I have just read this book straight through three times in a row in an effort to make sure that I was not missing or misunderstanding anything. Each time I came away with a greater horror at the inaccurate picture which it creates. Please note that the statements in Mr. DeFrancis book, which is a sustained argument of close to three hundred pages, cannot be thoroughly refuted in a review of the length permitted by this website; what follows is a mere summary of my principal objections to this book.
1. The arguments are circular. We get a pretty strange example of this almost at the outset. Linguists, we are told, prefer to use the word 'language' exclusively to mean the spoken, not the written, word. This is to avoid confusion. Mr. DeFrancis, who has a habit of falling back on every-day common sense only after a couple of pages of mind-numbing obfuscation, ultimately declares that he himself will not follow the advice of these anonymous 'linguists' in his own book: indeed his very title indicates a refusal to follow that advice. Instead, he promises to keep the distinction between writing and speech clear in his own writing, and to use (in what he makes sound like a brilliant solution to a vexing problem, whereas in fact the problem would not have vexed anyone had he himself not brought it up) to use the expressions "spoken language" and "written language". Why then bring up the problem at all? He and his fellow linguists are, it seems, engaged in a "persistent but largely unsuccessful battle against the confusion resulting from the popular use of the term" to mean both writing and speech. He gives two examples of this 'confusion'. First, the statement of a textbook writer that "two thousand [chinese characters] are sufficient for the speech of a well-educated man." DeFrancis comments: "This comment evokes a picture of our 'well-educated man' parading about like a comic strip figure with cahracter-filled balloons coming out of his mouth." Now, is it unclear to anyone that the speech of the well-educated man, according to the textbook writer, would, if written down, require about two thousand characters? Where is the confusion? Every day we see some writing which we cannot decipher and ask: "What does this say?" I suppose DeFrancis would chortle at that and then point out sententiously that it doesn't "say" anything as it is writing, and writing doesn't speak. He goes on to say:
"More typically misleading is the frequent dinner-table situation in which Chinese guests, when asked about their language, blandly assume that it is an inquiry about Chinese writing (which may indeed have been the case) or simply do not recognise the distinction and thus regale their listeners by dragging out the shopworn example of how the character for 'woman' and the character for 'child' are charmingly combined to form the character for 'good'. Incongruities and muddleheadedness of the kind just noted irritate scholars...."
Let us look closely at this passage; it exhibits three of the most common weapons in the DeFrancis arsenal. First, where is the muddleheadedness? We were told that using the word language to mean only speech, excluding writing, was necessary to avoid confusion, and that this anecdote would be an example of the confusion that might result if we did not make that distinction. On the contrary, however, it is only when we have such a distinction in mind--when we already expect "language" to mean only "speech"--that there is anything wrong with the response of the Chinese dinner guest above. Furthermore, the anecdote is highly suspicious. The imaginary Chinese dinner guest appears to have been asked something quite vague: "Tell us about your language." The scholar who asks so general a question as that has no right to be irritated when the answer does not happen to correspond to the aspect of the Chinese language he had hoped to hear about. Why, then, does this imaginary guest get painted as a muddle-head?
2. DeFrancis makes a highly reductionist argument. In what starts out as a highly accurate, scientific account of the nature of the Chinese writing system, he eventually ends up creating the false impression that Chinese characters are almost all phonetic. This may in itself be a useful counter-weight to the tendency, which indeed existed and exists, to treat Chinese as if it were exclusively ideographic, and which DeFrancis is right to oppose, but he practices an astonishing sleight-of-hand when he leaps from the fact that Chinese characters are largely phonetic to the suggestion that therefore only the phonetic principle deserves attention. He belittles the fact that even the most phonetic characters have a semantic component as well; he utterly neglects the fact that many phonetic elements are not exclusively phonetic at all but have a semantic aspect as well; and, worst of all, he dramatically underestimates the importance of the characters which are not phonetic at all but either pictographic or ideographic. (the sort of "woman+child=good" compounds whose introduction into dinner-table conversation he so deplores.) The way he does this is one of breath-taking cunning. Elsewhere, he rightly condemns the practice of taking the whole contents of large dictionaries as representative of the language, and highlights the importance of looking at characters or words which actually come up in typical prose instead. But here, he simply points out that in the enormous KangXi dictionary, 93% of the characters are of the phonetic-semantic compound variety. This, again, besides ignoring the fact that all of those characters have a semantic component, and that in many of them the phonetic component also has semantic value, also overlooks the fact that in normal Chinese prose characters of the purely ideographic sort will come up far more than 7% of the time.
3. There are also some strange bits of dishonesty here. He makes much of the rebus principle: the character "lai" for "come" is written with an old pictograph for wheat, as the word for wheat was pronounced the same way. He repeates this simple fact a number of times throughout the book, thus working exactly as he accuses his opponents of doing: exaggerating the importance of a small number of characters and ignoring the other types. But the real dishonesty comes when he blithely refers to a large number of simple pictographs as "independent phonetics". By this, he means something like the character "ma" for horse. In the words for mother, and ant, and question, and scold, the "horse" pictograph is indeed phonetic: it tells you how the character is pronounced. But in the word for horse itself, the character is neither phonetic nor following the rebus principle. It means horse, and derives from a picture of a horse. Now, you could argue that once we are familiar with the phonetic nature of the characters which include this element, we can read that information back to the original character itself, but to simply include a great mass of non-phonetic characters among the phonetic ones, in an author who makes so much of strict accuracy and who is so hostile and, I must say, rude to everyone who does not share his own rather sterile concern for semantics, is unforgivable.
4. I want to look more closely at the way he constructs his 'phonetic' argument. Typically, he debunks a somewhat false impression created by the loose talk of people who are not approaching Chinese characters in the spirit of a math teacher: that someone who knows Chinese well can guess the meaning of a character even if he's never seen it before. He refers to a guessing game or a game of twenty questions. But he ignores three things. First, the semantic elements, the radicals, are, if not so helpful as others might have suggested, more helpful than DeFrancis leads us to suppose. Imagine seeing the sentence "There was an X in the garden", X being a word I have never seen before. In English, or in Chinese written with pinyin, there is very little in the context to tell me what X is. In Chinese characters, on the other hand, there might be a great deal. Is there a grass radical? X is almost certainly some kind of plant. An insect radical or a mammal radical? Almost certainly an insect or reptile, or a mammal. I can attest from my own reading of Chinese books that this sort of thing is enormously helpful with some regularity. Secondly, he writes as if the only value of the characters in their semantic aspect is to help someone who has never seen a character before guess its meaning; he ignores the great usefulness of the semantic aspect in remembering the meaning of a character which we may have seen before but not deeply learned. Finally, and amazingly in a book that includes a whole chapter debunking the so-called "monosyllabic myth", he here behaves as if Chinese were indeed monosyllabic: apparently it suits him. Many Chinese words are made up of two characters. Even if I know both characters, I don't necessarily know the two-character word they compose. But perhaps I can guess. Now there are dozens of characters pronounced "yi" and dozens pronounced "shi" and so on. If I simply see pinyin it is highly unlikely that I will know which "yi" is meant, but if I see the character, and I know it is the "yi" that means medicine (not the one that means justice, or interest, or already, or any of the others) I have a good chance of guessing the meaning of the word. If I see pinyin I have no chance.
5. After all of this preparatory obfuscation, he descends into the lowest point of the book. He conducts two
"little tests" to find out whether Chinese readers use the ideographic elements in reading. First, he writes a simple sentence that means "I have one older brother and two younger brothers.": "Wo you yi ge ge ge, liang ge di di." Five of the characters he writes correctly; three (two are the same character reduplicated) he writes with another character pronounced the same way. Unsurprisingly, his Chinese readers were able to decipher the sentence. Why didn't he try simply writing every word with the wrong character? Amazingly, he later on contradicts his own point: he tells a story in which, instead of writing the characters "fen dou" meaning "manure bag", some urban cadres wrote the wrong characters: ones meaning "conflict". Of the remaining three characters in the sentence, two were correct and one had the equivalent of a minor spelling error. Nobody, it seems, understood the sentence. So which is it, Mr. DeFrancis? If I write a Chinese sentence with characters which are homonyms of the ones I mean, will people understand it or not? Apparently, when he wants to argue that only phonetics counts, they will; when he wants to show how hard and unreasonable Chinese writing is, they will not.
6. He then proceeds to a "test" which I defy anyone to absolve of gross dishonesty. He takes the first sentence of a story of Lu Xun and rewrites it two ways: one containing only the semantic, and one containing only the phonetic information. He triumphantly emerges with the information that the phonetic elements went much farther to suggest meaning than the semantic ones. (never mind the fact that the whole point of Chinese characters is for phonetic and semantic elements to work together.) But wait--he has dishonestly claimed phonetic information for no fewer than ten of the characters in the sentence. One example: Duo, meaning many. It consists of two characters both pronounced Xi. Where is the phonetic information telling us that the word is pronounced Duo? Perhaps he is using his own (false) principle of reading phonetic information back on a simple character used phonetically in
later compounds? But although there is one rare character, having the duo element as a phonetic, which is indeed also pronounced duo, the others in the group are pronounced respectively die, yi, and chi. The mere fact that one single character uses duo as a phonetic element is not enough to reflect a phonetic nature back onto a character which is purely ideographic. Even this shadow of justification does not extend to his inclusion of jin (today) tian (heaven) yue (moon) and the others. On the other hand, he ignores obvious semantic clues. Guang, in which fire and legs, suggestive of a man holding a torch, combine to mean light, is a clear example of the sort of ideographic compound whose existence DeFrancis would love for us to forget: in his supposed semantic rewriting of the sentence, he only gives legs, ignoring the fire!
7. It would be tedious to proceed in this way through the morass that is the remained of this book, but DeFrancis supposed debunking of the "universality myth" has only to be read by a careful, independent-minded critical reader to be laughed at for the absurdity it is. The literate product of Hong Kong, Tokyo, Shanghai, or Seoul, educated traditionally in his own language, can communicate with someone from one of those other cities, even if the two do not speak a single word of the same language, in writing. I myself have profited from this: In Tokyo I was able to ask for a bank, and tell people in a hotel that I wanted tea, and several other things, simply by writing Chinese. DeFrancis preposterous attack on this simple fact consists of saying that if the man from Hong Kong and the man from Beijing are illiterate, they have to spend several years learning to write (and therefore, first, to speak) each other's languages before they can communicate in writing. Since for a Spaniard and a Frenchman the task would presumably be easier if the various Chinese-character-using languages used alphabets. This is the first hint of the obsession with illiteracy which takes over the second half of the book, but I wonder how a publisher let this nonsense get past his desk with a straight face. The Spaniard who is literate in Spanish needs to learn French--still a task of a year or so however closely related the two languages are--before he can correspond with the Frenchman. The literate, educated Japanese can correspond with the literate, educated Beijinger without learning a day's worth of spoken modern Chinese. What part of that does DeFrancis not understand?
8. From someone who knows as much about language as DeFrancis, we had the right to expect the chapter on the
"monosyllabic myth" to contain fascinating and penetrating analysis: it could have done so. Instead, DeFrancis gives us an absurdly simplistic picture. In English there are "bound forms", like the two syllables in "coral": they cannot exist without each other; "semi-bound forms" like the -er in teacher, which can exist in other words but never on its own; and "free forms" like the "teach" in "teacher", which can exist on its own. DeFrancis, obsessed with proving that there is nothing in Chinese which we don't already have in our plain English, tries to force Chinese words onto the procrustean bed of this scheme. But already his analysis of English is flawed. The two syllables of "coral" are not at all unable to exist apart from each other; they are, indeed, found in many other words. It is only that in the word coral, neither "cor-" nor "-al" has any independent semantic
significance. Now, there is no Chinese character, or at least there are very, very few, which have no semantic significance. Even the two characters in the Chinese word for coral cannot be compared to this: the very nature of the Chinese character, with its semantic element which DeFrancis is trying to hide from us, preclude it. DeFrancis compares "Yuan" to the "-er" in teacher, but it would be better to compare it to the -man in workman: the equality which the Chinese written character imposes on both halves of the word, both in that they are each pronounced with one syllable and in that they take up the same amount of space on the page, militates against DeFrancis' attempt to compare any chinese character to the very subordinate, secondary role played by an English suffix of prefix. (Another quibble here: DeFrancis blithely labels the Chinese "jiao", teach, as a phonetic word. This is insupportable. Half the character is purely semantic. The other half may be described as basically semantic as well, the whole character thus forming an ideographic compound with a phonetic hint in addition, or as basically phonetic with a semantic hint in addition. Every Chinese dictionary I have describes the etymology of the character in purely semantic terms. Even to describe it as phonetic-semantic exaggerates the phonetic element, but to call it "phonetic" pure and simple is a lie, like so much of this book.) What DeFrancis ignores here is the degree to which Chinese differs from English: the hundreds of thousands of words or phrases which are composed of two characters, combined and recombined again and again to form discrete meanings. These are fifty/fifty partnerships quite distinct from the major/minor partnerships of English words with a primary root and a suffix or
prefix, and totally unlike the majority of English words whose individual syllables have no individual semantic significance. For DeFrancis to write a whole chapter on the topic and ignore all this is astonishing.
9. DeFrancis behaves as if the "six myths" program absolves him of the responsibility of building an argument in which the conclusions rise logically from the premises. His basic system is, from the fact that there is a phonetic element and the fact that a number of writers who were concentrating on other things did not talk about this fact, to leap to the false conclusion that Chinese is purely phonetic. He next, using a sort of gospel of efficiency which suggests that a written language cannot do anything other than transcribe sounds, blithely assumes that since an alphabet does that job better than Chinese characters can, an alphabet is superior. The second half of his argument is purely bizarre: he devotes essentially the whole of the second half of his book to the question of illiteracy. But, however important the question is, to discuss language policy and take nothing but that question into consideration, would be like writing a book about the whole nutritional policy of a large nation and
discussing it purely from the point of view of the nutritional needs of people suffering from celiac disease. Secondly, while the major effort of extending literacy must surely involve the teaching of children, he focuses most of his attention on the learning of writing by illiterate adults, a difficult task in any language. But even here his argument is bizarre. Surely the impressive literacy of Taiwan demonstrates that Chinese characters do not pose an undue burden on the aquisition of the ability to read by ordinary people? No, he says; in a laughably short and dismissive section, he simply declares that it is Taiwan's middle class status and government efficiency which makes it possible. But do sane and honest debaters not believe in comparing like to like, apples to apples and not to oranges? The illiterates of China are a) mostly poor peasants living a lifestyle in which literacy is not obviously necessary and is not encouraged by any local tradition and b) the survivors of the triple trauma of the the wars of the first half of the century, the famines associated with the Great Leap Forward, and the chaos of the Cultural Revolution. Therefore, says DeFrancis, if they are illiterate, it is the Chinese written Character which is to blame? And if Chinese Characters have not hindered the 23 million people of Taiwan from becoming literate, it is not the Characters which are to be credited but their middle-class lifestyle? How did this book get published?
10. On a sidenote, the supposed debunking by DeFrancis of what he calls the myth of Japanese literacy amounts to little more than that people attain the literacy they need for the life they choose to live. How literate does he think the average American is? He calls the Japanese partly illiterate because not many of them can write on a sophisticated level? How many Americans does he think can even read at that level?
11. DeFrancis knows that a great deat of Chinese cannot be written alphabetically and understood; there are simply too many homonyms. (See his attempt to deny this fact for an egriously bad arguement--even at the end of my second year of basic Chinese, I already had more than a dozen fourth-tone "yis" in my notebook.) He tries to solve this problem by declaring that spoken Chinese can be written alphabetically, whereas classical Chinese "of course" cannot (of course we get no analysis of the illuminating facts behind that "of course") and the blithe suggestion that all we have to do is expel those nasty vestiges of classical Chinese from the modern language and all, alphabetically speaking, will be well. Having argued this he has a responsibility to analyze the language for what can and cannot be written alphabetically; he does not do so and does not even approach the task. One gets the impression that he simply has no idea of the gigantic impoverishment his blithe suggestion would wreak on a great language.
12. The argument from "waste of time" is ludicrous. Even allowing that the Chinese child spends more time learning to read and write than does the child in an alphabet-using country, a couple of points bear keeping in mind. DeFrancis behaves as if, once the alphabet is learned, the American (or French, or Russian, etc.) child has completely his task. Remember that the acquisition of Chinese characters keeps pace with that of vocabulary; once you remember how many discrete vocabulary words the learner of a European language must learn, and how, in doing so (unlike in learning Chinese) there are no semantic or ideographic elements in the characters to help, the difference is far smaller. But again, even allowing that the Chinese child does spend more time learning to read and write, where does DeFrancis propose that all those "wasted man-hours" be spent? The last time I looked, the child, both in China and everywhere else, was already using many of his precious "man-hours" playing video games, watching TV, and moping about the house complaining of boredom. Is it for more of such pursuits as those that DeFrancis hopes to free some of the hours now "wasted" on the acquisition of a writing system which is one of the great treasures of world culture? Give us a break. I am reminded of the man in Saint-Exupery who bemoaned the many minutes we "waste" drinking water and hoped to invent a pill to end all that inefficiency.
13. The depressing effect of reading this book comes from several things. First, there is the mind-numbing repetitiousness. DeFrancis on page 69 gives us a quote from Boodberg. He repeats the quote on the bottom of page 144 and then repeats it again, a mere two pages later, at the top of page 147: the total effect of this sort of thing is to imply that he imagines his readers to be nitwits who cannot remember anything from one page to the next. Then there is the unwholesome air of mockery which constantly underlies the surface geniality of his prose: the anecdote at the very beginning, in which an imaginary Chinese dinner guest making what turns out to be a perfectly sane response to a vague question is called "incongruous" and "muddleheaded" is only the beginning: I have never seen such a procession of phantoms and straw-men, created only to be mocked and debunked. The real people behind the supposed myths DeFrancis attacks are a couple of French sinologists, Margulies and Etiemple, and H.G. Creel, a giant of the field. Never does DeFrancis really share with us the pith and marrow of those men's arguments so that we could objectively analyse the arguments DeFrancis makes in reply. Even worse is the total absense of a reasonable, balanced, and varied picture of the views on this subject matter of literate people in China and in the Chinese-character-using world. There is a very small handful of Chinese proponents of romanization whom DeFrancis quotes again and again--it seems that one of them occurs on virtually every single page of the closing section of the book. This is the way cult-members write: in their minds the leaders of their cult loom so large, and everyone else so small, that it seems perfectly appropriate to support an argument by simply quoting again and again the one or two fellow-travellers with whom they share it. Someone who wanted to write a useful and unbiased book would have interviewed all sorts of people all over the Chinese world, as well as in
Korea and Japan, about all aspects of the Chinese character.
14. The worst thing about this book is the depressing inability of its author to see anywhere outside of the narrow circle of his breath-takingly limited view of efficiency. No doubt there are writers who have carelessly made unprovable claims about great food: that it elevates the soul, brings joy and peace, and promotes warm relations between human beings. Supposing I wrote a hundred pages debunking these myths, showing that great food, just like mediocre food and bad food, went into the digestive tract, served a certain biological purpose, and then got expelled from the body; and if I then proceeded in another hundred pages to prove that thefore, since mediocre cooking and fast food do all the same jobs with far less wastage of time, we ought to get rid of that inefficient time-waster, good cooking; or if I said that great architecture was less efficient than the cubicle, or that the robot was more efficient than the human being, and that therefore men and women and great works of architecture ought to be jettisoned in favour of the robot and the cubicle; and if I maintained this, in a common-sense sort of tone, mocking everyone foolish enough to have used imprecise of unscientific language in the false praise of those inefficient things, human beings and great architecture and good food--this would be the sort of thing which DeFrancis has done to the Chinese written language. Perhaps before DeFrancis wrote there really was so much imprecision in the use of certain terms that something like the glass of cold water which his book represents was necessary as a corrective, but having said that I find nothing else good that I can say about it. Can we now have some people writing about the Chinese language who actually have a minimal amount of human feeling for their subject?