"A word is a bridge built between myself and another... a territory shared by addresser and addressee."
–Mikhail Bakhtin

Bakhtin might say – I'm sure he does somewhere – that all dialogue is an act of translation. Locked into our tiny Leibnizean monads, we realize ourselves by shaping personal identities out of everything from the sound of our mother's cooing to the clichés of Hollywood films. According to Bakhtin, the creation of individual language – which is really the creation of the self– is an act of willing submission: "One's own discourse is gradually and slowly wrought out of others' words that have been acknowledged and assimilated, and the boundaries between the two are at first scarcely visible."

But how can I translate you? That is, how can I render the uniqueness of your particular language, create an "accurate" image of you, not in your language but in mine? In order to get close to your language, shall I be your lover, your partner, or your amanuensis? How can I go into your skin and become you, when I know that no matter how much the creation of an individual language is a social act, discrete voices can never be congruent, individual selves can never quite merge?

In the spirit of the impossibility of translation, I'd like to reveal some dirty secrets – secrets that every translator knows, but not all will admit to.

No translator can translate every author equally well. The problem is that you don't know whom you can and can't translate until you try, and by then it's too late.

Case in point: the translation of War and Peace by Constance Garnett, the great twentieth-century translator from Russian to English, is a masterpiece; the 13 volumes of her Chekhov short stories are awful. No doubt Mrs. Garnett loved Chekhov as much as she loved Tolstoy, but they did not love her equally.

Over the period of a few years, my ex-wife and I translated two Swedish novels into English. (Actually one was Finland-Swedish, which could be compared, somewhat clunkily, to American vs. British English.) We admired both books, though they were very different. P.C. Jersild's After the Flood (Efter floden) is a dystopian novel about a young, Huck Finn-like innocent who travels a post-atomic world trying to carve out for himself a space of love and safety. The Thirty Years' War (Trettioåriga kriget) by Henrik Tikkanen is a sharp, ironic fable about a stubborn Finnish soldier who was never told that World War II ended.

When we started, I was sure that Jersild's book would be the "easier" to translate. The narrative voice is steady, consistent. The pathos of the young man's predicament is accessible, and Jersild fills the book with interesting, touching characters. Tikkanen's satire, on the other hand, is all about shifting tones: in a single sentence he can be witty, nasty, melancholy – and elegant. How could a translation capture all that?

In the course of translating After the Flood, we sensed pretty early on that things weren't going well. Sentences, sometimes whole paragraphs, had to be rewritten and reshaped. Scenes that seemed well-written in Swedish came out over-long in English. I wondered if it had to do with the fact that Jersild's Swedish syntax and punctuation were heavily influenced by English; his language and ours may not have been different enough to create Bakhtin's "bridge." The Tikkanen translation flowed smoothly; every one of his changes of tone seemed to survive translation. I'm still not sure why. I suspect it was because Tikkanen's moody, mercurial literary persona "fit" ours more than Jersild's rather noble one; we were more at home in Tikkanen's skin than in Jersild's. In any case, our translation of After the Flood was just passable, while our translation of The Thirty Years' War really sings.

A lot gets lost in translation. An awful lot. The good translator knows when to throw up his hands and move on to the next sentence.

A little story: at 16, I had a bit part in a high-school production of a charming comedy called Teahouse of the August Moon. It's about a well-meaning American lieutenant in Occupation Japan who helps a village rebuild their bombed-out teahouse. I played an old peasant who approaches the American officer to ask him for a lift to the village. My character doesn't want to appear needy or to lose face, so he doesn't ask for help directly: he simply asks for the time. An interpreter introduces the peasant to the lieutenant, and soon he's riding on the officer's jeep. My sole line, as I remember it, sounded something like "Ima nagi kai'ina."

Jump to 30 years later. I'm coming out of a restaurant with some friends, among them a Japanese jewelry designer. We're late for a movie, I've forgotten my watch, and the line from Teahouse pops out of my mouth: "Ima nagi kai'ina." Kazuko looks at me and tells me the time. Then looks at me again.

"Did I mispronounce it?" I say.

"Not at all. But you asked it in a very curious way. You sounded like an old man from a lower class speaking to a younger person from a higher class. But you were proud, and you didn't want to seem inferior."

Not even Shakespeare could have conveyed all that information in a phrase, and the judicious translator shouldn't even try. Class, emotion, age difference were packed into a language as old as amber. Instead of trying to crack the amber open, better to ask Kazuko to put it in a silver setting and wear it around one's neck.

Some languages don't translate well into other languages at all.

I know in my bones that this is so, but I haven't a clue as to why. One could say that in general, Romance languages translate better into English than Germanic ones, which might have to do with the preponderance of Latinate words that William the Conqueror brought over from France. But Slavic prose works particularly well in English, while poetry by Arabic and Sanskrit masters sounds embarrassingly corny. How come?

The real mystery to me is German. Why is it nearly impossible to do a good English rendering of the classic German authors? Goethe, for example. On the basis of existing translations, Goethe's reputation is completely baffling. Faust in English is postured Romantic drivel. The Sorrows of Werther and Elective Affinities are barely readable. Some of the shorter poems rendered into English are pretty good, but what's all the fuss about?

Goethe isn't the only problem. Brecht's skill in manipulating tone and vocabulary from many different strata of German society comes out in English sounding heavy-handed and polemical; much of the sparkle and linguistic fun doesn't come through. Kafka's ability to combine exquisitely melodic language with wisdom and irony so that it virtually shimmers is largely lost in English. When I'd learned enough German to read some tiny Kafka parables, I remember thinking, "But his language is so light!" And Rilke? Soppy, sentimental – that is, in most English translations.

I have a feeling that German is a smarter language than English. That as well as being rich enough to give voice to a variety of classes and regions, it's able to accommodate the blending of thought and feeling more seamlessly than English can. We Anglos think of German as a hard-sounding, ruthlessly logical language; yet the evidence points elsewhere. It appears that there isn't a lack of subtlety in German, but in us.

Self-translation is a very bad idea.

There are exceptions: Beckett comes to mind. But he may be an exception to the exceptions, since he wrote his manuscripts as "original translations", as it were, and then retranslated them back to the original language – or at least back to the source.

But many contemporary authors – and editors – think that they can translate their own work, or push their translators to absurd lengths, always with disastrous results. Some years ago, we translated a Danish thriller written, like a lot of contemporary Danish fiction, in short, blunt sentences. In Danish you can get away with a series of short sentences; in fact, it builds suspense. But even Hemingway, known for his staccato rhythms, varied the length of sentences; much of the power of English is in the flexibility of its syntax, and varying the length of sentences is one of the most effective ways of changing tone and emphasis. But the Danish editor felt that we'd ruined the author's "style". She couldn't let go of the original Danish; it was still in her ears, and she insisted that we "restore" the length of his sentences. A big mistake. Syntax, phrases, words, sentence length – one must let go of it all in translation. The only thing to hold onto is the essence, the spirit of the work – in Bakhtin's terms, my dialogue with you.

A translation can only be better or worse than the original, never just as good.

Worse, to be sure, but better? Isn't the text itself, respectable again after the age of deconstruction, the be-all and end-all? Translation can only be an approximation, most likely a diminution. The translator can only be the Master's servant, and is often treated worse; not only is he paid execrably, but if the book/play/poem is bad, it has to be his fault.

In my experience, a few good authors – the best ones, I'd venture – treat their translators very nicely. Not only because they're aware that if they are too intrusive, the translator can mangle the work into something unrecognizable, but because they know that a well-stroked translator can bring out qualities only implied in the text itself.

Why this should be so is easy. Think of the original text as a detailed outline or blueprint. The author/architect has noted down very specific instructions, but every word of the final product belongs to the translator/builder. All the details are in the architect's blueprint, but in a thousand invisible ways the builder can enhance the vision, actualize it.

I know this is true – not because I've done it, but because I've had it done for me. I write an occasional column for a Copenhagen daily. But I write the text in English, and put it in the hands of a skilled translator named Niels Ivar Larsen. "How well you write Danish!" say readers and friends who don't read the fine print. In a sense they're right. I do write Danish well, when Niels Ivar pulls out of my murky prose things I've intended but haven't given full voice to. Not that he adds anything. Not exactly. It's all sort of there. But somehow – and this is the magic of intellectual meiosis – something better can be created in that meeting on Bakhtin's linguistic bridge.

For really the operative metaphor isn't dialogue or architecture, but birth. From the meeting of author and translator, monad to monad, emerges the work, a squirming, squealing collection of ideas, sensations, possibilities: texts and subtexts. One always hopes that the baby will be better than its parents; but whoever it turns out to be, one can be sure that it will be different from both.

Published 2007-05-21

Original in English
First published in Kritika & Kontext 33 (2007)

Contributed by Kritika & Kontext
© George Blecher/Kritika & Kontext
© Eurozine