Hacker News Reader

Comment by hackshack | Hacker News Reader

Between this, and icon-only toolbars and ribbons, I think we're reinventing Chinese, badly. Ideographic characters can often convey meaning succinctly.

My vote is to either go back to picture icons, or use Chinese characters with localized pronunciation, so 車 or 车 is car, and so on.

View on HN

Chinese is not in any way ideographic unless you are already partially literate.

They are of you go far enough back. This is what 车 looked like around 1000 BC: https://commons.wikimedia.org/wiki/File:%E8%BB%8A-oracle.svg...

It's still ideographic but not legible, right?

Just like most software icons are not legible without prior knowledge like arrow down mean to save, a circle with a line mean power on/off, etc. Both are ideographic, and I guess some software icons might be a bit more pictographic (like a cogwheel meaning settings because you are interacting with the machine).

Incidentally, the largest group of Chinese characters are phono-semantic e.g. encode both meaning and pronunciation. Over half of all Chinese characters are in that bucket. That actually allows speakers to have some ability to guess both pronunciation and meaning of characters they have never seen. There are rules to guide this.[0]

0. https://en.wikipedia.org/wiki/Y%C7%92u_bi%C4%81n_d%C3%BA_bi%...

...in Standard Chinese only, other varieties having very different pronunciations, right?

In Classical Chinese actually. Mandarin, which I assume you mean, is not the language these characters were designed for. But it is related enough that the phonetic hints often (but not always) help.

Classical Chinese had a much larger phonemic inventory than modern Mandarin, and notably no tones. Below are a collection of Classical Chinese reconstructions in IPA that are all pronounced yì in Mandarin today. (like "ee" for English speakers). The creation of tones and other sound changes were fairly predictable, so as you say, the hints often still help today.

- ŋjajs 議; 'discuss' - ŋjət 仡; 'powerful' - ʔjup 邑; 'city' - ʔjək 億; '100 million' - ʔjəks 意; 'thought' - ʔjek 益; 'increase' - ʔjik 抑; 'press down' - jak 弈; 'Go' - ljit 逸; 'flee' - ljək 翼; 'wing' - ljek 易; 'change' - ljeks 易; 'easy' - slek 蜴; 'lizard'

Nit point (I'm not sure it's relevant), but we don't know to what degree Old Chinese did or dit not have tones. The very first work to say anything at all about pronunciation is a Middle Chinese text from ~600AD, which already did have a system of 4 tones, albeit a different 4-tone system than Mandarin. Old Chinese pronunciation is a reconstruction from very limited data, not unlike proto-Indo-European, despite being considerably closer to the present.

I just looked it up and the phonetic markers are only like 20-30% reliable. I am shocked at this number as in my experience I would have thought it higher (I would have guessed 60-70%), but it is definitely hit-or-miss. I've never found tones to be predictable.

That might be a better word to use, maybe. But I'm not sure there was an adequate word for the point I was trying to make.

The linguistic definition of ideographic is that it is a language which uses symbols to represent concepts, rather than just literal pictures (pictographic) or sounds (alphabet or syllabrie).

Linguistics textbooks as far as I'm aware do not define symbol in this context, but generally a symbol seeks to represent the concept. Emoji are great symbols - you see an emoji and you largely understand its meaning, even if you have never seen it before.

The modern Chinese writing system is so abstracted that even an otherwise highly educated person that just lacks exposure to Chinese written script would have absolutely no idea what any of the characters mean. 一, 二, 三, sure. Beyond that, no fucking clue.

So yeah, they wouldn't be legible. Because as symbols, they objectively suck until you learn the basic components, structure, and patterns of organization of the characters.

So to the extent that an ideographic language conveys words as ideas through symbology, and to the uninitiated these symbols lack all meaning, it's not really ideographic is it?

But yeah, not legible might have gotten the point across better.

> So to the extent that an ideographic language conveys words as ideas through symbology, and to the uninitiated these symbols lack all meaning, it's not really ideographic is it?

If I write math equations in an unfamiliar and inscrutable notation does that somehow make them "not math"?

I don't think ideography is in the eye of the beholder but rather the creator. Using the uninitiated as your standard doesn't seem to work very well for most things beyond the absolute basics.

The key observation here with relevance to the original topic would probably be that icons that are legible to the uninitiated are likely to be of benefit. Even if you don't really care to accommodate them it's still going to help you to get your choices adopted.

Thus an amusing thought occurs to me. If we did want to switch to Chinese characters for icons it would probably make sense to do so gradually, one app every six months or so.

Many characters aren't ideographic at all. Nothing at all about the structure of 的 (genitive case marker), 是 (be), or 有 (have) hints towards their meaning. A number of others like 好 (good) are ideographic only through convoluted and unintuitive etymologies.

If you put the history of the orthography in reverse, it looks like someone getting good at drawing.

That's a bit harsh, I'd say no less than 80% of HN commenters are partially literate.

So... Like the current icon set? E.g. how is a picture of a gear connected to "Settings"?

Icon - Ideographic character is a really interesting connection I've never seen made before that seems to capture what is going on. Don't agree with your conclusion to "use chinese characters" though. I don't think it's easy to tell what they depict.

Reply on news.ycombinator.com