If natlangs were conlangs

[akam chinjir]

Though Mandarin is unusually suited to that approach

Why do you think this is?

Just that I think it's fairly rare to have multiple tones, with their pitch contours mapping so easily onto distinct common diacritics.

Though actually, if you use -h to distinguish the low register (as in the Yale romanisation of Cantonese), and pinyin-style tone marking within registers, that might actually cover most cases (in languages without coda h, anyway).

[Salmoneus]

And the current Hmong and Zhuang orthographies, which as I mentioned in my previous post use letters which are otherwise unused instead of numbers, are worst — not only are they not mnemonic, they look exactly like regular letters.

Pah! These systems are cool, to be sure, but not as cool as mediaeval Arpitan. Mediaeval Arpitan used stress letters.

Your word has penultimate stress? Stick a -z on the end! Final stress? Add on an -x!

This is why the French Alps are still full of towns with names like "Verjux" and "Loisieux", "Unieux" and "Nandax", "Vulvoz" and "Mizérieux", "Excevenex" and "Saraz", "Saint-Paul-de-Varax", "Versonnex" and "Optevoz". They even liked the system so much they added the letters to some monosyllables, like "Oulx" and "Vaulx"...


I haven't yet used this system in a conlang, but one day...



[the problem with using 'mnemonic' diacritics for contour tones is that the contours rarely match the small range of available diacritics, which makes them misleading]

[Pabappa]

[the problem with using 'mnemonic' diacritics for contour tones is that the contours rarely match the small range of available diacritics, which makes them misleading]

Yeah, I've noticed. In pitch accent systems, sometimes the marked tones are all acoustically high, so the familiar grave/circ/acute setup gets repurposed. My conlangs are modeled on Baltic , so à=high short post-glottalized, á=high rising,ā=high falling,â=pharyngealized. Also ă for stressed mid tones. I find it easier to maintain the symbols as langs change, so these labels are not accurate for all my langs. E.g. ā sometimes evolves to a flat high level tone.

[bradrn]

And the current Hmong and Zhuang orthographies, which as I mentioned in my previous post use letters which are otherwise unused instead of numbers, are worst — not only are they not mnemonic, they look exactly like regular letters.

Pah! These systems are cool, to be sure, but not as cool as mediaeval Arpitan. Mediaeval Arpitan used stress letters.

Your word has penultimate stress? Stick a -z on the end! Final stress? Add on an -x!

This is why the French Alps are still full of towns with names like "Verjux" and "Loisieux", "Unieux" and "Nandax", "Vulvoz" and "Mizérieux", "Excevenex" and "Saraz", "Saint-Paul-de-Varax", "Versonnex" and "Optevoz". They even liked the system so much they added the letters to some monosyllables, like "Oulx" and "Vaulx"...


I haven't yet used this system in a conlang, but one day...

That’s… fairly amazing. Do you have a source for this? I couldn’t find anything about it online.

[the problem with using 'mnemonic' diacritics for contour tones is that the contours rarely match the small range of available diacritics, which makes them misleading]

This is correct. However, my view of this is that it’s still more mnemonic than any other system.

[Richard W]

At least it’s better than the current Zhuang orthography, though. When they got rid of the weird tone letters, they decided to copy Hmong and replace them with normal letters. So the current situation with both Hmong and Zhuang is that an arbitrary subset of letters represents tones, without any external indication of this except memorisation. So you get words like (in Zhuang) mwngz /mɯŋ˧˩/, hwnj /hɯn˥/, max /maː˦˨/ where the last letter represents a tone.

Those systems seem fairly obvious if you know the syllable structure. For Hmong, all syllables are open. For Zhuang, if you know 'ng' is a digraph, then the only memorisation for your examples is that 'x' is not a coda consonant. And if you're to read Zhuang out, you have to know which tones the tone letters represent, and ipso facto you will recognised tone letters even if they also serve as onset consonants.

I'm not sure that tone diacritics stick in the memory well. I find it difficult to remember which tone diacritic a word has, and historically they tend to get dropped very easily.

[Raholeun]

And the current Hmong and Zhuang orthographies, which as I mentioned in my previous post use letters which are otherwise unused instead of numbers, are worst — not only are they not mnemonic, they look exactly like regular letters.

Pah! These systems are cool, to be sure, but not as cool as mediaeval Arpitan. Mediaeval Arpitan used stress letters.

Your word has penultimate stress? Stick a -z on the end! Final stress? Add on an -x!

This is why the French Alps are still full of towns with names like "Verjux" and "Loisieux", "Unieux" and "Nandax", "Vulvoz" and "Mizérieux", "Excevenex" and "Saraz", "Saint-Paul-de-Varax", "Versonnex" and "Optevoz". They even liked the system so much they added the letters to some monosyllables, like "Oulx" and "Vaulx"...


I haven't yet used this system in a conlang, but one day...

That’s… fairly amazing. Do you have a source for this? I couldn’t find anything about it online.

Wikipedia mentions it, maybe there's more information in one of the linked sources.

[bradrn]

Wikipedia mentions it, maybe there's more information in one of the linked sources.

I didn’t see that — thanks! I actually did look on that page, but didn’t think to look at that particular section.

[Linguoboy]

I'm not sure that tone diacritics stick in the memory well. I find it difficult to remember which tone diacritic a word has, and historically they tend to get dropped very easily.

Same. I thought Gwoyeu Romatzyh was ridiculous as hell when I first encountred it, but now I think that if I'd taken the trouble to learn it, I'd have much less difficulty associating tones with morphemes. Maybe it's different if you grow up with an orthography with mandatory diacritics but I'm too used to them being optional. (This screws me up with German as well.)

[Pabappa]

For me I think the lack of diacritics in English is a help.... there's nothing I need to unlearn. I picked up the Lithuanian system easily enough, even with à being a high tone, but now that my conlangs are based on that I have a hard time thinking of graves as low tone. (Though I'm sure I'd do fine if I had to learn a new language written that way.)

[Linguoboy]

For me I think the lack of diacritics in English is a help.... there's nothing I need to unlearn.

You didn't need to unlearn /ʤ/ for <j> or /j/ for <y> or /eː/ for <a> when learning foreign languages?

[Vijay]

I've been exposed to orthographies with mandatory diacritics at least since I was four and started learning to count to ten in Mandarin when I was five (thanks to my dad's Chinese co-worker who took the trouble to write them out both in characters and in pinyin with tone diacritics). Maybe that's why getting diacritics to stick in my memory doesn't seem problematic for me as long as I understand the phonology and what the diacritics indicate.

[Pabappa]

For me I think the lack of diacritics in English is a help.... there's nothing I need to unlearn.

You didn't need to unlearn /ʤ/ for <j> or /j/ for <y> or /eː/ for <a> when learning foreign languages?

please don't do this, linguboy. You know I meant diacritucs because I was replying to you.

[Xwtek]

My first choice for tones is diacritics that match the tone contour, as in pinyin

Though Mandarin is unusually suited to that approach

Why do you think this is?

and really only if you think of its low tone in terms of the contour it gets when spoken in isolation.

This is true. But tone sandhi makes everything more difficult, and I think using the isolated form simplifies the romanization. But if you really want to consider it as a low tone, it would be easy to just use a different diacritic like a̱ instead of ǎ, and write words like Pu̱tōnghuà, Guóyu̱, fùyo̱u, li̱xìng instead of Pǔtōnghuà, Guóyǔ, fùyǒu, lǐxìng. (I actually quite like this version, now that I’ve tried it out!)

You already answered that. Also, ǎ is not really a low tone because it's pronounced as high rising next to another syllable with the same tone.

[Vijay]

It's pronounced as rising in that environment. I'm not sure I agree it's high rising (although there probably are sources that say it is).

[Xwtek]

It's pronounced as rising in that environment. I'm not sure I agree it's high rising (although there probably are sources that say it is).

Whoops sorry, my Mandarin teacher used to say it's prononounced as the second tone. It's actually another simplification, like b and p is pronounced as each other. (In Javanese p has no aspiration, but b does)

[M Mira]

It's pronounced as rising in that environment. I'm not sure I agree it's high rising (although there probably are sources that say it is).

Whoops sorry, my Mandarin teacher used to say it's prononounced as the second tone. It's actually another simplification, like b and p is pronounced as each other. (In Javanese p has no aspiration, but b does)

Actually, I believe that a substantial number of Mandarin speakers pronounce the third tone with sandhi and the second tone identically. Some 2nd tone characters that are often followed by a 3rd tone character, like 潛/潜 (qián), are frequently mis-interpreted as a 3rd tone character with sandhi because they are pronounced identically by some speakers. A quick Google of "qián third tone" in Chinese (潛 三聲) shows that the confusion isn't limited to any specific region.

[Zaarin]

(Of course, then there’s Squamish, with its use of 7 for glottal stop. Let’s not talk about that one please.)

This seems perfectly reasonable to me. Non-Unicode Roman transliterations of Semitic languages routinely use <3> for ayin and <7> for aleph.

[Vijay]

I've seen people who work on Chatino use 7 for glottal stop, too. I thought it was a clever alternative to actually digging up the glottal stop from an IPA keyboard or whatever, and it also happened to be very convenient for a research paper I had to do involving some text in Chatino.

[Kuchigakatai]

(Of course, then there’s Squamish, with its use of 7 for glottal stop. Let’s not talk about that one please.)

This seems perfectly reasonable to me. Non-Unicode Roman transliterations of Semitic languages routinely use <3> for ayin and <7> for aleph.

Is this use of <7> common in Hebrew or Akkadian? In Arabic it's <2> that is generally used for the glottal stop, <7> generally standing for /ħ/ instead.

[Zaarin]

(Of course, then there’s Squamish, with its use of 7 for glottal stop. Let’s not talk about that one please.)

This seems perfectly reasonable to me. Non-Unicode Roman transliterations of Semitic languages routinely use <3> for ayin and <7> for aleph.

Is this use of <7> common in Hebrew or Akkadian? In Arabic it's <2> that is generally used for the glottal stop, <7> generally standing for /ħ/ instead.

I don't think Akkadians text or chat much these days (and they also lacked a glottal stop, as far as we know). But I probably mistook the <7> used for heth based on my knowledge of Native Amerian languages (not just Squamish) that use <7> for a glottal stop. <2> for aleph certainly makes more sense.