It used to be that what a translator did was take a text from one language and write a version of it in another language that was as close to the original as possible, but was still accessible to readers in the second language. Critics could argue over which side of the equation deserved more weight, but few disagreed on the principle. No one expected a perfect copy, a text in the target language that perfectly conveyed, word for word, the exact meaning of the text in the source language. I mean, individual translators aimed at that kind of perfection, some believed they had achieved it, but most were content to have come as close as they believed was possible. They accepted that compromise and approximation were irradicable elements of translation. Matthew Arnold held that although exactitude in translation was a noble aim, dubious choices and even the odd howler were bound to creep in, and a translator ought not to worry overmuch about them. The proper aim of the translator, he wrote in “On Translating Homer,” in 1861, was “to reproduce on the intelligent scholar, as nearly as possible, the general effect of Homer.” By “scholar,” Arnold was not favouring academics, he did not mean quibbling pedants; a scholar was simply a generous-spirited reader who had a degree of “poetical feeling” and who “knows that Homer is Homer by his general effect, and not by his single words.” In other words, a scholar was anyone who agreed with Arnold’s definition of translation, and a pedant was anyone who did not.
In Arnold’s day, it seems, there were more pedants than scholars, but over time the scholars have been the ones who have approved of the best translations, even those with superficial wounds. Vladimir Nabokov, a prolific translator of Russian verse as well as an innovative novelist, forgave the occasional lapse: in “The Art of Translation,” he identified “three grades of evil…in the queer world of verbal transmigration.” The first, and lesser, “step to Hell,” he wrote, “comprises obvious errors due to ignorance or misguided knowledge.” For example, translating the word “queer” today without acknowledging or even being aware of its contemporary connotation was a venal, but not a mortal, sin. Languishing somewhere in Limbo was the “the translator who intentionally skips words or passages that he does not bother to understand, or that might seem obscure or obscene to vaguely imagined readers.” It was the translator who pandered to the “notions and prejudices of a given public” who occupied Nabokov’s ninth ring of Hell: “This is a crime to be punished by the stocks, as plagiarists were in the shoebuckle days.” In “the general public” he included “the publisher who does not give a damn for such niceties and always prefers an adaptation anyway.”
In case we are inclined to believe that things have become more stringent in the world of the translator, that the kind of errors pardoned by Arnold and Nabokov are now punishable by banishment, consider this statement by Michael Cunningham, in his introduction to a new translation of Thomas Mann’s Death in Venice, which appeared in 2004: “Any assertion,” he wrote in his introduction to that work, “that a translation can be rendered ‘accurate’ if its blatant errors are corrected underestimates the art and magic of translation….A handful of blunders does not seem likely to alter the fundamental tone of a book, or seriously subvert its meaning.” Which is another way of saying that his chief concern as a translator is to capture, in this case, not the exact words, but the “general effect” of Thomas Mann.
Cunningham went on, however, to explode the entire notion of what translation is. “All novels are translations,” he wrote, “even in their original languages.”
This is to say, in effect, not that translation is a more exacting occupation than even Arnold and Nabokov had imagined, but that all writing is translation. To write is to translate thought into language. In a recent conversation with the Acadian writer Antonine Maillet, three of whose books I have translated into English, she said that “all writing is translating God,” but I don’t think that’s what Cunningham meant. Cunningham is also the author of The Hours, a novel about Virginal Woolf, and I doubt that either Thomas Mann or Virginal Woolf saw themselves as taking down dictation from the Almighty when they wrote Death in Venice or To the Lighthouse. I think Cunningham meant something closer to what Octavio Paz had in mind when he said that “everything is translation: learning to speak is learning to translate.”
This opens up the definition of translation to such an extent that it is now extremely difficult to say what translators do that, say, everyone does every time they open their mouths. When everything is translation, how can Matthew Arnold’s poetically sensitive scholar distinguish between a translation and an adaptation – is Henry James’s short story “Paste” an adaptation of Guy de Maupassant’s short story “Boule de Suif,” or a translation of it? We can leave aside the question of good translations versus bad translations, unless we paraphrase T.S. Elliot by saying that there is no such thing as a bad translation, because a bad translation is not a translation. But if everything is translation, we can’t even make that simple distinction.
I am all in favour of broadening the definition of translation. I would like to be able to say that “Paste” and “Boule de Suif” are the same story without taking anything away from either de Maupassant or James. James must have written “Paste” with de Maupassant looking over his shoulder, and I think that if de Maupassant could have read “Paste,” he would have said, “Yes, Henri, you have understood my story very well. You have exactly conveyed my general effect.” And, on top of that, James wrote a brilliant and original short story of his own. One stone, two birds. Even Nabokov, who admitted that his translation of Pushkin’s Eugene Onegin was at least partly an original poem by Vladimir Nabokov, would have admired that kind of accuracy.