What We Mean by a Faithful Translation
If you are a translator, you are probably familiar with the Italian play on words "traduttore/traditore," or translator/traitor. Clichés such as this often gain currency as truth. Are translators traitors by definition? Certainly not. Translators take their job seriously and apply their skills in the source and target languages to the best of their ability. A pun the translator was unable to preserve, wording that doesn't capture all the nuances of the original phrase, an outright mistranslated word--do not justify the assumption that the entire profession is doomed to fail.
"Lost in translation" is another cliché often taken as inevitable fact. Translation does not lead to loss necessarily. In the excellent Sympathy for the Traitor: A Translation Manifesto, Mark Polizzotti puts it this way: "[I]t would be utopian to pretend that the reader of a translation is truly experiencing the original, or that in the reading of any translation there isn't a degree of difference--difference rather than loss--between the text being translated and the translation itself" (Polizzotti's italics). Of course all translators have come upon words or phrases in the original with no direct equivalent in the target language, and yet skilled translators will find acceptable solutions to such problems. The perfect translation doesn't exist. Excellent, faithful translations abound.
What do we mean when we say a translation, and literary translation specifically, is faithful to the original? To answer the question, I believe we need to start with an exploration of what the translation is faithful to. A literary text is composed of words and the syntax that organizes them, but it also contains subtext and is inserted in a sociocultural context. If we as translators focus on the accuracy at the word level but do not attend to syntax or tone or cultural references, the translated text is bound to be unfaithful to the original to some degree.
At the syntactic level, translators often have to strike a balance between maintaining the rhythm and style of the original work and making slight changes that will render the text more readable in the target language. In a number of languages, long, tortuous sentences are common, while in English readers and writers seem to prefer shorter sentences. Gujarati-to-English translator Jenny Bhatt has mentioned her care to preserve the run-on sentences that are an important element of Gujarati rhythm and at the same time to create an English translation that reads well. (I strongly encourage you to visit her website and sign up for her newsletter.)
I have written about the issue of sentence length in my essay Another Kind of Emergence, published by Hopscotch Translation. I have on occasion split a sentence that in the original Portuguese goes on for seven or eight lines to make the paragraph more manageable for English-language readers. However, I do refrain from splitting run-on sentences and comma splices if they occur in the context of, say, a parent's lecture to a child, a rant, an emotional confession, a ramble. In these cases, the long series of coordinated clauses in the original are meant to convey the spewing of words in urgent, breathless bouts, or the association of thoughts that flow from one to another and another in a stream of consciousness. Especially when clarity is not compromised, splitting these sentences into shorter ones may deprive the reader of experiencing a literary technique the author used deliberately and skillfully and may therefore amount to a betrayal of the original text.
The issue of fidelity may seem more straightforward at the word level, but I don't think it is. Lexical fidelity is not as simple as choosing the right word from a list of synonyms because it is intrinsically linked to fidelity to tone and voice in the original text. If an adjective can be translated as "bright," might other synonyms be more appropriate? Luminous? Radiant? Sparkling? Or simply well-lit or sunny? We need to consider the tone of the narrative and the author's style--serious, irreverent, sarcastic, humorous, polished, direct, lyrical, etc. In translating dialogue, we also need to attend to the relationship between the interlocutors. Is it a close one? Formal? Is there a power imbalance? What emotions are being evoked? Once these questions are answered, the translator may generate two or more adequate choices. Which one makes it onto the translated page is sometimes a matter of instinct--what feels right to the translator.
Some instincts, however, may need to be examined and kept in check. Regina Alfarano, a former professor of mine from long ago, shared in an interview that a translator of William Kennedy's Ironweed had omitted in her Portuguese translation all the swear words because she was a very religious woman and objected strongly to profanity. I have not read the original or the translation, but I believe that even if this translator was able to preserve most of Kennedy's meaning in her translation of Ironweed, she was not faithful to a significant choice he made as an author, and that to me is a failure in translation.
Preserving the tone of the original may indeed be the most important aspect of the original work we need to be faithful to. In an interview to Trinity News, the very accomplished translator Frank Wynne says: “The most basic element of translation is meaning. But it’s also the least important. (...) [Y]ou also need to be able to recreate the cadence and rhythms of the original language, the humour, the wordplay, the bathos…" In both narration and dialogue, language varieties need to be considered too: "The way the language is spoken is structured differently depending on the dialect. This includes the register, intonation, cultural references, and many other nuances."
I mention considerations of meaning and tone in my translation of Marília Arnaud's novel Suíte de Silêncios (Suite of Silence) in the essay I mentioned above. I explain that "[at] first, I translated ['quando o céu se abre em fendas iluminadas'] as 'when the sky opens up in illuminated slits' but reconsidered it. Too common and informal, the word 'slits' did not maintain the register of the original. And 'illuminated,' despite conveying register and tone, did not suit the rhythm and fluidity of the sentence. These I found instead in 'crevices of light.'" I was attempting to strike a balance between fidelity to the words and fidelity to the rhythm of the narrative and to the author's voice. This necessary juggling act often results in the best choice I can make given the resources at my disposal--my skills in both languages, my understanding of the work, and my sensitivity to the various elements that make it what it is.
We often ask if a translation is faithful to the original, and that is indeed crucial, but I find it is also important to ask, is the translation faithful to the target language? Here is an example: I recently read in the English translation of a novel written in Portuguese the sentence "I was encharged to deliver it." I didn't have the original at hand, but I suspect in Portuguese the sentence read, "Eu fui encarregada de entregá-lo." In what I assume to be the translator's attempt to be faithful to Portuguese, he was not faithful to English. The phrase "to be encharged with something" is archaic, while the language used throughout the novel is modern and free of anachronisms. In addition, to be "encharged to do something" sounds ungrammatical in English, while the infinitive in this case is correct in Portuguese. "I was put in charge of it" would have conveyed the meaning adequately and sounded more natural and fluid in the target language--it would have been more faithful to it.
Such linguistic traps are sometimes hard to avoid because translators are deeply immersed in the two languages they are working with. The constant shifting from one language to the other often blurs the border between the two. Speakers of multiple languages who often switch between them experience this too. For example, in a recent conversation with Brazilian friends, I said "prioritizar," then quickly corrected myself: "priorizar." "Prioritizar" is not a Portuguese word--it is an anglicism I had made up on the spot, unwittingly.
False cognates are another linguistic trap translators need to be attentive to. I have seen "estreito" translated as straight, when in fact it means narrow, and I know this was the case because the English and the Portuguese texts were presented side by side. Another potential false cognate between English and Portuguese is "delicado," which may mean delicate in some contexts, but when used to describe a person, it means polite. In the same translation where "encharged" appeared, "delicate" was used to describe a male character, and I suspect that may have been a mistake. I didn't have the Portuguese and English texts side by side, but subsequent mentions of the same character suggest he was polite rather than fragile.
The most competent translators make such mistakes every now and then. Translation work is intense, and in the back-and-forth between the two languages we may become temporarily blind to these problems. Or we may be faced with a fast-approaching deadline and may not have adequate time for our revisions. I have not found a publisher for my book-length translations, so I have not experienced this first-hand, but on my wish list would be the opportunity to work with an expert reader to check my translation for any inaccuracies.
And in full disclosure, I must say I do offer my services as an expert reader. I do not mean to sound self-serving here. I am making a case for collaboration in translation work. Writers in monolingual contexts work with editors who read in their language. In literary translation, it is not always possible to pair translators with editors who can read in the source language. I believe collaborations involving author (if possible), translator, and expert reader offer the best environment in which problems can be identified and corrected and the best quality in translation can be achieved.