The Wubi Effect

August 18, 2020 10:01 AM

How do you fit 70,000 Chinese characters on a keyboard? Radiolab delves into the history of writing Chinese characters on computer, from Wang Yongmin’s pioneering Wubi method that breaks down Chinese characters into components mapping onto standard Latin keyboards, to the cloud-based predictive typing system used across China on PCs and smartphones today.
posted by adrianhon (15 comments total)

A Chinese friend in Beijing also loved this episode and told me that the top cloud-based AI predictive keyboard for iPhones in China, SoGou, is in the process of being bought by Tencent.
posted by adrianhon at 10:01 AM on August 18, 2020 [1 favorite]

There’s a good research project out there on the impact in literature of going from a visual to phonetic typing method. I’ve always loved the concept of visual puns as they exist in glyph based writing, I wonder if these are becoming less prevalent with pinyin’s dominance in schools?
posted by q*ben at 10:43 AM on August 18, 2020

Wow, that was really fascinating. Especially the part where the Chinese government was pushing Pinyin as a way to standardize pronunciation across China. I thought it was useful for learners to be able to write down words before being knowing the character for them, in the way that a Japanese learner could use romaji or hiragana/katakana, but didn't think of the political aspect of something like this.
posted by any portmanteau in a storm at 11:05 AM on August 18, 2020

I was already somewhat familiar with how text input worked for Japanese devices but this is all post-2000 when personal computers and relatively smart phones existed. Thinking about how it would be done within the constraints of 1980s-era computers with the significantly more characters used in Chinese writing highlights what an amazing achievement Wubi is, although to me it seems like it would require more mental overhead to use Wubi as opposed to a Pinyin-based predictive method.
posted by any portmanteau in a storm at 11:24 AM on August 18, 2020

I can't wait to get to this episode.

Two things come to mind, tangentially related to the topic:

1. The inventor of Pinyin had been at odds with the CCP for years and began speaking out against the government after he turned 100. Seems like a very interesting person.

2. The history of Uighur language is interesting, and it's gone through a few alphabets. When the People's Republic of China was starting, they were forced to use a Cyrillic script, and then changed a few more times as a method of controlling the population in Xinjiang. I remember speaking with one elderly man in Kashgar who became literate under previous alphabets and could read English and Russian with ease, but struggled with the current Arabic-type script. His opinion was that the Han Chinese realized the Uighurs, with Roman and Cyrillic alphabets, had a leg up on the Hans in a world increasingly dominated by Western Europe and so changed the alphabet to force all the Uighurs into illiteracy and make it harder for them to learn major languages of commerce. I remember seeing hints of the older alphabets around Kashgar a decade ago, but definitely not a lot of it around.
posted by msbrauer at 12:24 PM on August 18, 2020 [6 favorites]

Previously on MeFi: The Chinese Typewriter.

I've got the eponymous book. It's a bit of a slog. The mechanism underlying the typewriter is fascinating.
posted by adamrice at 12:29 PM on August 18, 2020 [1 favorite]

How interesting that the guy who invented Wubi complains about Chinese culture being diminished when Chinese characters are inputted via Pinyin (roman letters) rather than by stroke or radical or whatever. I mean, the fact that Chinese is written with ideograms that have strong pictorial and phonetic associations is just a weird side effect of its inefficiency and has nothing to do with written communication. It's just a mostly useless extra layer of information that kids have to spend years in school studying. Perhaps I'd feel different if Chinese were my native language.
posted by jabah at 1:55 PM on August 18, 2020

Being the sort idiosyncratic weirdo who builds his own keyboards, I’ve frequently wondered how that sort of behaviour manifests itself in ideographic cultures. What does a perfect keyboard look like to a scholar of Chinese linguistic history?
posted by mhoye at 2:43 PM on August 18, 2020

If anyone is interested in the material but prefers to read rather than listen, there's a pretty-decent transcript on the Radiolab site - just click the Transcript link under the picture. (A couple of minor imperfections in there like a "pinging" which clearly ought to be "Pinyin" and a "opinion editor" which might have been "Input Method Editor", but very serviceable.)

I found it to be a pretty interesting discussion - I was vaguely aware of the use of Pinyin romanization for entering Chinese characters but I had assumed there was basically just that one system. The back-story of Wubi and the tension between it and Pinyin was fascinating to learn about.
posted by BuxtonTheRed at 2:59 PM on August 18, 2020 [1 favorite]

It's also possible to input via Bopomofo, although in the PRC that may be unacceptable, due to its association with Taiwan (incidentally, Bopomofo looks like an alternate-reality version of Katakana, which is kind of what it is).
posted by adamrice at 3:19 PM on August 18, 2020 [1 favorite]

mhoye, I'd say the ideal input system for a person devoted to Chinese writing traditions is a good big graphic tablet.
posted by homerica at 6:28 PM on August 18, 2020

I 've wished many times that I'd learned Wubi, as there are lot of Chinese characters that I know by sight, but not by their pronunciations. Those are usually fancy words that only appear on pages.

jabah, there are people who are convinced that abandoning ideograms and replacing them with Pinyin completely would modernize the language and lower the barrier to literacy, but they are a very small minority. Two immediate reasons for this I can think of: the language has a vast amount of homonyms or near-homonyms, and there are so many local dialects in China.

A third reason is how much people are still vested in calligraphy. Because the Gaokao system values good handwriting and neatly written exam papers get precious bonus points, schools put a lot of emphasis on handwriting training. Ubiquity of digital technology hasn't changed that.
posted by of strange foe at 1:29 PM on August 19, 2020

Getting bonus points on the exams that determine the rest of your life seems like a reason to get rid of calligraphy more than anything else. Have everyone type their exams and let them be judged by what they write, not how cultured their hands are.
posted by any portmanteau in a storm at 2:48 PM on August 19, 2020

of strange foe: I was just speaking of pinyin as a method of inputting characters.

But re: homophones...I'd wager that by the time you factor in tones, compound words, determiners and fixed expressions, that there is a lot less homophony in Chinese than appears on the surface. What I was really getting at is: Chinese characters are so stylized that the cultural attachment to how they are formed is not really worth the disadvantages of having a "pictographic" text input system. That seemed to be the bone of contention of the inventor of Wubi—that people inputting via pinyin were somehow "missing out" culturally. I just think simpler is better.
posted by jabah at 5:29 PM on August 19, 2020

Jabah, sorry I misunderstood. Simplicity is definitely winning. Wubi was invented before computers got smart enough to do Chinese handwriting recognition. Now that the latter has gotten good enough to serve that part of the population who never learned to use Pinyin fluently (such as my mom, who's in her 70s), Wubi is mostly for professional typists who care about speed.
posted by of strange foe at 7:42 AM on August 20, 2020