perhaps add breathy voice to another. If you keep contour tones, like zompist I would suggest making the contour tones long rather than short.

[Travis B.]

I would also suggest doubling vowels to mark vowel length if you are having trouble using diacritics to mark both tone and length.

[Rounin Ryuuji]

Zompist's already touched on this a bit, but, while tinkering with a pitch accent myself, I ended up developing three types of long syllable — long rising and long falling (both marked with a macron, and occurring in mutually-exclusive environemtns), and long rising-falling (marked with a circumflex) — from the coalescence of two short vowels. In monosyllables, a low-high coalescence (as [s̪ɔ̀ᵝ.wɔ́ᵝ] → [s̪ɔ̀ɔ́ᵝ] → [s̪ɔ̀ᵝː] sō) results in a falling pitch, and a high-low to a rising-falling (as [s̪ɔ́ᵝ.wòᵝ] → [s̪ɔ́òᵝ] → [s̪ôᵝː]). You could easily do something similar to create the long vowels you want and have them have a more unique feel (especially if you have a complicating factor like loss of voicing distinctions in stops, and maybe fricatives, phonemicising falling tone).

[linguistcat]

Losing length for the contour tones seems a little weird, probably because in some languages, like Ancient Greek and Vedic Sanskrit, contour tones appeared (in certain cases) when vowels adjoined.

I don't know much about tones as I admitted, but I did see a video from Artifexian that said if a system has only non-contour tones and simple contours, the simple contours tend to be lengthened, BUT if there are non-contours, simple contours and complex contours, only the complex ones have to be long although that doesn't stop the simple contours from also being on long vowels (or non-contour tones). Just that only the most complex tone(s) has to be long. But maybe I misunderstood or Artifexian was incorrect about that.

This strikes me as a case where you're running into limitations of romanization; maybe consider indicating length another way.

Thanks Zompist, although it was more that I was explaining why I didn't have long rising and falling contours, not that I couldn't find ways to write them. Would this be fixed if length just wasn't phonemic for these tones, possibly being short in closed syllables but long in open ones? Would that cause me more problems? I was already considering collapsing the short and long basic tones, so if that would help I could just, do that.

It also strikes me that you've copied Mandarin maybe too faithfully... maybe leave out one tone. Mandarin's falling-rising tone is pretty rare among Chinese dialects. Or consider Cantonese, which has low-rising and high-rising tones (24, 35) as well as low-falling and high-falling (21, 53).

This is more than fair. I'm not particularly comfortable with having too many tones and just wanted enough that I could get some variation in the descendants (gaining some tones, rearranging them, or turning into a pitch accent instead).

One thing I would do is simplify the tones themselves, but combine them with phonation, i.e. add glottalization or creaky voice to one register (and if you glottalize it, I would suggest making the vowel short) and perhaps add breathy voice to another. If you keep contour tones, like zompist I would suggest making the contour tones long rather than short.

I really should try this but I'm going to have to look into how those might affect daughter langs later on.

I would also suggest doubling vowels to mark vowel length if you are having trouble using diacritics to mark both tone and length.

I just didn't want those to be long personally, but if it's that weird I can just make them long by default. It would also make most of my vowels long, which seems weirder, or force me to use neutral short vowels more than I'd like, which I don't like.

... You could easily do something similar to create the long vowels you want and have them have a more unique feel (especially if you have a complicating factor like loss of voicing distinctions in stops, and maybe fricatives, phonemicising falling tone).

Any of that would have happened long before what I care about, so this is a stage of a language just before its daughter languages started breaking off.

I'm sorry if this sounds dismissive but I just have to think a bit about all of this and decide what/how much to tweak before even getting into creating daughter langs from this.

[foxcatdog]

Of note is east asian style tone systems with complex contours on monosyllables with no length distinction are less common on languages with multisyllabic words. In cases where complex contours do occur they are generally on long vowels (ie much of Africa and North America). Maybe if you are unfamiliar with tone it is better to make a language like this.

[WeepingElf]

African and East Asian tone systems are about as different from each other as either are from European accent systems.

[Travis B.]

One thing I would also suggest is that if you were to make a glottalized register is to require coda voiceless plosives to be accompanied by it - and disallow coda voiced plosives. Then, in some of your daughter languages I would suggest losing some or all of these coda voiceless plosives but leaving the glottalized register behind.

[zompist]

This is more than fair. I'm not particularly comfortable with having too many tones and just wanted enough that I could get some variation in the descendants (gaining some tones, rearranging them, or turning into a pitch accent instead).

If you want to develop the tones in daughter languages, I'd suggest looking more into Chinese historical and dialect linguistics-- Jerry Norman's Chinese is a great place to start if you can find it. It gets into how tones appeared, how the four tones of Middle Chinese got doubled, variation between the major dialects, and even the surprising variation within Mandarin. E.g. if you expect that the rising and falling tones in Beijing are that way all over... no, they're reversed in Chengdu.

To put it another way, Mandarin is a good model if you just want a tonal language... it's not a great model if you want a bunch of tonal languages.

I know very little about Mesoamerican and West African tone languages, except that they'd be good places to look also!

[foxcatdog]

Oto Manguean languages seem to have complex tone with disyllables at least.