The principal function of the Academy will be to work with all care and all diligence possible to give certain rules to our language and to render it pure, eloquent, and capable of addressing the arts and sciences.
—STATUTE OF THE FRENCH ACADEMY
On May 4th, 2004, one of the world’s oldest written languages underwent an abrupt official change. The government of Taiwan ordered that from then on, government documents in Chinese must be written from left to right. No longer could Chinese be written in its traditional ways, top to bottom or right to left. English characters and Arabic numerals running left to right, together on a page with Chinese characters running right to left, “looked confusing,” said a government spokesman. Modernization required a single standard. That standard, naturally, was that of the world’s dominant language, English—even though written English was centuries younger than written Chinese.
Most of the forms of top-down language policy—the kind that comes from governments or official and semiofficial elites—answer the question of what language, or what form of a language, will play a given role in society. What will be our official language, favored second language, liturgical language, language of education? But bureaucrats sometimes also sit down and make decisions on the actual form of a language—how it is printed, spelled, or spoken, which words are allowed and which forbidden, which alphabet must be used, and so forth. Linguists call it “corpus planning,” the creation of rules affecting the body of a language itself.
Using state power to compel people to speak and write their native language in a certain way is a modern phenomenon. For the tens of thousands of years humans have been talking, they have mostly just done as they pleased. Communication was the only important thing that mattered to almost anyone. True, we have seen grammarians and language scolds as far back as ancient Rome. But they were mainly people who tried to lay down rules for the literary and formal spoken language only, and then only by example. The idea of passing a law telling people how to spell, which words were forbidden, or which grammar innovations were acceptable would have been hugely odd until modern times.
But the gradual development of nations and nationalism changed that. Building languages was necessary for building nations. In the last chapter, we saw leaders banning competing languages or dialects, in the name of gathering one nation into one state and giving one chosen language top-shelf status. In the modern world, they have gone a step beyond that. The long arm of the modern government has been tempted to fiddle with the language rule book itself: governments have banned words and phrases, coined new ones from thin air, changed the writing system, and otherwise used the power of the state to influence the natural growth of languages. The rougher governments of the world threaten harsh penalties when their linguistic laws are not observed: using banned “impure” words or a writing system that has fallen out of favor, for example. This kind of linguistic activism by politicians has rarely been successful. Yet still they try.
One of the most common rationales for top-down efforts to shape a language is modernization. Citizens are often told, as in the case of Taiwan’s writing, that they must submit to an official change to their linguistic tradition because the demands of the modern world require it.
The world’s societies have developed unevenly in the past five hundred years. Beginning in the middle of the last millennium, the European powers became the world’s undeniable masters of technology and modern development. Later, their New-World offspring, and especially the United States, would join them in global dominance. From ocean-crossing vessels to railroads, from the telegraph to the telephone, from radio to television to the Internet, from the discovery of oxygen to the discovery of DNA, countries speaking European languages led the way in the world’s modernization.
This was unsatisfactory for two other kinds of societies. Postcolonial states born in the twentieth century wanted to show the world that they were every bit as sovereign as the countries that had formerly defeated, dominated, or colonized them. This meant having modern languages that could cope with all of the world’s technical and scientific challenges. The founders of Tanzania would seek, for example, to modernize Swahili—an East African lingua franca of African descent with a layer of Arabic vocabulary—to earn it a place among the world’s must-know languages.
Another kind of society felt the need for language modernization keenly too: not newly independent, identity-building nations like Tanzania but some of the world’s proudest countries, formerly powerful empires fallen on hard times. Countries such as China and Turkey could draw on hundreds of years of written history and a proud past of dominating huge patches of the world. To those former great powers, it was unacceptable to have their languages seem less than fully “modern,” capable of holding their own among the other great languages of the contemporary world.
In most of these cases, modernization has been synonymous with “Westernization” and sometimes even “Anglicization” or “Americanization.” The direction of a modernizing change is frequently in the direction of using the conventions, vocabulary, and sometimes even grammar of the main European languages, especially English. Traditionalists, of course, have resisted.
In other cases, language “reform” is really nationalist purification—the rejection of pronunciations, words, and even elements of grammar that come from another language. Such a reform is often sold as “modernization,” too, even when the offending elements come from “advanced” languages. Instead of modernization, these purifying reforms are really addressing identity and insecurity: the fear that if a language becomes impure by borrowing, the nation itself will become corrupted.
These mixed, sometimes conflicting motives of modernizing, Westernizing, and “nationalizing” of a language are shown vividly in the language reforms in Turkey beginning in the 1920s. Ottoman Turkey was one of history’s greatest empires, the equal of the Roman, Chinese, or British empires at its height. The Ottomans conquered Constantinople, a Greek-speaking center of Orthodox Christianity, in 1453. They renamed it Istanbul, turned the great Hagia Sophia church into a mosque, and proceeded from there to march farther into Europe, even reaching Vienna. The Ottoman state also expanded into the Middle East, extending nominal or real suzerainty over the Arab lands as far as Morocco and including the Levant, Syria, Iraq, and most of the populated parts of the Arabian Peninsula.
Like all empires of its size, the Ottoman Empire was heavily multilingual. Turkish, written in the Arabic alphabet, was the language of the Ottoman state. But Arabic, the sacred language of the Qur’an, was the indispensable language of religion. Despite their political dominance, the Ottoman Turks never sought to interfere with the crucial religious role of Arabic (and of course the empire included millions of Arabs). Persian, too, remained a major literary and diplomatic language, thanks to the presence next door of the culturally influential Iranian Empire. The empire’s other language groups included many others from the Berbers of North Africa in the west to the Armenians of the east.
The Ottoman defeat in the First World War finished the long dismemberment of an empire that had weakened over the previous centuries. The Ottoman state and its sultanate were abolished, along with the Islamic caliphate—the symbol of the unity of the worldwide Muslim community. After Greece invaded the remains of the empire, however, the Turks fought back. The nationalist officers who sought to reverse Turkey’s humiliation were ably led by Mustafa Kemal, who would later take the surname Atatürk (“father of the Turks”).
In the aftermath of the war with Greece, 1.3 million Greeks were removed from Turkey, while 400,000 Turks were expelled from Greece, a forced migration, though agreed upon by the governments, that would today be called ethnic cleansing. Following the Ottomans’ bloody campaign against the Armenians—which many consider the twentieth century’s first genocide—the new Turkish state was more Turkish than the Ottoman one had ever been. No longer a multilingual empire, Turkey was now a modern, ethnonationalist republic, based on the land of the Anatolian Peninsula and on the Turkish people itself.
Ottoman Turkish had been a mixed product. Turkish belongs to a larger group of Turkic languages with origins in Central Asia. But by the end of the Ottoman period, the language also included a large tranche of loanwords from Arabic and Persian. Those two languages, with their cultural and religious prestige, even influenced the grammar of Turkish. Persian, however, is an Indo-European language (related distantly to English), while Arabic is a Semitic one (related to Hebrew, Aramaic, Akkadian, and others). Reforms of the 1830s to 1850s had gone some way to removing foreign influence, but much of it remained, so much that even basic words such as “language,” lisan, were Arabic.
This was a prime situation for a would-be language autocrat: a newly independent nation, formed mainly of one ethnic group, with a proud history, a recent humiliation, and powerful, equally assertive neighbors. A newly purified language would be the vehicle for a thrusting new nationalism. Atatürk would separate Turkey not only from its neighbors but from the Ottoman past: he wanted a republic based on nationalism, not religion. So to those ends, he undertook one of the most abrupt and far-reaching language reforms in recorded history.
At a stroke, Atatürk banned the use of the Arabic script for writing Turkish. This former alphabet reeked too much of Turkey’s Muslim identity and of its connections to the rest of the backward Middle East. Instead, Turkish would now be written with the Roman alphabet of the Western European languages, of which Atatürk himself spoke French and German.
Language overhauls like Atatürk’s are usually made on purported linguistic grounds. Proposed changes are sold to the people as more logical, more expressive, or truer to the native genius of the people. Sometimes this has the virtue of even being true, and in this case it was. The Arabic alphabet was poorly suited to Turkish: only the long vowels a, i, and u are written in Arabic, and none of the short vowels are written at all. Turkish, however, features eight vowel sounds, today written a, e, i, ι, o, ö, u, and ü. A few Turkish consonant sounds also required repurposing some Arabic letters. Meanwhile, certain Arabic letters were not needed at all in Turkish.
But other languages, like the Urdu of India and Pakistan, and the Pashto of Afghanistan, are also unrelated to Arabic but nonetheless written in Arabic script. Modifications allow non-Arabic sounds to be written in those languages and Turkish too. And Ottoman Turkish, imperfect as its writing in Arabic script was, had served the empire for hundreds of years. Moreover, Turkish is genetically no closer to the European languages than it is to Arabic. So the move to the Latin alphabet wasn’t obvious on linguistic grounds alone; an alphabet could have been invented, for example, tailored to the sounds of Turkish. Atatürk’s real motive was purely political: by changing scripts, he wanted to bring Turkey out of the Middle East and into Europe.
To this end, he not only toured the Turkish countryside, teaching bewildered Turks their new alphabet on a chalkboard, he also decided that his language reforms needed the hard hand of state backing. It became illegal to write Turkish in the Arabic script on January 1, 1929. An alphabet was now a crime.
But the script was not the only object of Atatürk’s reform. Arabic and Persian influences went deeper, not only into thousands of important words (especially in written Turkish) but also even into grammar. To carry out the wide-ranging reforms that Atatürk wanted required a new Turkish Language Association. Its remit was no less than to make a new language.
Turkish grammar was so heavily influenced by Arabic and Persian that half of a 1904 grammar book of the Turkish language, written in English, focused on teaching Persian and Arabic grammar. For example, native Turkish word order places adjectives before nouns, but Ottoman Turkish often adopted the Arabic order of noun before adjective and the Persian way of inserting “i” between a noun and adjective. The very name for the “Sublime Porte,” the seat of the Ottoman government in Istanbul, was bab-i-ali: two Arabic words (“gate” and “high”), in Arabic order, joined using a Persian grammatical convention. (Imagine the White House being known as Das Maison Blanche.) Little wonder that a nationalist Turk could see his language as corrupted.
But Atatürk’s solution was one of the most concerted assaults on a nation’s linguistic heritage ever seen. The Language Association was tasked with coining new, purely Turkish words for Arabic and Persian ones. It took to its task with a vengeance. Words were either coined entirely from existing roots found in old Turkish writings or gathered in surveys of people’s speech from around the country. This resulted in more than 100,000 new words. Newspaper editors were required to comb through articles and replace borrowings with the Turkicisms prescribed in an official new book. But that book was too generous. To replace the Arabic borrowing for “pen,” for example, the editors had six choices, all equally obscure to most Turks. Chaos reigned.
Sometimes no Turkish word could be found. In those cases, the Arabic or Persian word was often kept—so long as it could be given a semiplausible Turkish etymology. This led to much amusing fakery. Atatürk himself was an exemplar; an amateur etymologist, he took on asker, soldier. It had been borrowed from the Latin exercitus via the Arabic ’askari. But Atatürk convinced himself that the true origin was the Turkish asik, “profit,” and er, “man.” A soldier, after all, profited his country.
The climax of this orgy of linguistic silliness was Atatürk’s “sun theory”: that Turkish was the world’s first language and so all words that had been borrowed into Turkish had really originally been taken from Turkish. The theory was patently absurd, but it did solve a problem. The linguistic renovation had gotten out of control. In the face of so much difficulty and confusion, the forced Turkization of the language could now be slowed.
Geoffrey Lewis, an Oxford scholar of Turkish, called Atatürk’s reform “a catastrophic success.” A new Turkish did indeed exist, but it was so altered and invented that it was no one’s native language. When the new Turkish was still young, Atatürk once even embarrassed himself by giving a speech in which he had replaced all foreign loanwords with his favored Turkish neologisms. An observer noted that he spoke “like a schoolboy who has just begun to read.”
But the overhaul of Turkish was nonetheless eventually successful in Atatürk’s stated aims. The reform did indeed rid the language of most of its foreign influences. The combination of the writing reform and his purge of the vocabulary and grammar left modern Turkish a substantially different language. Over time, thanks to the heavy but guiding hand of the state, Turkish editors of books and newspapers learned to use the new words and readers learned to read them. Turkish was reborn, but at a cost. Today, Turks cannot read most texts from the 1930s and before unless they are “translated” into modern Turkish.
Atatürk was a man so eager to break with the past that he ordered the banning of a hat—the traditional fez. This was just another manifestation of his urgent modernizing nationalism, willing into existence a new language that dragged Turks away from their heritage of the Ottoman Empire and the rest of the Islamic world. Though Atatürk’s interest in language was keen, his real motivations were not linguistic. Language reformers like Turkey’s first president talk about etymology and grammar, modernization and progress through a renewed language. But their real motivations are political. And in a dictatorship, like early Turkey, they can even succeed.
All this may seem distant and perhaps faintly amusing. But imagine that America had lost a war to France and, in the wake of its humiliation, tried to distance itself from the French language. Not only would obvious borrowings such as tête-à-tête have to go; a rich layer of vocabulary now considered perfectly good English—“royal,” “guarantee,” and so forth—would have to be replaced too. Vestiges of foreign grammar would have to be jettisoned as well. “Attorney general” and “surgeon general,” which reflect French word order, would have to be reversed. (Of course “general attorney” wouldn’t do either, as both words are Latinate; they would need replacing. “Head lawyer,” perhaps.) The suffix -ee, which makes possible words such as “employee” and “refugee,” would have to go too. Its closest native equivalent would be -ed. But of course “employ” comes from French too; an employee could now perhaps be a hired. Civilization itself would come to an end, since “civilization” had come from French. If all that seems too silly, just picture English-speakers not being able to read works as recent as those of Dickens or Henry James without a specialized education, or if their great-grandparents’ old letters were in a foreign alphabet. That is what Atatürk did for his countrymen.
He didn’t stop at encouraging patriotic Turks to use native words, he ordered them to, first banning the script and then forcing a new dictionary full of strange words on writers and editors. To ban an alphabet, and words themselves, is almost impossible to imagine in a democracy. Only a nation in the grip of both nationalism and anxiety would try it, and only a dictator could succeed.
Perhaps Atatürk dragged Turkey into the modern world. But he did so using means that no liberal-minded observer should approve. Turkey today remains plagued (though many Turks would say “preserved”) by Kemalism, the aggressively secular and Western-looking philosophy that the republic’s founder espoused. It remains illegal for women to wear the Islamic head scarf in parliament and in universities. The mildly Islamist prime minister, Recep Tayyip Erdogan, was once jailed for reading a poem that seemed to authorities to be too Islamic. Turkey has experienced several constitutional crises, coups, and coup threats as traditional forces close to Islam do battle with the Kemalist guardians of modern Turkey, represented by the army. But while Islam is curbed, secular nationalism is sacred. It remains a crime to defame the republic, “Turkishness,” or Atatürk himself, punishable by up to three years in prison. Ninety years on, the “success” of Atatürk’s language reforms remains in place, but the Kemalist foundations of the republic are cracked and in need of some renovation.
If Turkey’s language reform was an all-out assault on the past, the history of language planning in France is comparatively harmless. But while the Turkish case is little known in the West, the ineffectual meddling of the French is just too tempting for Anglophone journalists to ignore. For this reason, the French are known, not without some justification, as the most prickly protectors of their own language in the world.
In 1998, the BBC reported that the French president, Jacques Chirac, had received a letter that “demanded” the reversal of a creeping trend within his government. Politicians, of course, are accustomed to such letters. But this one dealt not with the economy, national defense, or education but with the weighty question of grammar—specifically, gender-specific articles. Women in the French cabinet had begun answering to madame la ministre, or Ms. Minister. According to the letter, written by the French Academy, the guardian of linguistic purity in France, they should be called madame le ministre. No matter that they were women and la is the feminine article in French. Ministre, whether male or female, was grammatically masculine, argued the Academy. Not only that, but allowing the la would only exacerbate the notion that men and women ministers were different. La ministre was an affront to equality. Man or woman, said the Academy, everyone had the right to be le ministre. But the government did not change its practice.
In 2007 the BBC was at it again, reporting that “a new French resistance” was under way. A center-right member of parliament was arguing publicly that the invasion of English words was very dangerous, because “the French language is the spirit of France and of every Frenchman.” A union leader bemoaned the fact that 7 percent of French companies were using English as their official language. The article cited les e-mails, le web, and l’Internet as proof that the English were invading back across the Channel. The British press loves a good tale at the expense of the French—perhaps in revenge for the Norman Conquest of 1066 and all that.
But this genre of story also works so well because the French passion for the language is so well known. French waiters and shopkeepers are thought to be rude to customers who don’t speak French. French writers publicly decry their eroding heritage, and French politicians sometimes refuse to speak English, even when they can, to the outside world. And then there is the famous Academy, which journalists so enjoy chronicling as it tries to stave off English words and natural language change.
All this overlooks the fact that the modern, standard French language itself is hardly the pure maiden its defenders would have it be. No doubt recent years have seen the import of Web and Internet, and earlier decades saw le week-end and le foot (soccer). But French borrowing goes back into history well beyond the supremacy of wily capitalist Anglo-Saxons and their devilishly successful ways.
One scholar has counted 2,613 borrowings from English in a modern dictionary of French words of foreign origin. But she also counted 1,012 from “Gallo-Roman dialects,” Romance languages related to French such as Provençal. She also found 694 words from Old German and 408 from modern German. The older borrowings include such venerably French mots as bleu, blond, and blanc, to name just color words, and soupe, haïr (to hate), and honte (shame). Whatever modern defenders of purity may think, France has been borrowing for a long time. L’Internet, in its way, is part of a grand French tradition.
This is all too easily forgotten by the modern era’s French elites, who, like declinists elsewhere, have for many years been convinced that their language is under threat. Unlike many others, they have been unusually active about addressing the “threat,” and the Academy embodies their determination to protect the language.
Cardinal Richelieu, the first minister to King Louis XIII, established the Academy in 1635 to promote the refinement of French letters. In 1694, it published its first dictionary of the French language. Its members have included Pierre Corneille, Jean Racine, Voltaire, and Victor Hugo. Limited to a total number of forty, they serve for life and are humbly styled “The Immortals.” They occupy not “chairs” at the academy but “armchairs” (fauteuils).* The Academy is thus sometimes referred to synecdochically as les 40 fauteuils. And well they might want comfortable chairs: in 2008 the average age of Academy members was seventy-nine years old.
Election to the Academy is an elaborate ritual. New members are chosen by the sitting members. Upon being selected to fill an empty fauteuil—which can take more than a year—the new member makes a speech eulogizing the member he has replaced. The new académicien must then listen to a speech by a sitting member. After eight days, the newcomer must once again make a speech, this time giving thanks for being chosen. Academy members wear traditional green garb, hats, and ceremonial swords. Perhaps unsurprisingly, a few iconoclastic or loner writers have refused to join the elaborate ritual; Marcel Aymé, a novelist, said upon being invited that he was a solitary type who wouldn’t know what to do in such company. Others, including Descartes, Molière, Balzac, Flaubert, and Proust, for one reason or other were never invited, were rejected, or died before they could take their seats; such eminent non-academicians are collectively known as the “41st fauteuil.” Besides them, there are two other groups that have rarely joined the Academy: only five women in its history (the first in 1980) and only two full-time language scholars in the last century.
A 1996 letter to the conservative newspaper Le Figaro by one of its members, a literary historian and essayist named Marc Fumaroli, captures how the average académicien sees his language in relation to the world’s others:
Well-taught French is more than just French: it is the human spirit placed in possession of a symbolic system which opens the door to all the others … the Latin of modern times is in itself a human education.
The Academy may seem merely stodgy to the outsider, but its habits cross the French political and intellectual spectrum. A detour in the Academy’s history is illustrative. After the French Revolution of 1789, the new republican government suspended the Academy in 1793, as it did all the royal academies. To the instincts of the revolutionaries, the Academy smacked of the Ancien Régime. (Later, Napoleon would restore it.) French political culture has long included an authoritarian strain that respects powerful men such as Napoleon, Philippe Pétain, and Charles de Gaulle and is conservative in believing in the virtues of tradition, the Catholic Church, and elites. The revolutionaries wanted to do away with all that.
But the revolutionaries introduced another cultural strain that, for the purposes of language, is little different from the elitist tradition it replaced. In opposition to the authoritarian streak, the revolution brought what the French call “republican” values. These include egalitarianism, secularism, rationalism, meritocracy, and science. Its radical form was Jacobinism in the revolutionary period and is reflected in the ongoing French sympathy for communism and socialism today. The products of French “republican” thinking range from the successful (the metric system, invented by the French revolutionaries) to the embarrassing (the new months and ten-day week of the “revolutionary calendar” and the guillotine).
What French republicanism and traditional elitism have in common is that they are top-down, statist, dirigiste. The revolutionary republicans were no less tempted to meddle with language than the elitists of the Academy. In 1794 theater directors were ordered to excise from the plays they produced the noble titles duc, baron, marquis, and comte, and even Monsieur and Madame. (Citoyen and citoyenne were preferred in place of the last two.) The order was repealed after a week. But a public notice in 1799 nonetheless told inhabitants of the capital, “The Citizens of Paris must reshape and correct anything which is contrary to the laws, to decency and to the rules of French.”
Language control was as important to the revolutionaries as it ever was to the old regime before it. According to the two traditional streams of French politics, society must be run either by a great man or by a group of self-selected rationalists who know what is best for you. There is little room in French political culture for the political strand that is called “liberal” in Europe and “libertarian” in America: leave people alone, let them work, live, and love, whether they succeed and fail, and use the power of the state as little as possible.
The two French strands—traditional nationalism and republican rationalism—occasionally join forces. The attempt to eradicate la ministre can be defended on grounds of French tradition (“We have always done it this way”) or on the basis of universalism and equality (“Le ministre can be a man or a woman”). Though France is internally politically divided between left and right, language policy is an area of broad agreement: from right to left, there is a national concord on the need to promote and defend French. The Academy’s members include a former center-right president, Valéry Giscard d’Estaing, and a well-known socialist (and formerly Communist) journalist, Max Gallo. The right and the nationalists support French language planning in the name of national prestige, while the left argues that French is a key to universal values like liberty, equality, and fraternity. But in the end, they support the same thing: government language planning.
The Italians had their Accademia della Crusca before France’s, but the French Academy’s model has by far the most celebrated and imitated. Spain was an early adopter; in 1713, King Philip V founded the Real Academia Española. Sweden founded an academy in 1786 and Prussia one in 1779. (Germans, such as the philosopher Johann Gottfried von Herder, have always had a soft spot for the notion that language represents the organic genius of a people.) And the trend has spread beyond Europe: Israel and Indonesia have language academies; Brazil’s Academia Brasileira de Letras, founded in 1897, so slavishly copied the French model that it has forty permanent members, called Imortais.
But how successful has the French Academy been? Its dictionary is not the most prestigious one in France, though it is widely respected. Its first grammar, having taken three centuries to appear, was considered a “scandalously poor piece of work.” Judging by its own standards, its success should be considered mixed at best. A few simple pieces of evidence demonstrate. At the time of this writing, the phrase madame la ministre appeared 321,000 times on Internet pages indexed by Google. Madame le ministre appeared just 87,400 times, or one-fourth as frequently.
Besides rare cases like madame le ministre, has the Academy managed to freeze “proper” French in place? Many foreign students who first try to use their book-language skills on the streets will answer in the emphatic negative. The ne that is supposed to go along with pas in negative constructions is routinely dropped, especially in speech: Je ne sais pas, “I don’t know,” is usually Je sais pas. The est-ce que construction in questions (Est-ce qu’il est parti?, “Has he left?”) is less and less frequent, and the more elegant Est-il parti? is rarer still. (Most people just say Il est parti?) Even a prestigious academy like the French one can’t simply change natural language behavior. The best it can do is freeze a formal written variety of the language that diverges more and more from the living, moving spoken language and everyday writing.
What about preventing infiltration by English? The Academy’s website frankly admits that the commission is charged with “forging new words and recommending French words in place of English.” In 2008, the institution announced its opposition to some five hundred English words, including blog, supermodel, and Wi-Fi. The Academy is assisted in its work by terminology committees in each ministry, which coin technical vocabulary, under the Délégation Générale à la Langue Française (DGLF). But the march of English has not been so easily stopped by the Academy’s elite or the ministries’ technocrats.
Technology is, of course, the area in which English words have had the strongest tendency to appear in other languages. Courriel, the official word for “e-mail” (borrowed, in fact, from Quebec but no less official in France for that), appears on 16 million French Web pages. But another borrowing is far more prominent: mail (words are often chopped or altered when they’re borrowed) appears on about 122 million French pages, an eight-to-one advantage. But mail in the sense of “e-mail” doesn’t appear in the section of the Academy’s dictionary that would contain it. (The dictionary is updated not all at once but in successive alphabetical volumes. Of the series of volumes making up the most recent full publication, “Logomachie” to “Maîtrise” was published in 2000.) There is an entry for mail—a kind of sporting mallet used for an old ball game, the game itself, and the field where it is played. No one plays mail anymore, while virtually everyone sends e-mail. But only the archaic, not the living, word was included in the Academy’s dictionary, which had not been published in full since 1932–1935. Maybe it will be included it in the next edition, half a century or so hence.
When I visit the French Academy for an interview, I am a little nervous. My French is perfectly workable, but it is hardly Proustian, and I know there’s a chance I will make small mistakes. If the French are famously insistent that foreigners use their language, how much higher a standard will the French language’s own guardian expect from me? Fortunately, my contact is Jean-Mathieu Pasqualini, the Academy’s surprisingly jovial chief of staff, a small, spry man who sounds on the telephone as if he’s about nineteen years old.
A former professor of philosophy, Pasqualini takes me in and defends the Academy’s positions while registering on his face that I don’t always buy them. Boasting that the Holy See uses French for its diplomatic correspondence, he says, “The reason given is the precision, the surety, the clarity of the French language” but goes on to offer, “I don’t know if attributing these qualities to the language makes sense on a linguistic level.” When he notes that the Academy’s dictionary is “normative,” not descriptive, he almost seems sorry to have to break the bad news to me. He thinks that courriel is a delightful word but admits that it is barely hanging on to life.
The French love their language. They even hold televised dictation contests, which bear witness to their fascination with getting their beloved tongue right. And they largely admire their Immortels. But the aversion to English, and the haughty, authoritarian tradition may be more of an elite phenomenon than a nationalist groundswell. Many of the French share Pasqualini’s cheerful realism. In 1994 the Ministry of Culture commissioned a poll, intended to show that the French were worried about “Franglais.” Instead it showed a mixed result: 60 percent said they were “strongly attached to French,” but only 44 percent said that the use of English words was “bad” and 42 percent said it was “good.” Asked to characterize the use of English words in French, 41 percent said it was “modern” and 30 percent chose “useful.” Just 14 percent chose “annoying” and 6 percent, “stupid.”
So the Academy and the terminology committees may be out of touch. The DGLF cannot keep le mail from dominating le courriel. The Academy’s dictionary tries to keep up with the times and accepts some English words (the dictionary currently being produced, in sections, includes dope and joint), but cannot accept le mail. The Immortals’ forceful pronouncement on madame le ministre is robustly ignored by most speakers. Seven of the Academy’s elderly membership died over eighteen months between 2007 and 2008, prompting national concern about how connected to the vital world of a living language its members were. The French may admire the Academy and say that they want to be told how to speak and write. But that doesn’t mean that they listen.
If the dented self-image of a Turkey or a France makes us see their language policies in a sympathetic light, we might have even more fellow feeling with newly independent countries. In the twentieth century, the number of the world’s independent states exploded. Most escaped colonial domination from a greater power. Many had only recently begun to write in their languages. When such nations achieve statehood, we can understand their efforts to shore up their languages, particularly against foreign elements.
The story repeats itself again and again. In Balkan Europe, the new states that broke away from the Ottoman Empire, like Bulgaria and Serbia, sought to purge Turkish words from their vocabulary. The Poles shunned Russian influences. Those dominated by the Austrian empire, like Czechoslovakia and Hungary, sought to cleanse German from the body linguistic.
Even one of the world’s most peaceful, prosperous, and homogeneous democracies has not been free from the purism born of language insecurity. Norway was formerly ruled by Denmark, and even though Denmark was not exactly known for its heavy colonial yoke, the nineteenth century saw Ivar Aasen, a Norwegian linguist, create Nynorsk (“new Norwegian”), a freshly minted standard based on various Norwegian dialects and Old Norse. Aasen’s real target was, of course, Danish. Nynorsk was successful, after a fashion. It has survived and is used by about 10 percent of Norwegians. But the rest use Bokmål, the older, Danish-influenced form. This leads to the absurdity that the modern and ethnically homogeneous state of just 4.6 million Norwegians has two written versions of its language. Divisions over which is the “real” Norwegian divide the country to such an extent that the youth wing of a political party burned a Nynorsk dictionary in one of its television commercials in 2005, causing a mini-scandal.
English has had some attention from purists, too. Great writers such as Winston Churchill and George Orwell have praised and encouraged the use of Anglo-Saxon words instead of their Latin or French equivalents. In the essay “Politics and the English Language” examined in chapter 4, Orwell memorably recast a famous verse from the King James translation of Ecclesiastes:
I returned and saw under the sun, that the race is not to the swift, nor the battle to the strong, neither yet bread to the wise, nor yet riches to men of understanding, nor yet favour to men of skill; but time and chance happeneth to them all.
In Orwell’s disdainful view, had that verse been written in the twentieth century, with Latin- and Greek-derived bureaucratic vocabulary in fashion, it would have come out
Objective considerations of contemporary phenomena compel the conclusion that success or failure in competitive activities exhibits no tendency to be commensurate with innate capacity, but that a considerable element of the unpredictable must invariably be taken into account.
Churchill was another fan of earthy, Saxon words. In one of his most famous Blitz-era speeches to his countrymen, he growled
We shall go on to the end, we shall fight in France, we shall fight on the seas and oceans, we shall fight with growing confidence and growing strength in the air, we shall defend our Island, whatever the cost may be, we shall fight on the beaches, we shall fight on the landing grounds, we shall fight in the fields and in the streets, we shall fight in the hills; we shall never surrender.
Just five of those seventy-five words are Greek- or Romance-derived: ocean, cost, defend, confidence, and surrender. (Even “France” takes its name from the Germanic-speaking Franks.) Churchill liked this rhythm; he said that “short words are best, and the old words when short are best of all.” The “old, short” words in English tend, of course, to be of Anglo-Saxon stock.
But the most active and elaborate English purist is little known and even less followed—which speaks volumes of the Anglophone world’s attitude toward purism. William Barnes (1801–1886) was a schoolmaster, minister, and poet from Dorset, in southern England. He was a keen student of languages—an admiring biographer says he knew fourteen fluently (though for someone who rarely left Dorset and Wiltshire, this seems far-fetched).
Barnes typified a populist purism that is in a way the opposite of the elitist prescriptivism of the French Academy. His most famous poetry was written not in standard English but in his Dorset dialect:
O zummer clote! when the brook’s a-gliden
So slow an’ smooth down his zedgy bed
Upon thy broad leaves so seäfe a-riden
The water’s top wi’ thy yollow head,
By alder’s heads, O,
An’ bulrush beds, O,
Thou then dost float, goolden zummer clote!
Barnes sought not to drag the barely worthy masses up to his lofty language, as did the French Academy. Rather, he wanted to purge Latin- and Greek-derived words from English, replacing them with coinages of Anglo-Saxon stock so the masses could more easily understand and learn literary language. “Photograph,” he though, should be sun-print. “Botany” was to become wort-lore and “enthusiasm,” faith-heat. When not inventing words, he used, and encouraged others to use, old country words that had been replaced with borrowings: inwit for “conscience,” earthtillage for “agriculture,” and bodeword for “commandment.” He used his invented language, too: he wrote an 1878 grammar called An Outline of English Speech-Craft and in 1880 An Outline of Rede-Craft (Logic) in which he employed his inventions liberally.
To those who might call him a rustic boob for writing in his Dorset way, he said, “I cannot help it. It is my mother tongue, and it is to my mind the only true speech of the life that I draw.” This argument is typical; countryside purists like Barnes portray the language of the peasant as the “only true speech,” not layered with foreignisms and fancy words like the language of the cities or the universities.
But even Barnes never sought to push politicians to impose his project. He was a Little Englander (before that term was common), anti-imperialist and apolitical, a member of no party. Deeply religious, he longed for a return to what he saw as an England uncorrupted by materialism and swaggering nationalism.
His work as an idiosyncratic rustic compares starkly with the elite language planners of other countries. The Anglophone world has long preferred a mix of nostalgia, shame, and patriotism to the setting up of a national language academy to rule on proper usage and allowable words. Something in the English mind resisted this even as other European countries set up their language guardians.
When the United States was born, it seemed briefly as though the new nation would engage in some of the kind of top-down language planning we associate today with France. After all, insecurity about a nation’s status in the world often prompts this kind of attempt, and America was just finding its feet.
Noah Webster (1758–1843), who gave his name to America’s most famous dictionary, decided that the new United States needed a new idiom to shore up internal solidarity and to distance America from other powers and its colonial past. He was honest about his motives, writing that even the seemingly prosaic matter of spelling “is an object of vast political consequence.”
But Webster’s results were modest. He introduced some of the best-known differences between American and British English, prescribing “color” for “colour,” “traveling” for “travelling,” “theater” for “theatre,” “check” for “cheque,” and the like. But English speakers today can flit from British to American texts hardly finding those words, or hardly noticing when they do. To the extent that American English differs from the British variety, this has little to do with such top-down reforms as Webster’s. Much more important was the natural additions to the language, from “bronco” and “lasso,” borrowed from Spanish, to Native American words such as “tomahawk” and “wigwam.”
Webster’s new nation flirted with the idea of an academy. The first proposal came even before independence, in 1774. The first bill to create such an academy was introduced, unsuccessfully, to Congress in 1806. Finally, in 1820, the American Academy of Language and Belles Lettres was created, and John Quincy Adams, soon to be the president of the United States, was president of the body. It seemed (“belles lettres” and all) that America might follow the French path.
Mostly ignored and unloved, the American Academy broke up after just two years. It left no mark to speak of on the American language or on English generally. It was never replaced, and America’s short-lived experiment with language planning is so obscure to modern memory that while the Klingon language has a 3,000-word entry on Wikipedia, as of this writing, the American Academy of Language and Belles Lettres has no entry at all.
It is tempting to argue that the French regulate their language only because they are insecure, while Britain, unconquered since 1066, and America, a global superpower, are not. But this is too simple, ignoring the chronology. America rejected government language planning well before it was a superpower. The White House was burned down by the British during the War of 1812, the same era during which the American Academy went belly-up for lack of attention. Meanwhile, the French Academy was founded when, under the “Sun King,” Louis XIV, France’s status as Europe’s greatest power was undisputed.
This leaves us seeking additional cultural explanations. France has both an authoritarian-elitist streak, reflected by its Academy, and a rationalist-“republican”-technocratic tendency, represented by the official terminology committees set up in each government ministry. There is nothing, it seems, that the French can’t improve by decree or regulation. But then again, France has had five republics, four monarchies, and a military dictatorship since its revolution in 1789. Britain has only gradually changed its constitutional form bit by bit since the Magna Carta of 1215, while America has made do with one Constitution, rarely modified, since 1789. If language policy reflects politics, as this book argues that it does, the different attitudes of the Anglophone and Francophone worlds provide a perfect example. The French relentlessly try to regulate and tinker, with mixed results. The Americans and British accept a flawed product, rarely if ever trying to change it with government action, under an attitude that if it isn’t broken, no law should try to fix it.
Though many top-down language policies are overtly nationalistic, many are fairly technical, presented on grounds of practicality. Often they even make sense. The Taiwanese example that begins this chapter may be one of them. It is hard to imagine many stable language situations that would tolerate writing left to right, right to left, and up to down, often on the same page. The Taiwanese government took a fairly sensible position: that in official documents at least, left to right would be the only way to go. The public will probably eventually go along, setting a new, accepted standard.
But even when writing reforms make sense, politics is rarely far from the scene. Usually, some groups win and other groups lose. Those in the dictionary business feed off of public demand for a codified, perfected language. Traditionalists, older people, and those simply opposed to government interference prefer to see language left alone. There is no purely technocratic approach to language.
If there were, we might expect to see it in a place like Germany: a modern, well-developed nation-state that has exchanged its history of nationalism for a reputation for efficiency and rational modern politics. If a country like Germany can’t impose sensible reforms on its writing system without chaos and howls of protest, we should generally be skeptical about virtually any government-run language reform. And, indeed, Germany can’t impose sensible reforms on its writing system without chaos and howls of protest, as the country’s response to its recent spelling reforms shows.
German, like most European languages, has a relatively straightforward letter-to-sound correspondence. The letter z is always pronounced the same way, and ei always rhymes with the English “eye,” never “see,” and so on. But most writing systems have their quirks, and German is no different. German famously runs words together into long compounds, but when the first part of a compound word ends in a doubled consonant and the second part starts with that same consonant, the possibility results of the same letter appearing three times in a row: add Kongress (“congress”) and Sitzung (“session”) and you get Kongresssitzung. Or maybe Kongressitzung—traditionally, triple consonants were reduced to two. Or is it Kongreßsitzung? Another quirk of German is the letter ß, called an “ess-zet” or sometimes a “sharp s,” sitting in for two ss.
Confusion about how to handle triple letters, when to use the sharp s, and a number of other wrinkles led to a movement to reform German spelling. The ß was to be abolished except after long vowels and diphthongs; that meant it would disappear from many common words like daß (“that”) and muß (“must”), which would become dass and muss. The reasoning is that in German, ending a syllable with a double consonant usually means that the previous vowel is short (hoff rhymes with English “off”), while a single consonant means that the vowel is long (Hof rhymes with English “oaf”). The sharp-s change was meant to bring s into line with the other consonants.
Elsewhere, the triple consonants would be restored, but in most cases, the compounds in question would be hyphenated (Kongress-Sitzung). Other changes were made, including to two-part verbs such as kennenlernen, “to get to know,” which became kennen lernen, and modifications to capitalization. But the sharp s accounted for most of the changes the average reader or writer would see or use.
The rules seem complicated to the foreigner, but mostly they made basic sense—as well they should, having been debated from the formation of a working group in 1980 to the announcement of the proposed changes in 1994. Germany, Austria, Switzerland, Belgium, Liechtenstein, and Italy (which has a German-speaking minority in Trentino–Alto Adige) agreed to the rules in 1996, planning to phase them in from 1998 over a period of years.
The intense pushback surprised the reformers. Several German states refused to implement the reforms in their schools. A newly formed group, Teachers Against the Spelling Reform, gathered to resist the changes, with intellectual support from a host of intellectuals including the novelist Günter Grass. (He did not, however, start writing his name Graß.) The state of Schleswig-Holstein held a referendum, which rejected the reform. A legal challenge reached the Constitutional Court, Germany’s supreme court. The court supported the reformers, agreeing that the state culture ministers could put the new spelling into place without further parliamentary approval.
In 2000, the Frankfurter Allgemeine Zeitung, the most prestigious daily newspaper in Germany, announced that it wouldn’t abide by the reform. In 2004, the Springer publishing group, along with Der Spiegel (Germany’s biggest newsmagazine), the Süddeutsche Zeitung (another major daily), and Die Zeit (a famed weekly newspaper) rejected it too. (To get a sense of their media heft taken together, imagine all the Dow Jones publications, plus The New York Times, The Washington Post, Time, and The New Yorker.) Parts of the reform were proposed for rollback, and a modified version of the reform later entered force—though only in fourteen of Germany’s sixteen states. Some three-quarters of Germans opposed the reform, according to one poll; the rest supported either the full reform or some “reform of the reform.” To this day, Germans put up with a partial reform and a chaotic situation; Germany’s neighbors have taken various stances, while the different German states have their own attitudes. The most prestigious publisher of language reference books, Duden, observes a partial version. And most Germans now think the whole thing was a bad idea from the start.
This result of Germany’s botched reform should be a cautionary tale to anyone who thinks that devising a “logical” set of rules—even for something as straightforward as spelling—and then imposing them through government fiat is a good idea. It can lead to disaster even in a modern, fairly homogeneous, and well-run state. Language is usually best left to its own devices.
Even though Germany’s spelling reform was modest and made a good deal of sense, it has produced little but confusion and anger. Why? Maybe language reform simply doesn’t work in a democracy. Whether people say they want to be told what to do by a body like the French Academy, or, as in Germany, they say they don’t, in practice, it’s simply hard to enforce language diktats. All things being equal, people will do what they have always done—write and speak the way they learned in school and go by their instincts when they don’t know the rules.
But we have also seen a language reform that was successful—“catastrophically” so—in Turkey. The difference lies in the political context, as well as the political system. Atatürk was revered. A hero during the First World War, he also turned back a subsequent invasion of Turkey by Greece. Plus, Atatürk governed a state that was, in effect, a one-party dictatorship. The popular autocrat was in a perfect position to drag his language into the future.
There is one major reform that might make more sense than almost any other reform to any language anywhere else in the world. But it will probably not happen anytime in the remotely near future, and only a dictatorship would be able to pull it off. This is the often proposed, but never implemented, shift to the Roman alphabet for writing Chinese and Japanese.
Everyone knows that Chinese and Japanese are written with a set of complicated characters that look nothing like Roman letters. Most know that there are quite a few of these characters. But there are many exoticizing myths about the writing of Chinese and Japanese that need to be dismissed before looking at Romanization.
One belief is that Chinese characters are “pictograms,” direct pictorial symbols of what they stand for. This idea is fostered by menus in cheap Chinese restaurants in America, which invite diners to kill time by reading about how Chinese characters developed from concrete pictures to stylized modern versions.
“Oracle bone”–style character for ri, “sun,” ca. 14th–12th century B.C.E.
“Seal script” character for ri, Qin Dynasty era, ca. 9th–3rd century B.C.E.
Modern character, ri.
This story is so easy and pleasing that most people won’t get around to thinking, before the hot-and-sour soup arrives, “Wait: I can see where it came from, but that last character doesn’t look anything like the sun, really. And how do they depict other things? Does the character for ‘car’ look like a car? And what does the character for ‘intuition’ or ‘to reconsider’ look like?”
Chinese characters are not pictograms for the simple reason that most things can’t be illustrated pictorially. A purely pictographic writing system would never be extensive or flexible enough for a natural human language.
A more tempting version of the legend of Chinese characters is that they are “ideograms,” with characters representing concepts, not concrete things, through visual symbolism. A simple ideogram might be the character for “one”:
Or “middle,” “inside”:
This notion gets more enticing when we hear intuitively appealing stories, as for example that the character for “good” combines “woman” and “child”:
“woman”
“child”
“good”
The relationship between a mother and child can certainly be very good indeed. But would you guess, if you didn’t know, that the abstract representation of a mother plus a child meant “good”? Or would you maybe guess it meant “motherhood,” “parenting,” “filial,” or something like that? What of the many other “good” things that could depict “good”?
If Chinese were really a system of symbolic representation, we should see those who don’t know it able to pick up on the symbolism to some degree. The research, however, shows no such thing. One experiment asked nonspeakers of Chinese to guess which of a pair of Chinese characters with opposed meanings corresponded with their translations. Shown forty-two pairs such as and 
they were asked to guess which meant “good” and which meant “bad,” and so on with beautiful–ugly, heavy–light, weak–strong, alive–dead, and a few dozen others. The average percentage of correct guesses was 54 percent—barely better than chance.
In another experiment with new subjects, the forty-two antonym pairs were lined up in the subjects’ native language (Hebrew) on one page and in different order in Chinese on the other. The subjects were then asked to find the pairs “good–bad,” “heavy–light,” and so forth from among the Chinese pairs. The average subject got just one of the forty-two guesses right. So much for universal symbolism.
But do those who speak Chinese and know the elements gain some kind of wisdom from how they are combined? Maybe the most common ideographic myth among Westerners is that the Chinese for “crisis” combines “danger” and “opportunity.” The appealing notion has been repeated by John F. Kennedy (on the Cold War), Al Gore (on climate change), Condoleezza Rice (on the Middle East), and countless investment advisers urging clients to buy during a dive in the markets. Perhaps the Chinese knew something, in that ineffable Chinese way.
But Victor Mair, a scholar of Chinese at the University of Pennsylvania, has shown it to be untrue. Weiji, “crisis,” is actually two characters, not one, as in the common rendering of the myth. The first does indeed mean “danger,” but the second, ji, means “incipient moment, crucial point,” and the like. So weiji means “the crucial point of danger,” or crisis. But ji can be combined with many other characters to mean a host of things and must be combined with a totally separate character, hui, to mean “opportunity.” Weiji and jihui share less than the English words “community” and “communism” do. To argue that weiji implies upside opportunity makes as much sense as arguing that people who live in communities are communists.
What Chinese characters most closely resemble are what linguists call “morphemes,” or the smallest unit of meaning. In English, “sun” is a morpheme (the smallest unit meaning what it means), but so are “un-” and “-y,” so that “unsunny” has three meaningful parts conveying “not,” “sun,” and “[adjective suffix].” Most Chinese characters are morphemes too. Since they are also all single syllables in Chinese, some linguists call them “morphosyllabographs.” The Chinese call them hanzi and the Japanese, kanji, but it is easiest just to call them “characters.”
How do they work? The truth is interesting but stubbornly unromantic. Once pictographic, Chinese characters developed over time through what is sometimes called the “rebus principle.” This is the process whereby a picture of an eye stands for the word “I,” the capital letter C stands for “see” and a picture of a bumblebee might represent “be.” Chinese characters similarly went from being pictographs to symbols that stood not for the original referent but for something that sounded like it.
As the Chinese characters evolved, many homophones (words pronounced identically, like “bear” the animal and “bear” the verb) had to be distinguished. The solution was to combine a “phonetic,” which showed how the character was pronounced, with another piece called a “radical,” which gave an idea as to the general domain of meaning. The character 
means “mouth.” But it is part of a common radical, 
, that represents words coming out of a mouth. That radical in turn is used in many characters like 
, “language.”
The same piece of a character can be both a phonetic and a radical: 
stands alone for “tree” (mù in Mandarin). It’s the radical in the character 
, bò, “cypress”—a kind of tree. But it’s the phonetic in 
, mù, “to bathe,” a semantically unrelated word but one pronounced just like mù, “tree.”
Even with a radical (conveying a bit of meaning) and a phonetic, characters can’t simply be decoded by anyone who knows the (huge) inventory of pieces that make up the characters. There are about 200 radicals and perhaps 800 to 1,000 phonetics, and it isn’t always obvious which is which in a given character. At best, there are a few rules, which often break down. Chinese learners must use the phonetics and radicals as mere helpers. These can aid in memorization, and they are useful in recalling hazily learned characters. But they will be unhelpful in recognizing or, especially, producing novel characters. Sometimes the radical is on the left and the phonetic on the right, sometimes vice versa. Sometimes the radical is on top, other times on the bottom. To make things harder still, not all characters have a radical or a phonetic. Some have just one, and some have neither.
Without the radical-and-phonetic “system” (such as it is), learning to read and write Chinese would probably be impossible. Very large dictionaries include about 60,000 or more characters. No one learns that many, of course. Estimates of the number in use average about 7,000. Learning all of them isn’t required for literacy, but at least 4,000 to 5,000 are required for any kind of advanced reading and writing. Knowing just a thousand characters would limit reading and writing ability to simple domains.
This makes it hard to estimate what is needed for “literacy” in Chinese. Italian uses just a couple dozen letters in predictable ways, allowing the Italian peasant who has learned them to read any word in his vocabulary, even if his vocabulary is small. A Chinese with a large vocabulary may not be considered truly “literate” if he can read only a fraction of the words he knows. And since the characters are hard to recall from memory, writing is even harder than reading.
If the thousands of characters of Chinese are daunting, the Japanese system is more difficult still. The government specifies 1,945 characters to be taught in schools, though a few thousand more than that are in common use. But though it uses fewer characters than Chinese, Japanese ladles on the complexity.
The Chinese characters were borrowed by the Japanese from perhaps the fifth century A.D. The Japanese, who had no writing system to that point, began to adapt the Chinese writing system to Japanese, despite the fact that Japanese and Chinese are utterly unrelated and have totally different grammar, vocabulary, and sounds. This adoption happened in several parallel streams, leading to the incredibly intricate modern Japanese system. In some cases, Chinese characters were adopted for native Japanese words. In addition, Chinese words made their way into spoken Japanese, bringing the corresponding characters with them.
For this reason, modern Japanese uses the Chinese characters with a wide variety of different “readings.” In Chinese, most characters have one pronunciation (though many have two, and some have more). Chinese characters in Japanese, however, have pronunciations (called on) derived from borrowed Chinese words, as well as native Japanese ones (kun). Learning a character means learning its meaning as well as its kun and on readings, which are usually not remotely alike.
One typical character, 
, has the on reading ei, and the kun reading kage.
Tōei suru
means “to project an image,” and
kage boshi ga utsuru
means “a person’s shadow is cast.” Someone learning the system, upon seeing the character, will need to know the meaning and readings of neighboring characters to know if he should pronounce it kage or ei.
A simple character has only two readings in Japanese; at the extreme, twenty readings are possible. The character 
, with the base meaning “birth” or “life,” has the readings sei, sho, umu, ikiru, ikasu, hayasu, haeru, shoujiru, shoujizu, nama, ki, and others. On its own, it can mean to give birth, to be born, to be alive, to grow, and the adjectives “raw” and “fresh”; it is also part of compound words including sensei, “teacher,” and gakusei, “student.” Meanwhile, forty-nine different characters can be pronounced “shi.”
To add to this, the Chinese characters coexist in Japanese with two syllabaries, hiragana and katakana, simplified from Chinese characters centuries ago. Hiragana and katakana have no meanings, only pronunciations; unlike alphabetic writing, though, each symbol represents a syllable (shi or ma or yo) rather than a single sound (there is no way to write just the sound y). There are only forty-six characters in each syllabary. In Japanese writing, the Chinese characters are the main parts of words: key nouns, verb stems, and so forth. The hiragana denote word endings, tense, possession, and other crucial bits of grammar. (At a glance, it is the presence of the simpler, gently curved hiragana that distinguishes Japanese from Chinese.) The other syllabary, katakana, is used only to write foreign words and names, to write certain brand names, for onomatopoeia, and for emphasis.
If Chinese and Japanese sound daunting for natives to learn—not to mention for foreigners—they are. Students in both China and Japan spend thousands of hours studying their character sets to be able to read and write. This is, needless to say, time that could be spent on other subjects. Some Western linguists even think that so much rote learning teaches Chinese and Japanese students to value form and process over creativity, though this is controversial.
Do Chinese and Japanese convey advantages, though? Some nationalist linguists in both countries claim that they do. Higher-level words in English are usually built from Latin or Greek roots. “Mesolithic” is from mesos (Greek “middle”) and lithos (“stone”). This is a fancy word pertaining to the Middle Stone Age that English-speakers must learn from a dictionary when they first encounter it. Similar learned words in Chinese and Japanese can be built from two characters that the reader already knows—or so the theory goes.
But William Hannas, an expert in Asian languages for the U.S. government (he speaks Japanese, Mandarin, Shanghainese, Taiwanese, Korean, Vietnamese, and Tibetan—“all well!” his former Ph.D. adviser, Victor Mair, says with amazement), has dismantled the theory. One Japanese study “proved” that the writing system helps Japanese readers learn advanced words, but Hannas, a former academic, showed how it relied on cherry-picking a few erudite English words that had simpler equivalents in Japanese. Meanwhile, some Japanese and Chinese native words can’t be figured out by knowing the individual characters—but Chinese or Japanese readers will try to analyze them anyway, making mistakes.
Other defenses of the system are vaguer: that the effort spent learning them trains the mind generally, for example. This notion enjoyed a vogue period in Japan when that country was booming economically but has lost its shine with Japan’s economic stagnation. The fact is that the enormously difficult Chinese and Japanese writing systems have not withstood careful scrutiny of their supposed advantages. And the disadvantages are more than obvious.
There are alternatives. Both Japanese and Chinese have standard systems for Romanizing their languages. Japanese has no sounds that can’t easily be written in Roman letters using only the existing alphabet, plus the macron for long vowels (like ā, which can also be written aa). The system, called romaji after the Latin alphabet’s home city, is seen all over Japan, and Japanese children learn it early, as an aid to learning the Chinese characters.
Chinese has a Romanization system too: pinyin, established by the Communist government in the 1950s. It is more complicated than Japanese. Mandarin Chinese has four tones, which are as crucial to Chinese meaning as the vowels and consonants of an English word. For example, ma with a rising tone and ma with a falling tone mean “hemp” and “scold,” respectively. In pinyin, the four variants must be written as ma, má, ma, and mà. (Pinyin replaced earlier systems, which is how Mao Tse-tung became Mao Zedong.)
Computers have made typing Chinese characters easier—but they have ironically also made Romanization more tempting than ever. Obviously, keyboards don’t have thousands of keys, so Chinese and Japanese users employ various workarounds. The most common is simply using the Roman letters on the standard keyboard. To type “I am Chinese” in the standard Microsoft Windows Chinese support system, you type wo in Roman letters, then pick the right wo character from the offering, and so on through each of the characters wo shi zhong guo ren. Modern computers are clever enough to guess at the most commonly sought characters, so users can also type woshizhongguoren in one go, and Windows will offer the correct characters. This won’t work for more unusual sentences and characters, however. Japanese works much the same way.
But if Chinese and Japanese both use Roman letters to get to Chinese characters, why use the Chinese characters at all? The pressures to make reading easier go back a long way and often come from nationalists with impeccable credentials. But Romanization has failed again and again. And it has failed thanks to political concerns of identity and cultural heritage, not the purely linguistic concerns that are sometimes trotted out.
Japan’s first notable advocate of Romanization, Nanbu Yoshikazu, proposed accepting the alphabet as far back as 1871—just decades after the American Commodore Matthew Perry steamed into Tokyo’s harbor and forced Japan open to the world. The Japanese leadership realized that the country was well behind the West in scientific, technological, and military advancement. The push for Romanization thus came when many other Western technologies were being eagerly adopted, during Japan’s spectacular drive to modernity.
Romanization’s fans competed with advocates of another radical simplification: those who, first organized in 1866, favored adopting the kana syllabaries for Japanese writing. They are simple to learn and have the political advantage of being locally derived. But the kana-only movement fell out over arguments about how Japanese-derived versus Chinese-derived words should be written.
Movements in favor of kana or Romanization ground to a halt in the 1930s, as radical nationalism took hold of a Japan that was fighting wars on several fronts. (Japan even stopped teaching English as a foreign language in schools.) In the postwar period, however, Japan was defeated and occupied by the Americans and was rapidly remade in the image of a Western democracy. An acclaimed novelist, Shiga Naoya, even proposed that Japan adopt French.
So the advocates of Romanization and kana had another chance. The Americans were divided on what to do about the writing system. Some supported a switch to kana or Roman letters, on the grounds that it would symbolize a break with the past, render wartime propaganda harder to read, improve education, and make censorship easier. But the Americans ultimately decided against imposing a reform, and once again the opportunity slipped. Japanese proponents fell out among themselves—advocates of one romaji standard (called Hepburn) even denounced other romaji supporters as wartime collaborators. While the romaji camps squabbled, a new national language body appeased critics of the writing system by abridging the kanji set to 1,850 characters (later expanded to 1,945).
Japan still has proponents of kana and Romanization, and it seems as though the increasing computerization of everything will gradually turn up the pressure. But the “inevitability” of Romanization or adoption of kana only has been confidently predicted since just after the war. No one can say that it won’t happen or that it will, but the forces of linguistic conservatism are powerful and determined.
Where does the resistance come from? Modern Japan isn’t particularly xenophobic regarding foreign words. The language has a big, and old, layer of high-level vocabulary from Chinese. And in the modern era, Japanese have happily borrowed words from aisu kariimu and kohii (ice cream and coffee) to boifurendo and garufurendo (boyfriend and girlfriend). Such “Japlish” is the source of many jokes on both sides of the Pacific, but it is a simple, mostly uncontroversial fact of modern Japanese.
Despite this openness to new words, the characters persist. Why? Besides nationalist motives, psychological factors favor inertia. All the bureaucrats and lawmakers in a position to reform the writing system have already learned it—indeed, their presence among the elite required it. People are strongly attached to things they’ve given a great deal of time and painful effort to learn. This kind of thinking is well known to everyone from soldiers looking back on the screaming and yelling of basic training to a senior fraternity brother hazing his pledges: “I went through this, and believe me, it was no party. But it made me who I am today, and it bonded me to this community. It didn’t kill me, and it won’t kill you either.” With this argument in place, secondary ones for Chinese characters, such as “It trains the mind for higher thinking” and “It is good for instilling discipline,” settle in unchallenged. If the infinitely more modest German spelling reform is any guide, any change in Japan would have to be demanded from below; to introduce it from above would be to spark ferocious opposition. Japanese adoption of the Roman alphabet is a long way off, if it is ever to come at all.
In contrast to Japan, China, a one-party dictatorship, could introduce pinyin in place of Chinese characters any day it liked. Yet China is unlikely to Romanize soon for its own reasons, some of which do overlap with Japan’s. The Chinese believe to some extent that the characters are the language and that understanding them means understanding Chinese. Western linguists wouldn’t agree—they tend to see writing systems as artificial structures built on top of the real language, which is the organic product of people’s brains and mouths. But changing Chinese minds on this is unlikely.
A more important reason China will not adopt pinyin is that it logistically can’t, not without exploding a notion to which the Chinese are inalterably attached: that there is one Chinese language, of which different varieties, such as Mandarin, Shanghainese, and Cantonese, are “dialects.” This is simply untrue. The major different classes called “Chinese” are, without any serious debate among linguists, mutually unintelligible different languages, at least as different as the major Romance languages of Europe are from one another. Saying that Mandarin and Hakka are dialects of Chinese is like saying that French and Italian are dialects of Latin.
Max Weinreich, a Yiddish linguist, noted that in the popular imagination, “En shprakh is en dialekt mit en army en flot”: “A language is a dialect with an army and navy.” How different two varieties should be to be called “languages” rather than “dialects” is not fixed, because the main criterion—whether two speakers can understand each other—isn’t an either- or phenomenon but a continuum. Speakers of Danish can understand Swedish, and Swedes can (with somewhat less success) understand Danish. To some linguists, that is a case for calling these varieties one language (“Scandinavian”) with several dialects. But for most people, the real criterion is political: since both countries have armies and navies, we consider Swedish and Danish different languages.
Any objective look would render the opposite judgment of China. It has a tremendous army and a growing navy. But, simply put, a Mandarin-speaker from Beijing and a Cantonese speaker from Hong Kong, if they have not learned the other language, cannot understand each other. Cantonese has six tones to Mandarin’s four; it can end syllables with the consonants p, t, m, and k, while Mandarin has just two syllable-ending consonants, n and ng. The actor Chow Yun-fat, a Cantonese-speaker, had to learn Mandarin for the movie Crouching Tiger, Hidden Dragon, and he was ribbed by some Mandarin-speaking viewers for his accent in Mandarin. (The movie is also available dubbed into Cantonese.) The same levels of difference, broadly, exist among the other major Chinese languages, including Wu, Hakka, Min, Gan, and Xiang.
The differences among the different “Chineses” are so well known that Mandarin-speakers have a rhyming saying:
Tian bu pa, di bu pa, zhi pa Guangdong ren shuo Putonghua. (“I fear neither Heaven nor Earth, I only fear Cantonese speakers trying to speak Mandarin.”)
The Cantonese have their own version:
Tin mh geng, deih mh geng, ji geng bak fong yahn gong gwong dung wah mh jehng. (“I fear neither Heaven nor Earth, I only fear Mandarin speakers speaking Cantonese so inaccurately.”)
Note how different they look.
But the Chinese nonetheless insist on speaking of one Chinese language with different dialects. Why? It’s not a government plot to deny China’s linguistic divisions and keep the country together. Han Chinese (the majority of Chinese, but not speakers of smaller languages such as Mongol, Tibetan, or Uighur) do share a strong identity, bolstered by China’s millennia-long history of political independence. Mostly they really do believe that they speak the same language.
Besides the strong sense of national identity, one reason so many Chinese are in denial is that the writing system hides the differences. Written Chinese, based on Mandarin, is a standard that speakers of all the Chinese languages are taught in school, even though they pronounce the individual characters completely differently in their own languages. The fact that a Mandarin- and a Cantonese-speaker read the same characters tricks them into thinking they speak the same language, when really they are both just reading Mandarin.
Romanization would change all that. No longer would written Chinese unify the country, because pinyin is a Mandarin-only system. (The Cantonese equivalent is pingyam, used in the “Heaven and Earth” saying above.) Writing Mandarin in pinyin would expose to all non-Mandarin-speakers the fact that they are looking at a foreign language. This is a headache that China’s authorities, already fearful of non-Han separatist movements in Tibet and the Muslim, Uighur-speaking region of Xinjiang, can do without. Stoking nationalism among the Han majority is increasingly tempting for the Chinese leadership, to head off challenges to its authority in the face of corruption, inequality, environmental degradation, and the like.
Surprisingly, Romanization nearly succeeded in China. One 1934 experiment with an early Romanization system “proved beyond a doubt that ignorant farmers and laborers need only a hundred hours” to learn alphabetic Chinese, according to proponents. But the nationalist Kuomintang government lost interest.
After the Communists’ 1949 victory in the Civil War, the political will for reform came in a new form. Mao Zedong, now the unquestioned leader of China, had said earlier that alphabetization was inevitable. After taking power he seemed to change his mind, saying in 1951 that “The writing system must be reformed, and it should take the phonetic direction common to the languages of the world; it should be national in form; the alphabet and scheme should be designed according to the existing Chinese characters.” (Emphasis added.) A more concrete statement could not be possible: Chinese must be phoneticized, though the system should be Chinese-based.
It looked as though some radical simplification was inevitable, and a national commission was set up to study the question. After mulling over Chinese-derived options and one using the Cyrillic alphabet, the commission recommended pinyin. (It considered, perhaps, that a made-in-China system, even Roman-based, would be “national in form.”) But Mao’s interest had moved on. His premier, the powerful Zhou Enlai, elaborated the new policy: that Chinese characters would be simplified, not abandoned, and pinyin would be used only for teaching purposes. Meanwhile, a standard language based on Mandarin was to be promoted. One famed poet and scholar, Chen Mengjia, dared to oppose the abandonment of traditional Chinese characters during the brief “Hundred Flowers” movement of intellectual openness in 1957. After the “Hundred Flowers” experiment ended, he was labeled a “rightist” and exiled to three years of farm labor in Henan. He killed himself during the Cultural Revolution.
The real reform was now the simplification of the characters. Many of the hanzi were stripped of brushstrokes, making them easier to write and learn. The simplified character set is now standard in the People’s Republic, though Hong Kong still uses the traditional characters, as does Taiwan and as do most overseas Chinese. The reform was ultimately conservative, and full Romanization was dead.
China still may introduce pinyin, but only if an important precondition is met. The government strongly promotes putonghua, the official “common tongue” based on the Mandarin of Beijing. Signs in Shanghai read “Speak Mandarin—be a modern person.” Toilets in Beijing’s Capital Normal University have signs above them reading “Your excellent Mandarin provides convenience to everyone around you.” (“Provides convenience” puns on a Chinese euphemism for going to the toilet.) And the government has even objected to the dubbing of the American children’s cartoon Tom and Jerry into Shanghainese for children: in 2004 the state broadcasting authority ordered an end to the show, though it later relented.
As China’s economy booms, the spread of Mandarin may accelerate: more people will move around the country and need a common language to communicate. Education is in Mandarin, and the spread of computing, broadcasting, and other technologies may put stronger wind into Mandarin’s sails. Those same forces may also militate in favor of Romanization. The more people use computers—using Roman letters to call up a menu of characters they can choose from—the more they forget how to write them from scratch and think in pinyin all the while.
It would be ironic if China switched to pinyin before Japan adopted romaji or kana. The characters, after all, come from China, not Japan. It’s not inconceivable that China may make the switch for practical, economic, and technological reasons someday, and it has a dictatorship to force it through. Meanwhile, Japan hangs on to Chinese characters for the sake of its national identity, a position encouraged by nationalist politicians. Neither China nor Japan looks as if it will make any changes soon. Politics, as ever, trumps clear thinking about language.
Several themes emerge from the different stories of language planning. Usually the planners make some form of argument based on the state of the language itself, whether its integrity, expressiveness, or modernity is at stake. The linguist George Thomas has collected the metaphors that language nationalists use, and each involves something reminiscent of physical wholeness or purity, often biological: language is an ore that needs refining, a grain that needs milling, a genealogical organism that needs pure breeding, a tree that needs pruning, a body that needs surgery, and so forth. Foreign elements—Arabic in Turkish, Anglicisms in French, Latin and Greek words in William Barnes’s view of English—these are somehow poisonous, contaminating, corroding. And change is usually equated with decay or rot, rarely considered neutrally as mere change.
But the real motivation behind most language planning is usually political. Language reforms have winners and losers, and so appeasing the “losers” (including those who have already mastered an old system and will have to learn a new one) makes it tempting to appeal to linguistic factors, to convince everyone that the planners know best about how the language should work.
If forcing through change by fiat is usually a bad idea, sometimes it makes some sense. Newly independent countries establishing their identities may have the most sympathetic arguments for planning their language: coining and codifying native elements, choosing an appropriate writing system, trying to keep foreign elements out. But even when we are sensitive to those political motives, they should be made clear. Language planning is almost always about identity creation, the gathering of an in-group and the exclusion of an out-group, the process called “nation building.” Nation building happens through the creation of artificial symbols such as a flag, a national anthem, or national myths. There is nothing wrong with these and nothing necessarily wrong with nation building through language, either. But language planning will be not only more honest but also more successful when its nationbuilding motives are front and center, rather than dressed up in bogus arguments about language itself.
* Early in its history, only the director had an armchair, but one ailing cardinal demanded something more comfortable than his ordinary chair, and Louis XIV ordered armchairs for all forty.