Axunashin logographs

[tezcatlip0ca]

You're saying that there are 770 logographs and you didn't bother to put up a list or something?

[So Haleza Grise]

I know! Didn't Mark think of our god-given right to know what they are right now for free? How damned inconsiderate, frankly. I demand that all work on every other aspect of Almea or the Incatena cease this second.

[tezcatlip0ca]

Sorry, Haleza Grise, I didn't mean it that way. I meant that Axunašin is already finished, and the logographs aren't that many, so I thought Zomp had already created them or something. They're supposedly derived from the Wedēi logographs, which also aren't up yet so...

[Thomas Winwood]

770 logographs isn't that many?

[Mornche Geddick]

770 logographs isn't that many?

Not compared to hanzi (47035 glyphs), or kanji (50000), or Ancient Egyptian writing (5000+), or even Sumerian cuneiform (1000 at its peak). Modern day logographic writing systems seem to require you to learn 2000 - 3000 glyphs in order to be literate.

Admittedly, asking Zompist to match 3 millennia of hanzi development would seem to be a little unreasonable!

[Bob Johnson]

770 logographs isn't that many?

Not compared to hanzi (47035 glyphs), or kanji (50000), or Ancient Egyptian writing (5000+), or even Sumerian cuneiform (1000 at its peak). Modern day logographic writing systems seem to require you to learn 2000 - 3000 glyphs in order to be literate.

1. If you're going to enumerate hanzi and kanji separately, you should count instead traditional Han characters, jiantizi (Chinese simplified), and shinjitai (Japanese simplified), plus Japan-specific characters like 込 -- all this despite the vast majority of characters being the same in simplified Japanese and traditional Chinese usage, and even many simplified Chinese characters have only one component changed in a predictable way.

2. Your numbers for Han characters are inflated compared to the others because it treats e.g. 神 as one new glyph instead of a combination of ネ and 申 -- this is a linguistic viewpoint rather than a graphic one. By this methodology, Hangeul has 11582 glyphs rather than 24 combined into syllable blocks. The main difference here is that the Han characters have the determinative (radical) merged in, whereas other languages treat that as a separate glyph. (Cases such as 品 〜 口口口 are unclear.)

I'm not sure how Loprieno gets the 5000 figure for Egyptian, but he, likewise, shouldn't count the same glyph with and without logographic vertical stroke as two glyphs for the purpose of enumerating glyphs or for learning them; the same applies to the dual and plural markers, and the tripling of glyphs.

You are right about the 2000 for literacy; the Jōyō list is now 2136 (the new ones look like mostly name characters, though I see claw, tiger, and fist in there). However, this isn't 2000 completely different blobs or squiggles: most characters break down into an arrangement of familiar parts, radicals still give some idea of meaning, and the rest of the character usually helps. Compare this to sight-reading English and still getting the pronunciation right.

3. Nagging people is not necessarily a good way to get them to do things for you for free. Doing a good job of a ~30 character alphabet is hard enough, all 770 would be masochistic.

[zompist]

Though it might be fun sometime to really geek out and actually produce the full list, I'm not likely to do so soon... it doesn't seem like something people would actually read. (At least some people will read a grammar; few will read a dictionary.)

[tezcatlip0ca]

At least I asked about Axunašin and not about Uyseʔ. Axunashin is built up of straight lines and circles, which I guess make it easier. I also asked Zompist about Uyseʔ a few months ago.

[zompist]

Not really— each of the lines-and-circles graphs has to be based on an earlier Wede:i form.

[tezcatlip0ca]

Why do you still use colons for Wede:i? You replaced them in Cuêzi...

[Yiuel Raumbesrairc]

2. Your numbers for Han characters are inflated compared to the others because it treats e.g. 神 as one new glyph instead of a combination of ネ and 申 -- this is a linguistic viewpoint rather than a graphic one. By this methodology, Hangeul has 11582 glyphs rather than 24 combined into syllable blocks. The main difference here is that the Han characters have the determinative (radical) merged in, whereas other languages treat that as a separate glyph. (Cases such as 品 〜 口口口 are unclear.)

What.

神 is not simply a combination of ネ and 申, linguistically, unlike the Hangeul combinations to which you compare them. Sure, graphically, it is, but linguistically, the combination is not simply them, but it gives a whole product. Take your second case; three things put together. The meaning is not simply "three things together", its "product", a meaning which is produced from the combination of the three boxes. It's not mere concatenation. (Hangeul, on the other hand, is pure concatenation and combinations don't add any information.) All the reason to treat 神 as a distinct character, even though it is a combination graphically.

(Japanese takes the concept and runs away with it in ways unforgivable. 明日 /asu/ or /asita/ is the result (though this particular case is rather easy). The characters are treated as meaningful morphemes but are stripped away of any proprer pronounciation, the spoken language supplying another reading only used when both characters are together.)

[Yiuel Raumbesrairc]

Why do you still use colons for Wede:i? You replaced them in Cuêzi...

What about tradition?

* I learned esperanto on the internet, and I never bothered to install a software to write the accents. So I always typed "x" when I needed an accent. I got so used to it that, today, I find the x-style writing prettier than the accents or the use of "h", and I'd be hard pressed to write esperanto with the accents on the internet.

[Bristel]

(Japanese takes the concept and runs away with it in ways unforgivable. 明日 /asu/ or /asita/ is the result (though this particular case is rather easy). The characters are treated as meaningful morphemes but are stripped away of any proprer pronounciation, the spoken language supplying another reading only used when both characters are together.)

Isn't that more exactly: /asɯ/ and /aɕita/? (just asking, I'm not sure about the exact phonemes)

[Thomas Winwood]

明日 /asu/ or /asita/

Isn't that more exactly: /asɯ/ and /aɕita/?

He's using Nihon-shiki (as is standard in Japonic linguistics today) as a phonemic transcription.

Phonemically, Japanese's compressed unrounded high back vowel constitutes the /u/ in its five-vowel system /i e a o u/. (You don't go for phonetic exactness in a phonemic transcription, you use whatever is most convenient without being ambiguous.) [ɕ] and [s] are allophones of /s/ before /i/ (the yoon sounds sha, shu, sho are analysed as underlying /sja sju sjo/).

[Bristel]

明日 /asu/ or /asita/

Isn't that more exactly: /asɯ/ and /aɕita/?

He's using Nihon-shiki (as is standard in Japonic linguistics today) as a phonemic transcription.

Phonemically, Japanese's compressed unrounded high back vowel constitutes the /u/ in its five-vowel system /i e a o u/. (You don't go for phonetic exactness in a phonemic transcription, you use whatever is most convenient without being ambiguous.) [ɕ] and [s] are allophones of /s/ before /i/ (the yoon sounds sha, shu, sho are analysed as underlying /sja sju sjo/).

Ok, that makes sense. I didn't know that there was a separate transcription for Japonic languages.

[Mornche Geddick]

As I said, nobody could reasonably expect Zompist to do all 770 logographs. The syllabary is a fine piece of work on its own.

[Yiuel Raumbesrairc]

As I said, nobody could reasonably expect Zompist to do all 770 logographs. The syllabary is a fine piece of work on its own.

The problem, I think, wouldn't be about creating a logographic system. As zompist himself explains clearly in his demystification of Chinese writing, The whole system can be broken down to three hundred elements that combine pretty much the same way a language works, and three hundred elements, while more than the 150 characters of the syllabary, isn't that much impossible.

The bigger problems are :
- Combinatory : The logographic combinations, while usually sensible, are completely arbitrary. There is no reason Axunashin characters would actually map the combinatory found in Chinese, Egyptian or Sumerian. Therefore, zompist would have to create a whole new combinatory, which is a lot more difficult.
- Usage : Phonemical writing systems are easy to apply to a language. While you can add exceptions (to the point of ridiculousness like French), there's just that much detail you can put. Not so with logographic systems. The very power of logographic systems is how the characters can be completely seperated from their initial spoken language value. I quoted Japanese where characters are used purely for their lexical meaning, without any phonological pronounciation. This would happen for Axunashin (or I suppose it would, since it did happen in Japanese and Akkadian) and that creates a lot more work. Characters may have more than one pronounciation, characters may have contextual pronounciations, then you have all the possible exceptions and ashita-like constructions and who knows what else. (Interestingly, this process, with combinatory, is close to create a completely new language; indeed, that's what would have happened in Japanese if the process I alluded to could have been completed. So it does take a lot of work.)
- Aesthetics : Though a detail, it's all too easy to follow the aesthetics of the systems we know, but we shouldn't. Again more work.

I should probably write a whole article on "So you want to create the New Logographic System".

[Nortaneous]

I should probably write a whole article on "So you want to create the New Logographic System".

Definitely. I've started reading a bit about Tangut, and I think I'm going to give Serhes Kettw a partially logographic orthography. I'm not entirely sure how it'd work, since Serhes Kettw is pretty much agglutinative, but Japanese sort of got away with it, so I can too.

[Bob Johnson]

神 is not simply a combination of ネ and 申, linguistically, unlike the Hangeul combinations to which you compare them

Yes. Exactly. That's what I said. And as you said:

The whole system can be broken down to three hundred elements that combine pretty much the same way a language works, and three hundred elements, while more than the 150 characters of the syllabary, isn't that much impossible.

[zompist]

The whole system can be broken down to three hundred elements that combine pretty much the same way a language works, and three hundred elements, while more than the 150 characters of the syllabary, isn't that much impossible.

300 sounds way too low for hanzi. After all, that's not much more than the traditional number of radicals alone! DeFrancis talks about an underlying syllabary of about 900 glyphs.

[So Haleza Grise]

Sorry, Haleza Grise, I didn't mean it that way.

I'm sorry, too; I was a little snarky.

[Mornche Geddick]

I should probably write a whole article on "So you want to create the New Logographic System".

In the meantime, do you have a take on Geoffrey Sampson's theory, that English spelling has some of the features of a logographic system?

I'd also like to hear what Zompist has got to say about this, because he recommends his book.

[zompist]

It's true enough, so long as it's taken as Sampson intended it— basically, that our spelling system gives distinct shapes to homophones, avoiding ambiguity and facilitating holistic recognition. (E.g. we immediately detect the difference between to, two, too or night, knight. Experienced readers don't sound out a word letter by letter but grasp it as a whole, so differences in the shape of the word are useful.)

(It doesn't of course mean that English spelling resembles logographic systems in any other way— it's not syllable-by-syllable, for instance.)