by David Bellos
Strachey took it for granted that psychoanalysis was a science. Scientific terminology in English traditionally relies on Latin and Greek roots to forge new words for new concepts. However, Freud himself wrote in a language that uses compounds of quite ordinary words in the natural and social sciences. Thus, where in English we use bits of Greek for hydrogen and oxygen, German uses only “plain words”: Wasserstoff is “water stuff,” Sauerstoff is “sour stuff,” but such terms are no less technical and precise than their Greek-based counterparts in English. Consequently, where Freud says Anlehnung (“leaning on”), Strachey coins anaclisis, and for Schaulust (“see-pleasure”), he invents scopophilia. Many now common words of English—ego, id, superego, empathy, and displacement, for example—were all first invented in Strachey’s translation of Freud, to replace the equally technical but less recondite neologisms of the original: Ich, Es, Überich, Einfühlung, and Verschiebung.6
Strachey’s approach is quite unexceptionable if Freud’s writings are seen as contributions to social or medical science. We can test that in a back-translation exercise. What could Freud have written had he wanted to coin a term in German for the English neologism scopophilia? The norms of German-language science writing of his era would have led him inevitably toward a compound noun such as Schaulust.
If, on the other hand, works such as The Interpretation of Dreams are assimilated not to science but to literary creation, then Strachey’s English, which gives a version that is tonally and stylistically distant from the original, could easily be seen as a misrepresentation.
In France, a large and coordinated team has been engaged since the 1980s in producing the first “Complete Works” in French. The enterprise aims to restore the German specificity of Freud, treating him less as the inventor of a new science than as a writer of a particular (and rather strange) kind of literary prose. Indeed, the team’s leaders have declared that Freud didn’t write German at all but “Freudish,” “a dialect of German that is not German but a language invented by Freud.” The result is widely regarded as incomprehensible in French—but then, if “Freudish” isn’t German, it wouldn’t have been easy to read in the original, either …7
The tangled disputes over Freud in English and French would not arise if it were clear how to categorize the field to which his work belongs. In most social-science translation, the problem does not arise. Because it is believed in many places that the best work in social science is done in the United States, translation of social science from English typically retains some linguistic features of the original, to authenticate the quality of the work. But in literature, there is no such collective agreement about where the “top model” lies. Should a new foreign novel in translation conform to the manner and style of some existing writer of English prose? Some would say, Of course not! What we want is something different from the familiar patterns of Philip Roth. Others would say, Of course it should! We want to read something that matches our existing conception of novelistic style in English prose. The book may have been written in Albanian or Chinese, but if it’s a good novel, then it should sound like one—of the kind we know.
There is no resolution to this squabble. You could say that literary translation is easy because, in the last analysis, you can do what you like. Or you could say that literary translation is impossible, because whatever you do, serious objections can be raised. Literary translation is different from all other kinds. It serves readers in a quite special way. Modestly, often unwittingly, but inevitably, it teaches them on each occasion what translation is.
TWENTY-EIGHT
What Translators Do
Speakers of any natural language repeat themselves and others all the time, and to do so they use their natural facility to rephrase together with a well-filled toolbox:
• they can replace one word with another of like meaning (synonymy)
• they can take one part of the expression and replace it with a longer and more elaborate one (expansion)
• they can take one part of the expression and replace it with a dummy, an abbreviation, a short form, or nothing at all (contraction)
• they can take one part of the expression and move it to a different position, rearranging the other words in appropriate ways (topic shift)
• they can use the relevant tool from their language kit to make one part of the expression stand out as more important than the others (change of emphasis)
• they can add expressions that relate to facts or states or opinions implicit in the original in order to clarify what they (or their interlocutor) just said (clarification)
• but if they try to repeat exactly what has been said with the same tone, pitch, words, forms, and structures, they do not succeed (unless they are also gifted, sharp-eared, and well-trained impersonators, and probably employed in the music hall)
Translators do exactly the same things when they repeat the words of another, and the fact that their “afterspeech” is in what we call another tongue makes no difference at all to the range of discursive devices they use.
But they use these tools to support an overriding aim that is not necessarily relevant to voluntary or inadvertent repetition in interaction in the same tongue. They seek to preserve the force of the original utterance—not only the overall meaning of what has been said but the meaning that the saying of it has, and to do so in a way that is appropriate to the specific context in which the second formulation is to be heard or used. They are not trying to change anything—whereas when we repeat something without translating it, we usually intend to make some small or large difference to it.
Here’s a tiny example of the kind of changes translators make in order not to change anything much at all. In the multilingual “in-flight magazine” supplied to travelers on the Eurostar train, a page is devoted to graphics demonstrating the size and achievements of the whole enterprise of high-speed rail through the Channel Tunnel. One of the bubbles features “334.7 km/h,” which is glossed in English as “The record breaking top speed (208 mph) a Eurostar train reached in July 2003 when testing the UK High Speed 1 Line.” It is followed by the following French text:
Le record de vitesse d’un train Eurostar établi en juillet 2003 lors du test d’une ligne TGV en Grande-Bretagne.
The suppression of the “miles per hour” speed in the French translation might be seen as simply conventional—but the obvious reason for its omission is that it is of no relevance to French readers, who do not generally know how far a mile is anyway. More interesting is the French assertion that 208 miles per hour was the top speed of the train doing the test, whereas the English asserts that the train’s top speed broke a record. What record? Well, in Britain, just about every record—no train had ever gone faster on a British track. But it’s not a record for France, where TGVs have exceeded that speed many times. So for the French not to be frankly counterfactual, the translator has to rephrase and recontextualize. However, the real subtlety in the recontex-tualization is when the “UK High Speed 1 Line” becomes just “a high-speed line in Great Britain” in French. French readers do not need to know the embarrassing fact that Britain still has only one such line, when the French have many, and so they had also better not be told the proper name of a piece of railway engineering that is unique exclusively in British terms. Now linked more closely than ever by a fast train, Britain and France still provide two quite different contexts of use for even the simplest expressions. Translations naturally rephrase the message to adapt it to its alternative context of use.1
Literary translators have a less clear idea of the “context of use” of their work than translators of all other kinds. Actually, they don’t know for sure that it will have any end use at all. Many translated works (including many of great merit) sell pitifully small numbers of copies and disappear into a black hole. The only real “client” of a literary translation is an imaginary reader—the reader that each translator invents in his head.
That’s the real reason why, when it co
mes to the transmission of cultural goods, translators tell themselves they are trying to produce an equivalent effect.
There are two difficulties with this commonly mentioned criterion of translation art: “equivalent” and “effect.”
Translations do have effects. They may make readers laugh or weep or rush to the library to find more books of the same kind. They can even have quite sinister effects, as the following historical anecdote shows.
In 1870, Otto von Bismarck, the German chancellor, released a statement to the press about his sovereign’s negative reaction to a request from the French ambassador that the German royal family should commit itself to never accepting the throne of Spain. The statement also reported that the Kaiser didn’t want to talk to the French ambassador again and had sent him a message to stay away by the hand of the “adjutant of the day”:
Seine Majestät der König hat es darauf abgelehnt, den französischen Botschafter nochmals zu empfangen, und demselben durch den Adjutanten vom Dienst sagen lassen, daß Seine Majestät dem Botschafter nichts weiter mitzuteilen habe.
The “adjutant of the day”—Adjutant vom Dienst—names a high-ranking courtier, an aristocratic aide-de-camp. But it happens to be almost identical to a word of French—adjudant. When Bismarck’s statement was received in Paris it was instantly translated by the Havas news agency service and wired to all newspapers, which reprinted it in the “special extra” that went on sale straightaway. In the Havas version, Adjutant is not translated, but left in its original form. The effect of that one word was enormous. French adjudant means “warrant officer” (“sergeant-major” in Britain). It therefore seemed that the French ambassador had been treated with grievous disrespect by having had a message from the Kaiser taken to him by a messenger of such low rank. The French were outraged. Six days later, they declared war.
It’s likely that the overall effect—the outbreak of war—was what Bismarck intended at that time, but it is implausible that he sought to achieve it by drafting a statement in such a way as to lead to its being misunderstood through the existence of a false cognate of a German word in French. After all, Bismarck didn’t decide to leave Adjutant in German in French translation—the Havas agency did.
In life generally, and in translation in particular, we are not very good at calculating the effects that our words and actions will have.
When translating a crime novel by Fred Vargas, I came across a comically grandiloquent passage of direct speech that recycled a famous line from Victor Hugo. To re-create what I thought would be an equivalent effect of misplaced hyperbole I substituted a barely altered quotation from a speech by Winston Churchill. It didn’t work. A reviewer reprimanded me for inserting Churchillian language where the original had none. Can I blame her for not knowing what motivated the effect that I sought? Of course not. Using “Churchill” for “Hugo” was just an entertaining mind game. You can’t require readers to notice that the switch was supposed to produce an equivalent effect, because there’s no way of assessing whether it does that or not.
A similarly futile submission to the doctrine of equivalent effect can be found among the cans of sound recordings used by Jacques Tati for his Oscar-winning movie, Mon Oncle. Before it was released, Tati conceived the ambition of producing an English-language version himself. He reshot several scenes that included public signage, painting over École, Sortie, and so on with School, Exit, et cetera. It was then pointed out to him that the change of visible language would create confusion as to where the action was really located. His solution to that problem was to change the background music track of the English-language version to make it sound more French, and that’s why the Tati archive contains cans labeled ambiance française pour version anglaise—“French atmospheric music for the English version.” That didn’t work, either. Despite the care with which it was done, My Uncle never had an “equivalent effect” because distributors and audiences loved the French original so much. The English version with its “French effects” ran for a few weeks in a single movie theater in New York and then disappeared for fifty years.
Servile adherence to the ideology of equivalent effect can lead translators a merry dance and give rise to unforeseen effects—if they are seen at all. The investigator at the center of an unfinished “literary thriller” by Georges Perec called “53 Days” is looking into the disappearance of a thriller writer by the name of Serval. He comes across Serval’s last unfinished novel on the writer’s desk and is told by the typist that one chapter of it at least was copied out from another book. The investigator looks more closely at the two texts—Perec gives us the two-page original, which he invented—and notices that some of the words have been changed in the plagiarized version. Oddly, they are all twelve-letter words, and there are twelve of them. He writes them out in capital letters, and they naturally make two word-squares:
LAMENTATIONS RESURRECTION
CALLIGRAPHIE STENOGRAPHIE
SECHECHEVEUX TAILLECRAYON
SACHERMASOCH ROBBEGRILLET
MITRAILLEUSE KALEIDOSCOPE
READERDIGEST HEBDOMADAIRE
CARICATURALE PAROXYSTIQUE
INTEMPORELLE METAPHYSIQUE
FOOTBALLEUSE OCEANOGRAPHE
HAMPTONCOURT CHANDERNAGOR
QUELQUECHOSE JENESAISQUOI
FORTDEFRANCE SALTLAKECITY
The investigator stares at the two lists for a while, but as he can’t see any sense in them, he puts them aside. End of chapter.
One day, when I had already started translating the novel, a graduate student burst into my office in Manchester to ask if I had noticed that the diabolical Perec had actually placed a huge clue in the word list printed in the left-hand column (shown above). Reading one letter per row from top left to bottom right in a diagonal line, you get the name of a mountain massif in southeastern France that is also the first word in the title of a famous novel by Stendhal. I hope you can see. At the time, nobody—not even the editors and publishers of Perec’s posthumous novel—had seen it. Bravo! I said to the student. So what am I supposed to do?
What I did in mindless implementation of the idea of equivalent effect was this: I doctored the English translation of the pseudo-extract to make it include twelve twelve-letter words that, when written out as a list, preserve reference, self-reference, and truth value with respect to Perec’s left-hand column:
LAMENTATIONS
CALLIGRAPHER
FACUPROSETTE
SACHERMASOCH
MORTARBARREL
NEWYORKTIMES
EXORBITANTLY
CRAFTYARTFUL
HUNDREDMETRE
HAMPTONCOURT
CLEARLYGUESS
FORTDEFRANCE
But having replanted the invisible clue, and feeling rather pleased with myself, I went one further and invented a purely fictional list to stand in lieu of the twelve words that Serval had used to mask the original. These words had to fit plausibly into the same places in the plagiarized text, so my choices for List 2 had a retroactive effect on List 1 and consequently on the sentence formulations in the translation of the supposed source. Rome wasn’t built in a day. But because the task was so mind-bendingly tricky I decided to give it a personal point that is not present in the French. Here are the two lists in English:
LAMENTATIONS BENEDICTIONS
CALLIGRAPHER PENCRAFTSMAN
FACUPROSETTE KALEIDOSCOPE
SACHERMASOCH CARLOFRUGONI
MORTARBARREL DEDIONBOUTON
NEWYORKTIMES SMITHSWEEKLY
EXORBITANTLY TOOEVIDENTLY
CRAFTYARTFUL STUPIDFUTILE
HUNDREDMETRE TRAMPOLINING
HAMPTONCOURT TRIPOLITANIA
CLEARLYGUESS ALMOSTINTUIT
FORTDEFRANCE NORTHDETROIT
Is the effect “equivalent,” after all that work? I’m not aware that my simulation of the game Perec played has had any effect on readers at all. Or else the fan mail is twenty years late.
An even more obvious trouble wit
h the idea of an equivalent effect is that there’s no scale available for measuring equivalence. “Effects,” especially holistic impressions left by extended works, can’t be extracted from people and measured against one another. Nor can any one reader give an independent measure of the effects made on her by two language versions of the same text. That’s because a reading of a text always happens in a language—not in between. The distinction between language A and language B is problematic enough, but one thing is sure: there is no linguistic no-man’s-land in the middle, just as there is no midpoint between Dover and Calais where you can stand on the water and look on French and English from the outside at the same time.
A bilingual reader may have a perfectly trustworthy judgment of whether a translation communicates the same meaning as its source. But can such a person, however smart and subtle, ever reasonably say that this “Baudelaire in German translation” has on her an effect equivalent to the effect that that “Baudelaire poem in French” has? Such an assertion would be radically unverifiable—and in my view it is also a meaningless string of words. “Baudelaire in French” has a whole range of different effects on me at different times, and it surely has an even wider range of effects on the community of readers as a whole. Of which one does the “effect” of a translation aim to be the equivalent?