by David Bellos
“My word” is simply my saying of it. In this usage, word means not a unit of speaking but the act of speech itself. Similarly, when I call a friend “a man of his word,” I make no reference to his using some particular lexical item. I mean to say that whatever my friend undertakes to do by an act of speech is to be taken seriously, because it was he who said it.
In French, the distinction between “word as act” and “word as unit” is made clearer by the general use of parole for the first and mot for the second. In German, too, there is a trace of the fundamental divide in the meaning of the word word in the two different plurals of Wort—Worte for acts of speech and Wörter for entries in a Wörterbuch.
There is of course a real connection between these two divergent associations of word. Both of them name the smallest handy unit of speech. It’s just that since the invention of alphabetic script we’ve grown completely accustomed to thinking that the true form of what we say is the way it looks when written down. “Scriptism,” as Roy Harris called the illusion that a language consists of things called words, has served us well for a few thousand years, but it has a downside as well. It makes it harder to understand what translation does.
The uses in many Western languages of words meaning “word” to refer to acts of speech are perseverating traces of primary orality. The status of any utterance in a mental world without script derives mainly from the identity of the speaker, much less from the “meanings” of the “words” that are spoken. The concepts in scare quotes are probably not even thinkable without writing. The indeterminacy of the flow of speech and the dependence of meaning on the human context in an oral culture are pinpointed with affection and insight by Tolstoy in his portrait of the illiterate peasant-philosopher Platon Karatayev, in War and Peace:
Platon could never recall what he had said a moment before, just as he could never tell Pierre [Bezukhov] the words of his favourite song … He did not understand and could not grasp the meaning of words apart from their context … His words and actions flowed from him as smoothly, as inevitably and as spontaneously as fragrance exhales from a flower. He could not understand the value or significance of any word or deed taken separately.4
“Translating” in this kind of cultural circumstance calls for a special kind of trust. If the force of an utterance is intimately linked to the identity of the speaker, then it can’t be conveyed by any other speaker. That fundamental rule has to be suspended for oral translation to come into existence, since it requires the listener to take the words of the translator as if they had been uttered by the speaker of a foreign tongue. Oral translation in a world without writing creates and relies on a fiction—perhaps the earliest fictional invention of all. The first great leap forward in the history of translation must have been when some two communities found a way of agreeing that the speech of the translator was to be taken as having the same force as the immediately prior speech of the principal.
It’s not hard to account for the existence of bilinguals in early human societies: taking brides from different communities and taking slaves from vanquished enemies are ancient practices, and both of them can easily result in people who understand two different languages. But there’s a great difference between bilingualism and translation. For the latter to exist, huge intellectual and emotional obstacles to taking the word of another for the word of the source have to be overcome. They can be overcome only by a shared willingness to enter a realm in which meaning cannot be completely guaranteed. That kind of trust is perhaps the foundation of all culture.
But that trust is never granted without reservation. To conduct negotiation or trade between two communities speaking mutually incomprehensible tongues, the principal relies on the translator and is in his power, just as the translator serves one master only and is entirely in his power. The situation is guaranteed to create anxiety, suspicion, and mistrust.
The fear of imperfect or deceptive performance by an oral translator affects the translation protocols for private meetings between world leaders today. Each side brings along his or her own oral translator. When the British prime minister talks to the French president in confidential, face-to-face encounters, the person employed by Her Majesty’s Government speaks in French on behalf of the prime minister, and the French translator similarly speaks back the French president’s words in English. Such two-handed, one-way speech translation, out of the mother tongue and into the foreign, is never seen in public.5 These arrangements hark back directly to the issue of trust in oral translation. Translators are no longer slaves, but states still have greater recourse against employees who have signed confidentiality agreements than against a translator hired by the other side.
This costly double dose of oral interpreting is rare, but not solely because it is expensive. Outside of private head-of-state encounters, almost all speech by politicians, diplomats, and public figures begins and ends its life on the page. Delegates at the United Nations General Assembly and Security Council, for example, read from prepared texts, and often the interpreters translating the speech simultaneously into (any five out of) English, French, Spanish, Russian, Chinese, and Arabic have the original text in front of them. All six language versions are recorded on tape and these recordings are used by the UN Documentation Division to produce the “Verbatim,” the official written record of what was said. This allows translation errors to be trapped and corrected, but, more significantly, it allows delegates to correct what they actually said. The “Verbatim,” the final official repository of UN proceedings, is not actually verbatim at all—it’s a rewritten version of a written text that passed through an untrusted oral stage in the interim. In large areas of national and international affairs, speech has now become a secondary medium, a by-product of writing. But this is a very recent state of affairs. Our thoughts and feelings about language and translation, together with many of the things we say about it, have much older sources.
Between the fifteenth and early twentieth centuries, the Ottoman Empire held in its not always steady sway mostly illiterate populations speaking a great number of different languages. Throughout these five centuries, the administration of this vast and elaborate state was carried out in Ottoman Turkish—a partly artificial hybrid of Turkish, Persian, and Arabic vocabulary held together by Turkish grammar, with some Persian syntax added on, written in an adapted Arabic script that was not particularly well suited to it. It was the official language of the court at Istanbul, but outside the circle of imperial grandees and civil servants Ottoman Turkish did not have many speakers. Its written form was of course used for the state’s labyrinthine archives—by some accounts, the Ottomans even kept records of people’s dreams.6 However, one characteristic of Ottoman society was a paranoid suspicion of forgery, and as a result writing was not used for all purposes of state. Strong residues of orality—of a trust in personal speech over the impersonal technology of writing—affected the management of public affairs and, most especially, its use of translators.
Ottoman society, like those of the Greeks and the Romans, made slaves of a significant proportion of its subjects, and it recruited translators from among the young boys sent back from the provinces to Istanbul as obligatory payment for the protection the empire provided. Most of these enforced bilinguals served the internal needs of the empire, since they spoke one of its regional languages and received an education in Ottoman Turkish. Its external translation needs for trade, war, and diplomacy were served mostly by other means.
The Ottomans were Muslims and could therefore communicate with many of the peoples on the southern and eastern borders of the empire in Arabic, which was either a native or a vehicular tongue over a wide area. But contact with Western Europe was not so easy. In no region of the empire were any of the Western languages taught. Initially, therefore, the training of cadres who could handle relations with the West was farmed out to the Republic of Venice, which had long-standing ties with many parts of the Mediterranean that had fallen into Ottoman hands.
From t
he late fifteenth century on, Venice dispatched plenipotentiaries on two-year postings to Istanbul to run the bailo, which was something like a translator’s school. It recruited adolescent apprentices called “language boys”—giovani di lingua, a translation of the Turkish dil olan—across the Venetian and Ottoman territories and turned them into loyal, Italian-speaking Venetian subjects capable of talking to the Turks. Many of the recruits came from the Greek-speaking Roman Catholic community that had settled in a quarter of Istanbul called Pera, or Phanari in Greek, and Phanariots eventually became a hereditary “translation caste” within the stratified world of Ottoman society. By the early seventeenth century, the whole business of translation at the highest levels of the Ottoman state was in the hands of closely linked families of Phanariots, whose status was partly protected by the fact that many of them also held Venetian citizenship by inheritance. But they did not translate very much into or out of Greek: they were trained to translate Ottoman Turkish into Italian, and sometimes Arabic as well. They became richly rewarded grandees. Based in Istanbul, they sent their sons to Italian universities before bringing them back to continue the family trade.7
Diplomacy, spying, and administrative intrigue were all part of the job done by these Ottoman translators, called tercüman. This Turkish term has come into English as dragoman, but in only slightly altered forms it can be found in dozens of other languages that had contact with the Turks. Azerbaijani trcüm∂çi, Amharic ästärgwami, Dari tarjomân, Persian motarjem, Uzbek tarzhimon (), Arabic mutarjim, Moroccan Arabic trzman, and Hebrew metargem ( ) are all sound translations of tercüman. But whether written as dragoman or as tercüman, the Ottoman word for “translator” is not a Turkish word at all. It is first found in a language spoken in Mesopotamia in the third millennium B.C.E., as a translation of the even more ancient Sumerian word eme-bal. Akkadian targumannu thus has a descendant by way of Turkish tercüman in an admittedly obsolete but still extant word of English—probably the only word with a stable meaning whose history can be traced in writing over a period of five thousand years.8 The spread of one of the most widely used root words for “translator” from one of the cradles of writing in ancient Mesopotamia can hardly be bettered as evidence for the immensely greater antiquity of the practice of translation itself.
Top Ottoman dragomans became the equals of ambassadors. The first to be granted the title “Grand Dragoman” by the sultan was appointed in 1661 under the reign of Küprölü Ahmed Pasha—the famous Albanian grand vizier Quprili, whose many adventures are turned into fiction in the novels of Ismail Kadare. 9 A later grand dragoman, Aléxandros Mavrokordátos, founded a dynasty that eventually acquired princely status. His direct descendants became the royal family of Romania.
Because they were diplomats and negotiators using speech and not writing for the most delicate matters, dragomans dealt with their written tasks along lines more characteristic of oral translation. Dragomans altered the pasha’s language to put it in a form best suited to performing the act that the principal intended. They did this in order to remain faithful to the sultan—for disloyalty was punishable by death, if not worse. Far from being “free,” the dragomans’ reformulation of the words of the source expressed subservience to their principal’s intention. Despite appearances to the contrary—substantial amounts of contraction, expansion, and recasting—dragomans stuck rigidly to their brief, which was not to translate the sultan’s words but his word.
For example, when Sultan Murad II granted permission for English merchants to trade in the Ottoman lands, his original letter in Turkish refers to Queen Elizabeth as “having demonstrated her subservience and devotion and declared her servitude and attachment” to the sultan. For onward communication to the English court the letter was translated by the grand dragoman into Italian, which was still the international language of the Ottoman Empire.10 In Italian, however, the letter doesn’t say nearly as much: it expresses the elaborate Turkish formula economically as sincera amicizia.11
Is this a “free” translation or an “unfaithful” one? I don’t think either term is appropriate. The dragoman’s occlusion of the words for “subservience” and “servitude” is not an expression of his freedom but of the political and administrative constraints of his own position. He knows that his own master will never regard the queen of England as a monarch of equal power; and as a seasoned diplomat he also knows that Elizabeth I cannot possibly accede to the expression of her “servitude” to the sultan, even in a conventional flourish.
Western embassies in Istanbul did not use the official court interpreters in the service of the Ottoman court, who were bound to be loyal to their sovereign. They employed less eminent, and mostly non-Muslim, bilinguals to be found in Istanbul. As they became less and less familiar with oral culture over the several centuries of Ottoman rule, Western diplomats increasingly described their Levantine intermediaries as unreliable and untrustworthy folk. In the first place, they grumbled, at least half of what they wrote and pretended to be translation “from English” was pure invention, in the following style:
Having bowed my head in submission and rubbed my slavish brow in utter humility and complete abjection and supplication to the beneficent dust beneath the feet of my mighty, gracious, condescending, compassionate, merciful benefactor, my most generous and open-handed master, I pray that the peerless and almighty provider of remedies may bless your lofty person, the extremity of benefit, protect my benefactor from the vicissitudes and afflictions of time, prolong the days of his life, his might and his splendor …
Also, every scrap of information they gleaned from translating for a foreign embassy was put up for sale. As one English ambassador put it, since these dragomans “with large families live upon a small salary and are used to Oriental luxury, the temptation of money from others is with difficulty withstood by them.”12
It’s easy to see why such dragomans should adapt their work to their audience—they were Ottoman subjects and stood to lose far more from displeasing the authorities than from misrepresenting their foreign employers:
Fear tied their tongues: they would much rather risk their employer’s displeasure than the brutal fury of an angry pasha … At times, ingenious interpreters … were known to improvise imaginary dialogues—to substitute speeches of their own inspiration for those really made.13
They were suspect in any case for the mere fact of working for a foreign embassy. Why double the risk by failing to address local potentates with the florid servility to which they were accustomed? Adding a few paragraphs of eternal devotion wasn’t mistranslation. It was life insurance. “All things considered, the wonder is not so much that Dragomans fulfilled their perilous task inadequately, as that they dared undertake it at all.”14
Fidelity was obviously a major issue for Ottoman dragomans, but it didn’t mean what translation commentators in the West seem to mean by “fidelity to the source.” Dragomans needed to prove that they were faithful to the padishah or to the particular Ottoman grandee they were addressing.
It was the grandest of the Phanariot dragomans who paid the highest price for suspected disloyalty. In 1821, the Greek provinces of the Ottoman Empire rose up in revolt. Because they were Greeks as well as Catholics, Phanariot families in Istanbul came under immediate suspicion. Their leader, Grand Dragoman Stavraki Aristarchi, was hanged for treason. Why? Because, as had long been said in the Ottomans’ international language, Traduttore/traditore! Translators are traitors anyway!
This exotic adage has percolated into all Western languages, in Italian and in translation, and has become one of the most commonly touted pieces of expertise about translation in circulation. But save in quite exceptional cases it is wrong, and always was. The translation practice of the dragomans was generally subservient to an outstanding degree—subservient to the purpose of the original, and subservient to the dragomans’ real masters. Treachery was what the masters feared, not what the translators performed. But even if Phanariots did on occasion make d
eals for themselves by misrepresenting their commissioners, the connection between “translating” and “treachery” is of no relevance to modern, thoroughly print-based societies. In a world where you can check the translation against the original, even when it has the form of speech (thanks to the sound-recording devices we have used for the past one hundred years), the principal grounds for the fear and mistrust of linguistic intermediaries that is endemic to oral societies no longer exist. Yet people go on saying traduttore/traditore, believing they have said something meaningful about translation. A thoughtful translator such as Douglas Hofstadter still feels he needs to counter it with a pun in the title of an essay, “Trader/Translator.” 15 We may now live in a sophisticated, wealthy, technologically advanced society—but when it comes to translation, some people seem to be stuck in the age of the clepsydra.
Traditional mistrust of oral interpreters in the Middle East affected Western tourists when visits to the region became practical and prestigious for individuals in the nineteenth century. Tourists had to rely on local intermediaries for contact with the authorities, and hereditary dragoman families turned themselves into guides, guesthouse brokers, and go-betweens for the purchase of antiquities and other delights. As they performed their tasks according to their own traditions of highly adaptive translation, they were despised and scorned. “Dragomania,” the fear and loathing of the intermediaries who ran rings around all but the most canny Western travelers, made a major contribution to the stereotype of the “wily Oriental gentleman” of colonial-era travelogues.16
The tropes of “fidelity” and “betrayal” in translation commentary do not come to us only from a vanished Ottoman past. In seventeenth-century France, several translators of the Greek and Latin classics thought it best to amend the originals to make them correspond more closely to the standards of politeness that ruled behavior and writing at the Court of Versailles. Swearwords and references to bodily functions were simply cut out, as were whole passages referring to drinking, homosexuality, or the sharing of partners. Confident in the absolute rightness of the courtly manners of France, these translators tried to produce translations that were fitter for their target audience, and also (in their view) better and more beautiful works. They were saving the Greeks from themselves by editing out all those primitive blemishes. Purposefully and intentionally adaptive, these many classical texts refashioned for courtiers (or for children) were dubbed les belles infidèles, literally, “beautiful unfaithful [ones] [feminine].”