Grasping in the dark: Self-taught Chinese.

My recent interest in the Chinese language was sparked by the appearance last spring of an unusual volume of news reports out of China. At that time I realized to my vexation that I knew almost nothing about Chinese geography.

I’d at least heard of Wuhan. It gets a few paragraphs in Paul Theroux’s 1988 travel book Riding the Iron Rooster. He’d visited there years before and remembered it as “a nightmare city of muddy streets and black factories, pouring frothy poisons into the Yangtze”. By his second visit it had cleaned up and sprouted new towers, but its lurches into modernity, Theroux grumbled, “were not necessarily improvements”.

I learn from Wikipedia that Wuhan played a major role in 20th century Chinese history, as the seat of the 1911 rebellion that ended the Qing Dynasty and brought Sun Yat-sen’s Kuomintang to power. The septuagenarian Mao Zedong’s famous exhibition of his vitality – swimming 15 kilometres in the Yangtze River (with the assistance of the current) – took place at Wuhan in 1966.

Nowadays metropolitan Wuhan has almost 10 million people, about the same as Chicago. And up until last year, I couldn’t have told you even roughly where it was.

Embarrassed by my ignorance, I began looking at maps of China. As some of those maps included the Chinese names for cities and other geographical features, I began to notice the recurrence of certain characters: the hai in Shanghai and Hainan; the an in Anhui and Xi’an; the bei in Beijing, Hubei, and…Taipei?

On the other hand, the jiang in Xinjiang was not to be confused with the one in Jiangsu and Heilongjiang.

Intrigued, I began drawing a map, labelling it in English and Chinese, looking up the meaning of the characters as I went.

Most Chinese placenames, I learned, are surprisingly straightforward. The provinces of Shandong and Shanxi are, respectively, “mountain east” and “mountain west”. Shanghai is “upon the sea”. Beijing is “northern capital”. (The kanji characters for Tokyo, “eastern capital”, are in Mandarin pronounced Dongjing.)

It didn’t take me long to learn the dozen or so characters that reappear again and again in Chinese placenames – words like river, lake, sea, mountains, forest, and the four cardinal directions.

As a strategy for actually learning Chinese, my map studies weren’t that effective. Suppose you were a foreigner trying to learn English from the names of English towns. You’d learn a handful of useful geographical words like the ones above, a few of limited everyday utility like “ford” and “shire”, and a bunch more that are fossils of long-extinct words, like “wich” and “caster”.

Nevertheless, my efforts weren’t entirely wasted. Most Chinese characters are built out of smaller pieces, most of which started as simple pictograms – a hand, a tree, an elephant, and so on.

Once you start to recognize these smaller pieces and their variant forms, the characters become less baffling – you can discern order in what appears at first glance to be a pile of squiggles.

***

I try to be explicit, whenever anyone asks: I’m not “learning Chinese”. As I explained in a post early last year, I was daunted by the prospect of absorbing a new vocabulary and grammar, plus a new writing system, plus the “tones” without which (we are told) any attempt at enunciating their language will be met with puzzlement or laughter from Chinese-speakers.

I wondered instead whether I could break out the most immediately useful of the above challenges, learning the meaning of the characters alone, with the goal of being able to decipher the signs on local Chinese storefronts.

Like most westerners, I had a vague idea that Chinese writing is purely pictographic – that it conveys sense only, not sound. This isn’t really true. Many characters are pictograms – simplified pictures of things. Others are ideograms – graphical representations of concepts. Anthony Burgess in his 1992 book on language, A Mouthful of Air, described the character 不, or “not”, as

a sort of plant with a line above it. The plant is trying to grow, but the line is stopping it. This is a little poem of negativeness, a metaphor of notness.

Still other characters are ideogrammic compounds – two or more ideograms combined to illustrate a concept. Take 意, which means “idea”.

On top is the character meaning “sound”: it depicts some mysterious force emitting from an open mouth (with a horizontal line in the mouth, just in case you’re unclear on the source of the phenomenon). Beneath it is the character for “heart”. An idea, then, is a “heart-sound”.

You don’t need to speak Chinese to absorb the meaning of characters like “not” or “idea”. If the ancient scribes had stuck to pictograms, ideograms, and ideogrammic compounds, written Chinese might have evolved into a truly universal language – something like what the 20th-century inventor Charles Bliss had in mind with his quirky system of Blissymbolics. (Which was inspired by his own experiments in self-taught Chinese.)

But despite the ingenuity of the ancient scribes, it was impractical to compose a little visual poem for every word. To make headway, they devised another, more efficient method of invention – mashing together two existing characters, one to define the broad sense of the new character, and another to suggest how it should be pronounced. These so-called phono-semantic compounds now constitute the bulk of the Chinese character set.

Here’s one. On the left side is the “foot” character, indicating that the compound will have something to do with feet. (The semantic component is usually, but by no means always, on the left.) On the right side is the character bao, meaning “to roll up”. (Yes, it also refers to the steamed bun.)

The bao part is semantically irrelevant – 跑 in fact means “to run”. But the Chinese-speaker will guess, correctly, that it should be pronounced something like bao: in Mandarin, pao.

Because at this point I’m not trying to learn to speak Chinese, these phonetic components aren’t very useful to me. I usually don’t bother trying to remember how the characters are pronounced – aside from those whose pronunciations are mnemonically useful. For instance, 糖, which combines the semantic component “rice” with the phonetic component for the Tang Dynasty.

The syllable tang naturally makes me think of astronauts and orange-flavoured juice crystals, and from there it’s a short leap to the meaning of “sugar”.

Luckily, it’s rarely necessary to smuggle in English words to make sense of phono-semantic compounds. The ancient scribes weren’t completely indifferent to the poetic reading of their new character mashups. Take 晚, or “night”. The semantic component is “sun” and the phonetic component means “to remove”.

Even where the phonetic component was chosen solely for its sound, it’s usually possible to construct a little mnemonic story to help cement the compound in your mind. A well-known example is 找, “to search”, which joins the semantic “hand” with a phonetic component representing a ge, or dagger-axe: just picture a hand grasping in the dark for a weapon.

I was always being stumped by the rather impenetrable glyph 用, meaning “to use”…until I realized that the same shape, with a little handle on top, is the phonetic component of “bucket”: a “use”-ful item.

I should reiterate that the phonetic components are only hints to pronunciation. Since the characters are so old, and the spoken language has drifted so much over the millennia, many of these hints are completely misleading. Which brings me to…

An aside about Japanese.

Prior to writing this essay, I hadn’t given any thought to how the Japanese pronounce the kanji that make up around half of their writing system. (Kanji, by the way, just means “Han characters”. The Chinese call them hanzi.)

I suppose I assumed they were reading the characters for semantics only, just as I have been; that when they spoke the words aloud, they were using “native” Japanese words; and that the Chinese phonetic clues were therefore just as irrelevant to them as they are to me.

Leave it to the Japanese to do things in the most complicated possible way. It turns out that nearly every kanji has at least two readings – one for its original Chinese pronunciation and another for its “native” Japanese equivalent.

Suppose that every English word based on a Greek root was not only spelled using the Greek alphabet, but could be read either as Greek or as an equivalent term derived from Old English: “psychology”, say, would be spelled ψυχολογία and could be pronounced either as “psychologia” or as “soul-learning”. That’s sort of how the Japanese have elected to treat their Chinese loanwords.

For a foreigner trying to learn Japanese, these multiple readings must be awfully frustrating. But native Japanese learners don’t need to memorize the various readings from textbooks or flashcards – they pick them up effortlessly through speaking and listening. The kanji merely bundle together, under one symbol, various words they already know.

Perhaps the Japanese are onto something. Why not have a single symbol to represent two or more English words with the same meaning – “sight” and “vision”, for instance? Something like…oh, I dunno, 視, maybe. If nothing else, it would save space – and one of the main arguments for kanji seems to be its efficiency.

Still, it’s hard to blame the Japanese girl in this video who says of kanji, “I don’t want to learn them anymore. We can just use hiragana.”

The trouble with simplified Chinese.

As you may be aware, there are two main varieties of Chinese characters: the traditional ones, nowadays used mainly in Hong Kong and Taiwan, and the simplified ones used in mainland China. Up till now I’ve been showing the traditional characters in this essay.

The simplified characters were introduced in the 1950s, not long after the Communists took over. Around the same time, the Japanese introduced their own, less comprehensive kanji simplification scheme. Many characters simplified by the Communists kept their traditional forms in Japan. Others were simplified in a slightly different way.

Most of the characters are the same in all three writing systems. But a few, including many of the most common ones, have two, sometimes three different forms.

Here in Vancouver you mostly see the traditional characters. But the simplified versions have gained ground as immigration from mainland China has increased.

When I started dabbling in Chinese I figured I’d stick to the simplified characters because, well, they were simpler. Then as I started being able to recognize the characters on storefront signs I realized that if I ever wanted to be able to read anything around here, I should learn both varieties.

SkyTrain and Chinese language signs, Richmond, BC.
© Steve Bosch / Vancouver Sun.

Then I discovered something surprising. The traditional characters were easier to learn.

***

Let me talk about nostrils for a second. The word “nostril” comes from the Old English word “nosthyrl”. The “nos” part means “nose”. “Thyrl” is an extinct word meaning “hole”. (It’s related to the word “through”.)

So a nostril is just a nose-hole. Makes sense. But for English learners it must be pretty confusing. Why “nos” instead of “nose”? What the heck is a “tril”?

I imagine English learners must memorize “nostril” the way I’ve been memorizing Chinese characters – by way of mnemonics. Maybe something like this: a “trill” is a high-pitched sound, like you make when you whistle through your nose: a “nose-trill”.

Now, suppose that in the wake of a future Communist revolution, our new government were to decree, “Henceforth, in the name of simplicity, nostrils will be called nose-holes.” This would be sensible enough. Immigrants and schoolchildren would happily adopt the new word. In fifty years or so, after most of us old-timers had died off, “nostril” would be as extinct as “thyrl”. Good riddance!

But most linguistic reformers have less totalitarian aims. Knowing that it’s difficult to change the spoken language by decree, and that it tends to make old-timers grumpy, they concentrate on tinkering with the written language.

So the Peeple’s Kommittee Too Standerdize Inglish probably wouldn’t abolish “nostril”. They’d just change its spelling to “nawstrel”, to conform with “jaw” and “law” and “raw”.

“Nose”, obviously, would become “noze”.

And future English learners would lose a vital clue to the relationship between the two words.

***

Looking at many Chinese characters, it’s easy to see why you’d want to simplify them. Have a squint at the traditional character for “medicine” or “medical”: 醫. Unless your browser is set to an unusually large font size, it looks like a squashed bug. Here it is, magnified:

Now you can see that it includes the “arrow” character, which is easy to remember because it looks like a guy with an arrow through his head. The bottom rectangle comes from the “liquor” character, because how do you treat a guy with an arrow in his head? Obviously, with booze.

The shu at top right unnecessarily complicates my mnemonic story. Did my imaginary arrow victim also get stabbed with an antique spear?

Lucky for me, the bureaucrats who overhauled the writing system in the 1950s agreed that the shu didn’t fit: the simplified (also Japanese) character for “medicine” is just an arrow in a box. In this mnemonic story, the guy got shot with an arrow and now he’s lying on a bed in a modern, uncluttered medical facility. (Wishing he had some liquor, no doubt.)

Okay! Good job, simplifiers. But let’s look at another traditional character, 長, which means “long” – as in Chang Jiang, the “Long River”, known to us as the Yangtze.

The strokes on the bottom appear kind of random, but actually they’re a standard stroke-shape that shows up in other common characters, like the ones for “clothing” and “to eat”. The three horizontal strokes at the top are from a character meaning “hair” – because 長 originally referred to long hair.

Not too busy. Easy to remember. The Japanese simplifiers had enough sense to leave this character alone. But the mainland Chinese turned it into this:

Much quicker to draw, true: it requires half as many pen-strokes. But now the top part is missing the three horizontal lines that clue you in to its meaning, while the bottom part no longer matches the “clothing” and “to eat” characters – which weren’t simplified.

The new character is nothing but an abstract shape which must be drilled and memorized, without reference to other shapes already learned – just as westerners must learn the abstract shapes A through Z. But we have only 52 such shapes (counting lower- and upper-case) to worry about.

Did simplification make Chinese characters easier for schoolkids to learn? Having fewer strokes to memorize seems like it would be a plus, but I’d argue that the elimination of visual cues for mnemonic association cancels out the advantage. That’s been my own experience, anyway, and it’s why I’ve switched to concentrating on the traditional characters in my own haphazard studies.

***

Have I gained anything by my self-directed Chinese studies? Not much. After more than six months of, admittedly, not very conscientious efforts, I still can’t read a primary-level Chinese text. I recognize many, sometimes most of the characters, but they remain stubbornly isolated from one another. I translate each one to its likeliest English equivalent and glare suspiciously at the resulting string of nonsense.

Imagine a foreigner tackling a simple English sentence: “I am going to the state fair.” The sentence begins straightforwardly enough. But where is the speaker going? What’s this “state fair”? A condition of beauty? An administrative district of moderation? A spoken-word exhibition?

Knowing the meaning of all the words isn’t enough. Knowing all the meanings of all the words may actually add to your confusion. You need to know the most plausible ways for the words to pair up.

(And obviously, when it comes to Chinese, I’m nowhere near knowing all the meanings of all the words.)

If I’d followed a curriculum I’d likely be a little further along. On the other hand, going my own way has allowed me to skip some of the off-putting stuff – like the tones – that might have caused me to dump the project entirely.

Perhaps once I’ve achieved fluency in reading Chinese – which at my current rate of progress could easily take a decade or more – I’ll try learning to speak it as well. But probably not. I’m not that enthusiastic about conversing in Chinese. I can barely be bothered to converse in English.

M.

—

My last attempt to explicate the Chinese writing system roped together an essay by Hillaire Belloc, a novel by Robert Graves, and the elephant of Han Fei. Nine years ago I took issue with Paul Theroux’s interpretation of the American bombing of Hanoi. It seems I’ve never mentioned Anthony Burgess before – except in a footnote to last month’s essay on Robert Heinlein’s The Puppet Masters.

The rectification of names.

The Chinese government seems to have been successful in its campaign to guilt us into replacing the logical, easy-to-remember “Wuhan virus” with the turgid, clinical “Covid-19”.

Apart from everything else, it strikes me as a blown marketing opportunity for the city of Wuhan. When international travel picks up again, western tourists who would otherwise hop straight from the Great Wall to the giant panda sanctuary at Chengdu might be convinced to add a stop at Wuhan Virusland. The mascot could be a pangolin wearing a surgical mask. Ozzy Osbourne could star in a promotional video where he dips into a bowl of delicious bat soup.

But if Beijing has its way, in a year or two Wuhan – that insignificant provincial town, home to a mere nine million souls – will recede into the obscurity it enjoyed before the virus made it briefly famous.

We in the west are pretty clueless about Chinese geography. It’s partly because China was closed to the outside world for 30 years, partly because their language looks so forbiddingly strange, and partly because, in a test-run of the Wuhan/Covid guilt trip, we went meekly along with their decree that we should junk our old, familiar names for their towns and provinces and replace them with hard-to-pronounce Chinese versions – so Tsingtao became Qingdao, Canton became Guangzhou, Amoy became Xiamen, and so on.

(In his 1988 travel book Riding the Iron Rooster, Paul Theroux is corrected by a government flunky when he refers to Peking and Canton. “I’m giving you their English names, Mr. Zhong,” he replies. “We don’t say Hellas for Greece, or Roma for Rome, or Paree, if we’re speaking English. So I don’t see the point–” But the flunky smoothly changes the subject.) [1]

Speaking of under-publicized tourist destinations, Zhengzhou is another huge city – almost six million people – that I couldn’t have placed on a map before the other day. That’s probably why I was unaware of this monument to the ancient semi-mythical emperor-heroes Huang and Yan carved into a mountain outside of town. Their faces are three times as big as the ones at Mount Rushmore.

Image by V.T. Polywoda.

Meanwhile in Changsha (population five million) there’s an oddly sexy 100-foot-tall bust of Mao Zedong. Or if you like your colossi a little shaggier, the 1200 year old giant Buddha statue near Leshan (a quaint village of 1.2 million) gives a preview of how Mao will look in a millennium or so, when the elements have done their work.

I was watching The Neverending Story with a friend a while back and when I saw the Ivory Tower – the fortress sprouting like a pistil from the shell of a hollowed-out mountain – I said, “How come our multibillionaires all live in boring suburban mega-mansions when they could be using their fortunes to erect cool fantasy architecture like that?”

But even if Jeff Bezos yearned to live in a hollowed-out mountain, he would never get away with it. For that matter, Mount Rushmore wouldn’t get the go-ahead nowadays. The local Native Americans would raise a fuss, protesters would converge, lawsuits would be launched, and after a few years the whole thing would be quietly dropped, as happened to that “grandiose” (actually, by Chinese standards, rather understated) statue of “Mother Canada” the Tories were talking about building in Cape Breton.

The Chinese, poor rubes, lack the sophistication to realize that enormous monuments to their heroes and heritage are gaudy and wasteful, and that developed countries have more important things to spend their money on, such as…wait a second, what are we spending our money on? Our infrastructure is rickety and inadequate. Our streets are full of homeless drug addicts. Our homes are full of cheap made-in-China crapola. Is it possible that all our extra wealth is going into inflated university degrees and pipeline litigation?

***

Ever since I moved to Vancouver from the Canadian prairies, I’ve had the vague intention of learning a little Chinese. Not enough to actually talk to people – I figure that’s unrealistically ambitious – but maybe enough to make out the gist of signs outside the many local Chinese businesses.

As I understand it – and I’m aware this is a gross oversimplification – Chinese characters, or hanzi, are built from ideograms representing ideas rather than sounds. Two quick strokes make a person; a few extra strokes denote a woman; two women side-by-side, hilariously, represent a quarrel. The concept of “big” is communicated by a little man, arms thrown wide, going “it’s this big!”

Thus speakers of mutually unintelligible Chinese languages – Mandarin, Cantonese, Fujianese, etc. – can still communicate by scrawling characters on a piece of paper. Chinese travellers in Japan and Korea can also get along, to some degree, without knowing the local languages because hanzi (or kanji, or hanja) form part of the Japanese and Korean writing systems.

I’ve heard mixed reports as to whether Chinese languages are especially difficult for westerners to learn. I assume they are: on top of the usual challenges of learning a foreign A) vocabulary and B) grammar, you’ve also got C) a completely alien tone system and D) at a bare minimum, a few hundred non-phonetic characters to memorize.

Maybe if your goal is to become a fluent Chinese speaker you need to learn A, B, C, and D together. But I’ve often wondered if it wouldn’t be useful to disaggregate the various off-putting features of learning Chinese. Maybe you could, for instance, acquire a basic vocabulary without worrying about tones.

Consider English: every word has a stress that falls on one syllable or other, sometimes according to a predictable rule but often not. We say “AUTomobile”, “autoMOtive”, and “auTOMoton”, which is just something foreigners have to learn – but we can still understand those words if all the syllables are stressed equally, even if the result sounds funny and robotic to us.

The go-to example for the Chinese tone system is the sound “ma”, which in Mandarin can mean “mother”, “horse”, “hemp”, and “scold”, depending which tone is used. But those are pretty distinct concepts – couldn’t the listener figure out by context which is intended, the same way we do with “be” and “bee”, or “high” and “hi”?

This Mandarin language teacher pretty much concedes my point:

[B]elieve it or not, people can mostly understand when foreigners speak without tones. Why? Because of context.

But before you become tempted to take this “shortcut” yourself…don’t! It’s a big mistake! You see, even though people might still be able to understand you if you don’t use tones, it’s not accurate Chinese. And the other person may have to try much harder to catch what you’re trying to say.

You’re basically limiting yourself to “complete beginner”.

But if “complete beginner” is all you’re aiming for – why not? There are a lot of people who, like me, might be interested in acquiring just a smattering of Chinese, who would be happy to take this shortcut if they knew it existed.

Likewise, maybe it would be useful to learn Chinese characters without learning a word of Chinese. Maybe we could absorb a limited set of hanzi into our language, which we could use to communicate across language barriers not only with Chinese, Japanese, and Koreans, but with Germans, Russians, Indonesians, and so forth.

You might say, why import a bunch of antique, abstract, overly complicated ideograms from a foreign country? Why not devise a brand new set of simple, rational ideograms of our own?

Well, consider the fate of Blissymbolics, the hanzi-inspired, hyperrational universal language invented by a disillusioned Eastern European Jew during World War II. (It was introduced in a book called Semantography: A Logical Writing for an Illogical World.) Blissymbolics caught on in a limited way as a method of teaching writing to handicapped kids in Canada, and nowhere else.

From The Book to the Film “Mr. Symbol Man”, by Charles K. Bliss. Image source. You can watch Mr. Symbol Man on YouTube.

That’s how it goes with a constructed language: absent a pre-existing population of speakers and a pre-existing body of texts, there’s little reason, apart from ideological enthusiasm, to learn it. With no-one to talk to and nothing worth reading, students grow bored and chuck it over. Whereas with Chinese you can just take the bus down Kingsway and every third or fourth storefront will present a new opportunity to test your vocabulary.

If our descendants ever do wind up adopting hanzi into the English language, it won’t be through the efforts of armchair theorizers like me. Attempts to benevolently direct linguistic evolution tend to backfire. For instance, the Chinese government “simplified” their writing system in the 1950s, reducing the number of pen strokes needed to draw many common hanzi. But in Hong Kong and Taiwan they ignored these directives, so that now many readers of “simplified” Chinese have trouble reading the “traditional” forms, and vice versa. Meanwhile the Japanese adopted some, but not all, of the simplified forms. (See also.)

This reminds me of the various ineffective attempts to preserve Canada’s endangered aboriginal languages. I can appreciate that aboriginal people would like to hang onto those languages. I think it’s a laudable goal. But to take a local example, there are 14 different Coast Salish dialects on or near the southern BC / Washington coast, distributed over an area smaller than Ireland. (The modern convention is to call them “languages”, but it seems that adjacent tribes could understand one another, though more distant ones couldn’t.)

The “7” stands for the number of people who can actually read this. Image source.

Left unmolested by Europeans, a single dominant dialect would eventually have emerged – or maybe the Coast Salish would have been conquered by some other, more unified tribe and had an alien language imposed on them, as happened to the Irish.

My point being, in my imaginary Coast Salish Republic, there’d still be at least 13 dialects regrettably falling into disuse, with old-timers in the sticks grousing that their grandkids didn’t know the words to the old folk songs anymore. But Coast Salish as a whole would stand a chance of survival. It would have enough speakers to sustain newspapers, a publishing industry, radio, TV, and so on.

My further point being, if there’s any chance of preserving Coast Salish now that its surviving dialects are mumbled by a handful of codgers each – it will be by picking one. But then, how do you get the 14 or more Coast Salish-speaking communities to agree to a strategy that involves 13 of them euthanizing an essential part of their culture for the good of the rest?

M.

1. Re Peking/Beijing, Kingsley Amis grumbled in The King’s English: A Guide to Modern Usage, under the heading “Didacticism”:

[T]hat right of the English language, as of any other, to devise its own forms for foreign names is under constant erosion. Peking was an English word for centuries before it was suddenly replaced by Beijing, however you pronounce it; Ceylon has notoriously been replaced by Sri Lanka; Lyons has reverted to Lyon (Lee-on(g)) and Marseilles (pronounced Marsails) to Marseille (MarSAY, often with an attempt at the French uvular trill in the middle); Seville and Genoa have come a step nearer being pronounced in the native fashion. What about Brussels and Brussels? Ah, that I predict will go on as before. The British/English form conveniently steers between Bruxelles and Brüssel, the Walloon and Flemish versions of the name of the Belgian capital.

Mark Steyn once referred to this trend as “the reflexive multicultural cringe that automatically assumes any new, less familiar (and thus less ‘western’) name must be more ‘authentic'”.