by David Bellos
NINETEEN
Global Flows: Center and Periphery in the Translation of Books
The Harvill Press was founded in London in 1948 to publish literary works of high quality from other languages, initially from Eastern Europe. By the time of its fiftieth anniversary, it was proud to announce that it had published English translations of works originally written in forty-three different tongues. In Paris, Ismail Kadare’s French publisher regularly informs readers of jacket blurbs that the Albanian novelist’s works have been translated into “more than forty languages.” Are these the same ones? With only a few exceptions, the answer is yes. Nowadays there are only about fifty languages between which imports and exports of translated books occur with any regularity.1 That represents a minute fragment of global linguistic diversity, yet it covers a large proportion of the population of the world. That’s because translation languages are, by necessity, vehicular ones, read (if not also spoken) by vastly more people than those who have them as their native tongues.
But what of the rest? All or part of the Jewish and Christian scriptures exist in nearly twenty-five hundred languages. Some of these also have translations of legal and administrative texts, and a few possess news or gossip magazines and a small quantity of popular fiction. But what’s obvious from these numbers is that more than half the world’s languages probably receive no translations at all, and all but fifty or so export almost nothing, either. Print translation happens only in special places. That’s not to minimize its importance but to point to the peculiarly asymmetric relations that have always obtained among the different forms of speech on this planet.
UNESCO, the cultural arm of the United Nations Organization, has attempted since its founding to keep track of the global flow of translations through the Index Translationum, which is now available as a searchable database on the Web. It can be used as a rough measure of the huge imbalances in translation in the world today.
Chinese is spoken by about a quarter of the world’s population, and in a well-balanced and reciprocating global society you would expect it to be the receiver of about a quarter of the translations done in the world. The truth is nothing like that at all.
Taking seven world languages of different kinds for the ten years from 2000 to 2009, Chinese is the receiving language of just over 5 percent of all the translations done in all directions among these tongues—barely more than Swedish, whose speakers number less than 1 percent of the speakers of Chinese. But the picture in the reverse direction is even worse. Only 863 books were translated from Chinese into Swedish, Hindi, Arabic, French, German, and English combined, whereas more than twice that number of books written in Swedish were published in Chinese, Hindi, Arabic, French, German, and English combined.
Books translated between seven languages, 2000–2009 inclusive
Nearly 80 percent of all translations done in all directions between these seven languages over a decade—104,000 out of 133,000—are translations from English. Conversely, barely more than 8 percent of all translations done in the same set are translations into English—whereas French and German between them are the receiving languages of 78 percent of all translations.
The asymmetry is striking and, in some senses, quite alarming. Granted, published books do not provide the only channel of intercultural communication; in addition, the data stored by UNESCO may not be complete, and its search engine may have its own quirks. But the overall picture—which is confirmed by what any traveler can see in any airport bookstore in the world today—must be broadly true. Translations from English are all over the place; translations into English are as rare as hen’s teeth.
It is neither accurate nor even interesting to pin the responsibility for our lopsided translation world on the Almighty Dollar alone.2 Translation flows measured in this way also fail to give a particularly convincing map of military power in our own or recent centuries. The initial spread of British English around the globe was certainly the fruit of colonial expansion—but the huge scope and increasing pace of its dominance followed the dismantling of empire that began in 1947. The imperial hypothesis fails to explain why French, Spanish, Portuguese, and Dutch, the languages of equally far-flung and densely populated empires between the sixteenth and twentieth centuries, are nowhere near the top of today’s global translation tree. For every work in Spanish translated into English in the first decade of the twenty-first century, fifteen were translated from English into Spanish. Yet there are almost as many native speakers of Spanish (around 350 million) as of English (400 million) on the planet today.
Translation DOWN often takes place for mostly practical reasons from the language of dominance to the languages used by peoples living within the field of domination. In the Hapsburg Empire, for example, laws, regulations, official announcements, and daily news were translated from German, the language of the court and imperial administration, into the seventeen official languages of that ramshackle state. But books didn’t follow behind to any great extent. No lively culture of literary translation sprang up into Slovene, Slovak, Serbo-Croatian, Ruthenian, Czech, and so forth. That’s because there was a much more straightforward way of becoming a cultivated citizen of the Austro-Hungarian Empire: by learning German. In like manner, many serious books in English about history, science, literature, and the arts cannot be commercially translated into Swedish, Danish, Norwegian, or Dutch because interested readers in these communities read them in English already. Economic, military, and cultural domination obviously affects translation flows, but typically not in direct or straightforward ways. A truly dominant language that has a great army and a well-filled treasury behind it—say, Latin throughout the period of the Romans’ domination of Europe and the Mediterranean—is the one tongue from which you do not ever need to translate. People just learn it, because without it their prospects are blocked. English does not dominate the world in the way that Latin did, because it is massively translated into vernaculars. Translation is the opposite of empire.
When speakers of Spanish, Portuguese, and English spread into the New World between the fifteenth century and the eighteenth, they did not initiate translations into any of the languages of the native inhabitants of the Americas. They created empires. But when Soviet Russia consolidated its hold on the many peoples of Siberia, the Caucasus, and Central Asia in the 1920s, it did so with firm political convictions that were explicitly anti-imperial. To demonstrate and implement those self-directed beliefs, the Soviet Union launched a huge program of translation from and into the indigenous languages of what were called “the nationalities”—Kazakh, Turkmen, Georgian, Azerbaijani, and so on. There was a good deal of hypocrisy in the Soviet stance, but the important thing to realize is that only translation could serve as a public alibi for what was in most other ways a classic instance of imperial expansion. Russian literary classics were made available in Kazakh, Ingush, Daghestani, and so on, but translation UP was a necessary complement. Two-way trade was needed to demonstrate the truly anti-imperial nature of the Union.
The problem Soviet language planners faced was that it takes a long time to establish functioning translation relations between two languages. Schools have to be established to educate a generation of bilinguals, who then have to develop their own translation tools and conventions. It can’t be done overnight, however great the need. But Soviet Russia was a revolutionary enterprise in a great hurry to usher in a new world. That’s why it began to cheat. Native poets in the main non-Russian languages were hard to find, and it was even harder to find Russian poets able to translate them. The Soviet solution was to invent them. Dzhambul Dzhabayev is the most famous example of Soviet pseudo-translation, partly because the deception was so long drawn out. A well-known Kazakh folksinger at the time of the Revolution, Dzhabayev was compelled to lend his name to patriotic poems written in Russian by a whole factory of hacks, who presented them as having been translated from Kazakh. Dzhabayev was translated into many other languages—from Russian, in fact, but always o
fficially from Kazakh. Because “Kazakhstan’s national poet” lived to the age of ninety-nine, the Moscow song factory was able maintain the illusion for many decades.3
However, not all empires treat the language of the conquerors as the conquering language. In many known instances, a culture of translation sprang up that gave prestige and authority to the language of the conquered. When the Akkadians overran Sumer around 2250 B.C.E., they did not sweep away the much older culture and language of their new subjects. They adopted Sumerian script—the wedge-shaped letters made by incising wet clay with the sharpened tip of a reed—and treated the (linguistically unrelated) Sumerian language as a cultural asset. Laws and legends, rules and chronicles were translated from Sumerian into Akkadian, and knowledge of Sumerian became the mark of an educated man throughout the many centuries of Akkadian and Assyrian civilization. Despite having ceased to be associated with any political, military, or economic authority, and progressively disconnected from any identifiable ethnicity as well, Sumerian went on being used as a sacred, ceremonial, literary, and scientific language in Mesopotamia until the first century C.E.—giving it a life span as a source of translations DOWN of approximately three thousand years. English has a long way to go to equal that.
Between the fifth and third centuries B.C.E., Greek-speaking seafarers spread their language far and wide in small pockets stretched out along coastlines from Marseilles to Odessa, and Alexander the Great took it overland as far as Egypt and Afghanistan. But the role of Greek as a source language for translation has nothing to do with Macedonian military might. Even before it conquered and occupied the Greek peninsula at the start of the second century B.C.E., Rome became eager to take possession of Greece’s culture and thought. In due course the language of the conquered was recognized as a repository of cultural and intellectual prestige. The learning of Greek became the main content of a proper education in ancient Rome and translation into Latin the main skill associated with high rank.
The stories of Sumerian and Greek oblige us to be more than doubtful about economic, military, and political explanations of the translation map of the world today. In one sense, of course, these ancient examples of the languages of subject peoples being invested with cultural prestige through translation are exceptional, because there don’t seem to be any good examples in medieval or modern times. When the Normans conquered England, they did not adopt Anglo-Saxon as their language of culture—they carried on using French and let the common folk speak a Franco-Saxon mishmash that eventually turned into English. When the French seized the throne of Sweden in the Napoleonic Wars, they didn’t start translating from Swedish. In fact, the new Swedish royal family and its court carried on speaking French for more than a hundred years—and their descendants still maintain a palace in Nice.
But we could just as easily take the opposite view. The general history of translation in the European sphere in the last few hundred years may itself be an exception to a longer-running norm. And even within that domain there are examples of languages of culture and translation being retained against political or military logic. Latin remained dominant, both as the source language of translations DOWN and as the target language of many vernacular texts principally for the purpose of retranslation into other vernaculars that had no translation relations between them for more than a thousand years after the fall of Rome. Jews have continued to use Hebrew for more than three millennia despite having had a thousand practical reasons for dropping it like a hot brick.
What makes a language culturally dominant, today as at all other times, has no relationship to the number of centurions, tanks, or missiles it has to back it up or the quantity of gold in its treasury. A culturally dominant language is one that maintains significant volumes of translation activity between itself and a significant number of languages that have smaller bilateral translation relations between them. The dominance of Latin in fourteenth-century Europe, for example, is not just exemplified by the way in which The Travels of Marco Polo was spread; it was created and maintained by just that kind of use, as the interlanguage that enabled the same (or a similar) text to be made available in languages such as Czech and Gaelic, between which bilateral translation skills were practically nonexistent. It had nothing to do with the economic or military power of “native Latin speakers,” of whom there were precisely none.
The position of English as source and target for the vast bulk of translation done in the world to date is brought out by ranking the most popular source languages for translations into any selection of languages you care to choose. This table shows the main source languages for translations of books into thirteen widely spoken languages since UNESCO first started keeping records:
Top four source languages
What’s clear is that English, French, and German dominate translation worldwide. Russian has a perhaps surprising role in fourth place, but the eight others that appear on this ranking—Spanish and Italian three times each; Sanskrit twice; Japanese, Finnish, Bengali, Arabic, and Malayalam only once each; and Chinese not at all—are peripheral to the global business of translating books.
The raw numbers of translated books on which this ranking is based produces an even more startling picture of the pyramidal structure of global translation today. Of the nearly 1 million translations used to compile the ranking, more than 650,000 are translations from English, and a further 10 percent of the total number consists of translations into English. English is the medium as source or target of 75.12 percent of all translation acts.
What these figures also show is that around 42 percent of all the translations recorded in the UNESCO database between the thirteen languages listed above have taken place in closed circuit between just three of them—English, French, and German. This is not an ineluctable consequence of the fact that of the million books we are dealing with, more than 47 percent were published in one of those three languages, too. Culture is not the prerogative of any part or place in the world, but book culture—and, within it, the culture of translation—is heavily concentrated in Britain, the United States, France, and Germany.
As a result, at any truly representative gathering of translators from across the globe, between 70 and 90 percent of delegates must be L1 speakers of a language other than English. To put it another way: if you would really like your children to earn their living as translators, you’ll give them a much better chance if you don’t raise them in Britain or America. This also explains why translation is much less easy to see and understand when you are based in the English-speaking world. You don’t meet many translators in the normal course of life in London, Sydney, or Cork—but they’re all over the place in Geneva and Berlin.
The flow of translations has always had a hierarchical structure: the present situation reproduces a pattern that can be observed many times in the historical past. Translation typically takes place not between languages felt by their speakers to be on an equal footing but between those that in some respect have a vertical relationship between them. Laws, commands, instructions, and treaties are translated DOWN—from Sumerian, Greek, and Latin in ancient times; from German in the Hapsburg Empire; from Ottoman Turkish in the long period of Ottoman sway in the Mediterranean basin—into vernaculars spoken by people who need to grasp what the rules and agreements that affect them are. Novels, plays, philosophical and mathematical treatises, and religious texts may accompany them, but not always. Out of these kinds of situations the world over have grown ideas among the speakers of culturally dominant tongues that their language is inherently superior and the only true vehicle of thought. In the Muslim world, for example, there was little doubt in past centuries about which language was top:
The perfect language is the language of the Arabs and the perfection of eloquence is the speech of the Arabs, all others being deficient. The Arabic language among languages is like the human form among beasts. Just as humanity emerged as the final form among animals, so is the Arabic language the final perfection of human language a
nd of the art of writing, after which there is no more.4
Seventeenth-century French grammarians made much the same assertion about French, and similar expressions of confidence in the superiority of Greek, Persian, Latin, Chinese, and who knows how many others among the world’s temporarily dominant tongues could easily be lined up.
Obviously, there are no rational grounds for such kinds of linguistic preference: all languages can be made to serve whatever ends their speakers wish to achieve. But the feeling that a difficult foreign text makes real and proper sense only when it’s been put into the language we prefer to use for thinking hard thoughts can easily ambush an otherwise sensible mind. Years ago I sat in a library in Konstanz trying to make sense of Hegel by reading him very slowly in German, with a pencil in my hand. It was hard going, and I never really got the hang of it. I sneaked a look at what the German student in the next carrel was reading. It was Hegel, too—but in English translation! Well, I thought to myself with relief, if even native speakers use the English translation as a guide to Hegel’s thought … Such experiences can easily lead you into a barely conscious, self-comforting persuasion that your language alone is the one in which real meaning is to be found. But however great the service that a clarifying, explanatory translation of a foreign text may provide, we should always resist the false conclusion that the target language—whatever language it is—is “better” at expressing this or that kind of thought.
Despite their numerical insignificance, translators into English play an important role in the international trade in books. Because it is the most translated language in the world, it is far easier to get a book into any other language if it exists in English already—whatever language its original language was. But English is by no means the only “pivot tongue” in the world.