Translating Subtexts:

What the Translator Must Know

by Arthur Goldhammer

Talk delivered to

BU Translation seminar

CUNY conference on translation

Brandeis English/Comp Lit Seminar






Translating Subtexts:

What the Translator Must Know

by Arthur Goldhammer

I come before you as a professional translator. I insist on this point because, if professionalism is generally considered grounds for claim on the attention of an audience, when it comes to translation this is notably not the case. As I began to think what I might have to tell you, I took down a favorite volume from my shelf of works on the craft, a collection edited by Reuben Brower called On Translation. In it I turned to Justin O’Brien’s piece, "From French to English," where one may read:

The truth is that no one wants to be a professional translator—partly because the guild enjoys no particular dignity or respect and partly because the occasional translator is generally better than the one who, making a living at his task, is obliged to accept almost any assignment. Now God forbid that, having said as much, I should lose my amateur standing by these remarks on the art of translating.

Tartly put. May I be just as tart in reply? O’Brien goes on to make many intelligent remarks about translating French, remarks that belie the fatuousness he displays here. His assertion is undercut by his own evidence, for he cites instances of bad translation by any number of presumably inspired amateurs. What underlies this nonsense is the theory that good translation requires some sort of mystical sympathy between author and translator. The amateur, not being obliged to accept whatever comes his way, can wait—so the theory goes—until sympathy strikes, until the divine afflatus blows his way.

Nothing could be further from the truth. I take a materialist view of the matter. The sinews and reflexes that translation requires are capable of development through exercise. They are best developed where there is resistance to be strained against. The besetting danger of all translation is that the translator will simply substitute his language for his principal’s. This is the danger to which Walter Benjamin, citing Rudolf Pannwitz, called attention in "The Task of the Translator": "The basic error of the translator is that he preserves the state in which his own language happens to be instead of allowing his language to be powerfully affected by the foreign language." The amateur, gentleman though he may be, is unlikely to suffer the indignities that true perfection in the art requires. He is unlikely to submit, promiscuously, to the domination of many, often antipathetic masters. But one does not know a foreign language until one submits to a generous spectrum of its voices. Translation is a rough trade.

So much for philosophy. No one—Benjamin is a case in point—can make a living preaching about translation. We professionals are paid by the word, and when we are not composing at the keyboard we often are engaged in the kind of finger exercises that build dexterity and speed, hence income. What I propose today is a sort of piano lesson. Together we can walk through some of the questions distilled from daily practice. We can consider some of the sour notes, the couacs, that even the best translators commit from time to time and ask how to lessen their frequency. For want of competence to speak about anything else, all my examples are from French. To those of you who don’t speak French, forgive me.

I’ll begin, as translating lessons so often seem to begin, with an error. This in itself is worth pondering. Some years ago I won a translation prize. At the award dinner both speakers felt it somehow appropriate to introduce the subject of translation by heaping ridicule on what they took to be hilarious gaffes—mercifully not mine. It was as if the very thought of translation elicited the cognate thought, "howler." Perhaps this is related to the vulnerability we all feel when speaking in a foreign tongue, as if each time we shattered the rules of syntax we simultaneously opened chinks in the armor of the self, as if each botched subjunctive were a self-maiming, as if every time we substituted one word for another we exposed some shameful private part. Freud, I once read, found that some patients lost their accents under hypnosis, as if accent were a consequence of embarrassment, of the need to contort our most expressive features to produce proper foreign sounds. There is a species of critic that delights in pointing out errors. I don’t deny the legitimacy of the activity; what I deplore is the delight in exposing deformity. Would that every critic might learn to emulate the tact of Jean Starobinski, who, each time I’ve translated one of his books, has written me using the same formula: "I see that the obscurity of my French has, in a couple of places, led you astray." In hindsight, of course, there was nothing obscure about his French. There is, I think, a lesson to be drawn from the very prevalence of translation error, and it is not that translators as a guild are incompetent swindlers. It is rather that misreading is more common than one might think, a fact that would scarcely have surprised I. A. Richards, to whom Brower dedicated On Translation. Literary theorists have attempted to exploit the evidence of misreading that translation provides in support of their theories. Impressed by their perspicacity, I remain skeptical of their reasoning, but that is a debate which must be postponed to another time and place.

I see that I’ve digressed. Revenons à nos moutons. What must the translator know? Many things often unlinked by rhyme or reason. It is literally impossible to know in advance, or sometimes even after the fact, how difficult a particular translation task is. That is why I’m always irked when a publisher calls and says, "We have a job for you, the language is straightforward, we’ll pay you X." X, of course, is roughly half of Y, the fabulous sum dangled before you as what would be paid if ever the publisher were to come upon a text he considered difficult. I read in The New York Times that Alfred Knopf once told Prof. Parshley, the translator of The Second Sex, that "translation has always been dog’s work—never well paid and seldom if ever bringing the translator any glory." As a long-time resident of the kennel, I’m of the opinion that the quality of a dog’s life is a reflection not on the dog but on its keeper. Be that as it may, there is a sense in which it is far more difficult to translate Le Nouvel Observateur than a work by Michel Foucault. "Straightforward" is not a word that can readily be applied to language, which, like a confidence man, is often most devious when it seems most plain. Consider, just to bring these abstract matters down to the level of concreteness, exhibit A, from a biography of Foucault, in fact, and written, as it happens, by a journalist from Le Nouvel Obs. The book begins:

Le décor est presque saugrenu. C’est un théâtre, situé au rond-point des Champs-Elysées.

I was asked to evaluate the work of another translator. The text began:

The setting was almost preposterous: a theater at the traffic circle on the Champs-Elysées.

What’s wrong with this? Nothing and everything. The translator has made a decision to use the English past for the French historical present, which is fine, though in biography sometimes the English present works better. He has combined the two choppy French sentences into one English sentence, which is excellent. The sense is almost right. But what are we to make of "the circle on the Champs-Elysées?" Rond-point certainly means "traffic circle": the dictionary says so. But hasn’t the translator ever been to Paris? The avenue boasts two famous "traffic circles," if you can call them that, one at Etoile, the other at Concorde. As it happens, the writer isn’t thinking of either of these. He means the Rond-Point des Champs-Elysées, the smaller, less famous circle where, under the glassy-eyed gaze of Le Drugstore, the great avenue joins the avenues Montaigne and Franklin Delano Roosevelt. Perhaps the French publisher’s copy editor is at fault, because Rond-Point should have been capitalized in the French. A pedantic point? Maybe, except for one thing: the real intention of the sentence depends on it. For what is preposterous about the location is that the theater in question is the Théâtre des Champs-Elysées, the first reinforced concrete building in Paris and the site of the famous premier of Stravinsky’s Rite of Spring. The gathering being held there, as described in the remainder of the paragraph, is a memorial colloqium in honor of the dead philosopher, a man more usually associated with such Left-Bank venues as the Collège de France than with this Right-Bank icon of modernism. The occasion being a memorial, the atmosphere was presumably decorously lugubrious, whereas the first performance of Le Sacre triggered a raucous riot. But how much of this can be got into a translation? Not too much. One doesn’t want to overinterpret or weigh heavily on a point the author would prefer to make lightly. Consider:

The setting was almost preposterous: the Théâtre des Champs-Elysées at the Rond-Point.

This modified translation supplies some, but not all, of the information implicit but unstated in the French. The translator needn’t flaunt her knowledge; the naming of the theater is enough to compensate for the English reader’s presumable unfamiliarity with Paris geography. What of the writer’s other assumptions: the Left-Right Bank distinction, the début of The Rite of Spring? These are signs, to those who perceive them, of the position from which the writer speaks. Some readers, whether French or English, will fail to perceive these signs, will know no more of Stravinsky in 1913 than the prince d’Agrigente at the Guermantes’ dinner party in Proust knows of Flaubert. But that is the point, really. These opening phrases establish a degree of intimacy between the writer and the knowing reader—the reader who, like the cocky duc de Guermantes, considers himself anything but what he declares himself to be: a pedzouille, a country bumpkin; and the translator, however much he may deplore the writer’s "insider" tone, the meretricious glitter of false sophistication, had better not interfere, for the manner in which an author strikes that distance from his reader is a fundamental trait of style. Some authors, used to the podium, can only lecture; others can only whisper in the reader’s ear. Here, I think, the manner is one of winks and nudges. I conjure up the image of two habitués of the Latin Quarter drinking espresso in the Café du Panthéon or Le Soufflot. "You’ll never guess where they’re holding the Foucault colloquium," one says. "Where?" the other asks. "C’est presque saugrenu. Le Théâtre des Champs-Elysées." "Mais c’est quoi alors?" "Ben, tu sais bien, cette espèce de blockhaus en béton près du Rond-Point." "Ah, celui du Sacre de Stravinsky?" "Ça y est, mon vieux, tu te rends compte?"

The translator who introduced the extraneous traffic circle into this dialogue of arched eyebrows was thus guilty of an error, but was it really an error of translation? Not, I think, in the ordinary sense of the word. The meaning of the sentences was rendered roughly correctly in English. The failure was not of translation, but of imitation. Let me illustrate what I mean by once again invoking Proust. Here, in my own translation, Marcel is describing Oriane de Guermantes’s talent for imitating the speech of people she knows:

Even if Mme de Guermantes did [her imitation] to perfection, the Courvoisiers would have been as incapable of appreciating it if they had been rabbits rather than men and women, because they had never noticed the defect or accent the duchess was trying to mimic. When she "imitated" the duc de Limoges, the Courvoisiers protested: "Oh, no, he doesn’t speak like that at all. I had dinner with him just yesterday at Bebeth’s. We talked all evening, and he doesn’t speak like that." Meanwhile, any Guermantes with any cultivation at all would exclaim, "Oh, God, that Oriane is a scream! The amazing thing is that while she’s imitating him, she looks like him. I could swear he was right there. Oriane, please, more Limoges!"

The translator must cultivate the gift of the duchesse de Guermantes because every stylist is like the duc de Limoges: it is the essence of style to deform language, to bend it away from its degree zero, its dead center of neutral speech. Style, therefore, resembles error, and no doubt that is why Proust, one of the greatest of stylists, also had such a weakness for linguistic error, which he notoriously and mercilessly mimics, whether it be the oddities of the foreign-born hotel manager at Balbec, the solecisms that Françoise brings with her from la France profonde, or the mispronunciations of the well-born, a signal perhaps of their belonging to an international caste defined by something other than language, a transnational brotherhood that Jean Renoir, in La Grande Illusion, signals by having the fin-de-race aristocrats von Rauffenstein and de Boeldieu lapse into English over dinner. Style is like error, and no doubt that is why a stylist like Proust is so drawn to what Gérard Genette calls the volupté immanente, the intrinsic sensual pleasure of certain deviations from the norm. The duc de Limoges, like the hotel manager at Balbec, errs repeatedly because we all, according to Freud, experience a compulsion to repeat the errors we would suppress. But the writer, like the duchesse, is a sadistic voyeur, drawn by the intrinsic pleasure of imitating the damage to language that these unconscious artists inflict. And the translator must learn to share that pleasure, even when it goes against the grain. Hence, as Proust says in the passage Genette cites as an epigraph to his article, "the duty and the task of a writer are those of a translator."


Since I’m speaking today to comparative literature scholars, I thought it might be useful to add a word about how the view of style I’m hinting at here relates to a view put forward by one of the founding fathers of your discipline, Leo Spitzer, many years ago. In his 1923 essay on "Pseudo-objective Motivation in Charles-Louis Philippe," Spitzer suggests applying "what we have learned from classical philology to stylistic criticism. When we notice something that seems conspicuous in a modern text, we shall, if we search carefully enough, find related phenomena in the same text. Once we have identified the similarities between these passages, we shall be able to draw certain conclusions about the author’s psychological makeup. What the ‘corrupted passages’ are to the philologist editing older texts, the ‘conspicuous passages’ are to the modern philologist intent on psychogenetic explications." He then goes on to apply this reasoning, in somewhat unconvincing fashion I think, to Philippe’s use of the phrase à cause de. Now, there are certain obvious analogies between Spitzer’s notion of the "conspicuous passage" and my idea that style is a deformation of or deviation from neutral diction, grammar, and syntax. But Spitzer, proceeding as a scientist would rather than a musician, focuses on the repeatable phenomenon, the isolatable instance. His approach is interesting but insufficient, I think. Spitzer, to put it in a pithy but perhaps obscurely psychoanalytic way, gives us style as neurotic symptom, whereas what I am proposing is a theory of style as persona. Whoever says persona, of course, says mask, and I think that perhaps the most useful starting place for a theory of style might be none other than Starobinski’s essay on Stendhal’s masks in L’Oeil vivant. So, for those who wish to reinscribe the anecdotal material I will be giving you today in the less jocular terms of literary theory, I give you these intertextual hints: read Spitzer against Freud, Freud against Starobinski, Starobinski against the musician Charles Rosen on The Classical Style, and Rosen against me. To develop the theoretical implications further, however, I would need at the least an NEH grant and run the risk of being called a professor, and as John Maynard Keynes, who was just a bursar and not a fellow of his college at Cambridge once put it, "I will not accept the indignity without the emoluments."


Errors can be signs. When we hear someone say, in English, "Thank you for inviting my wife and I," instead of "my wife and me," we suspect that, having grown up in a home where it might have been common to say, "Me and the missus thank you for the invite," the speaker has, by way of overcompensation toward the "cultivated" norm, substituted the nominative where the objective case is required—a case of misguided or jumped-up politeness. Similarly, when we encounter a translation that renders Rond-Point as "traffic circle," we diagnose a need for more street-prowling or at least for wider reading. And our diagnosis is confirmed when we read, a few pages later, in translation of the French:

Les élèves de seconde [sont] requis pour le "service rural": six semaines d’agriculture pendant les vacances d’été, avec la tâche notamment de détruire les doryphores...

this English:

High-school students were drafted for the "rural service corps": six weeks of farm work during summer vacation, chiefly a matter of exterminating potato beetles. ...

Again, the translation is as impeccable as the Courvoisiers’ appreciation of the duchess’s imitation. The translator has looked up doryphore in the dictionary—a poor one, probably, not Harrap’s or Robert-Collins—and found "Colorado beetle" and then decided that, this being France, he had better check a book on entomology, where he learned that a "Colorado beetle" is a kind of potato beetle with a spike on its proboscis. Laudable diligence. He has not, however, bothered to ask what could possibly be the significance of the three dots in the French and has simply reproduced them in English. Yet these three dots indicate not ellipsis but a pause for laughter. The correct translation requires an appreciation of the context. The passage concerns Foucault’s experience as a young student in the Unoccupied Zone during World War II. During the Franco-Prussian war doryphore became slang for a German soldier, perhaps because the German army subsisted on potatoes, perhaps, too—for Latinists at any rate—because the Germans wore helmets with spikes on top (etymologically, of course, a doryphore carries a spear or spike). It remained in the vocabulary of the 1940s along with other slang terms for the Germans: les frisés, les Fridolins, les Fritz, les Boches. At a time when youths elsewhere in Europe were killing not beetles but each other, the sentence is an ironic expression of Foucault’s distance from the war. Hence the translation should read:

High-school students were drafted for the "rural service corps": six weeks of farm work during summer vacation, with the primary task of killing not Boches but bugs.

Rond-Point, doryphore: the error, in both cases, is the same, and it is also Marcel’s error when he first encounters the duchess whose gift of mimickry he will later surpass: it is the error of naïveté. And of course we know that the purpose of style is to dispel naïveté, because Raymond Radiguet has said so, as Roger Shattuck in his book on Proust usefully reminds us: literature, Radiguet wrote, is a means to "déniaiser les lieux communs." Shattuck, incidentally, remarks that a phrase like this is a "translator’s nemesis—or fortune." He gives three translations: "To give meaning to the commonplace." "To teach clichés the facts of life." "To make a man of a maxim." I like them all, but may I be permitted to try out my own? Déniaiser les lieux communs: to wise up worn words.

Let us circle once more around the Rond-Point. You might think that ancillary information of the kind I’ve been discussing could be important only in the kind of quirky context that occurred at the beginning of this particular biography. It could hardly be germane to the translation of a great literary text. But consider:

L’un [des gens au café] racontait que sa voiture, se croyant arrivée au pont de la Concorde, avait fait trois fois le tour des Invalides; un autre que la sienne, essayant de descendre l’avenue des Champs-Elysées, était entrée dans un massif du Rond-Point, d’où elle avait mis trois quarts d’heure à sortir.

The Rond-Point again! Moncrieff translates:

One reported that his carriage, thinking it had got to the Pont de la Concorde, had circled three times around the Invalides, another that his, in trying to make its way down the Avenue des Champs-Elysées, had driven into a clump of trees at the Rond Point, from which it had taken him three quarters of an hour to get clear.

Moncrieff knows his Rond-Point from his Concorde, even if he unaccountably chooses to strip the place of its hyphen-passerelle. What bothers me, though, is that "clump of trees" for massif. Why does he add this precision, trees? Wouldn’t it be more likely for a carriage to become fouled in a hedge? Might we resolve the issue by searching for a picture of the Rond-Point circa 1895? Am I being pedantic? Perhaps. A critic who disliked one of my translations said that "pedantry is the soul of translation," and at the time I hotly disagreed. But that is not to say that pedantic attention to detail is always misplaced. I’m particularly persuaded of the importance of detail in this passage because, as you recall, what Proust is describing is an impenetrable fog that descended one night on Paris, a fog that made familiar landmarks unrecognizable, as in the sentences just cited, yet that brought out with previously unmatched clarity certain hideous features in the character of those assembled in the café. The fog, of course, is a metaphor for the Dreyfus Affair, or for race prejudice generally, which blinds us to the humanity we share while sharpening to perverse keenness our awareness of accidental differences and distinctions. Marcel, when he stumbles through the café’s revolving door, is sent to the "Jewish side" by the proprietor, who, unaware that the young man is accompanied by the most un-Semitic marquis de Saint-Loup, exemplifies the racist’s exquisitely blind insight. All of this episode has to do with the alternate obfuscation and revelation of identity, and details, sensual details, take on extraordinary symbolic importance. Some of Moncrieff’s lapses, which might be of less moment elsewhere, therefore have significance here:

Les jeunes nobles qui formaient l’autre partie de la clientèle [i.e., on the "non-Jewish side"] ... considéraient Dreyfus et ses partisans comme des traîtres, bien que, vingt-cinq ans plus tard, les idées ayant eu le temps de se classer et le dreyfusisme de prendre dans l’histoire une certaine élégance, les fils, bolchevisants et valseurs, de ces mêmes jeunes nobles dussent déclarer aux "intellectuels" qui les interrogeaient, que sûrement, s’ils avaient vécu en ce temps-là, ils eussent été pour Dreyfus, sans trop savoir beaucoup plus ce qu’avait été l’Affaire que la comtesse Edmond de Pourtalès ou la marquise de Galliffet, autres splendeurs déjà éteintes au jour de leur naissance.

The young nobles who composed the rest of its patrons ... looked upon Dreyfus and his supporters as traitors, albeit twenty-five years later, ideas having had time to classify themselves and Dreyfusism to acquire, in the light of history, a certain distinction, the sons, dance-mad Bolshevists, of these same young nobles were to declare to the "intellectuals" who questioned them that undoubtedly, had they been alive at the time, they would have stood up for Dreyfus, without having any clearer idea of what the great Case had been about than Comtesse Edmond de Pourtalès or the Marquise de Galliffet, other luminaries already extinct at the date of their birth.

Not one of Moncrieff’s prouder moments. Bolshevism has had many crimes laid at its door, but driving young nobles mad with dance, as though the prophet armed were some militant Dionysus, is not one of them. Nor does Proust say that these young swells were Bolshevists, merely bolshevisants. Waltzing left-wingers, not dance-mad Bolshevists. It would be more idiomatic, moreover, to say that "ideas sorted themselves out" rather than "classified themselves." The end of the phrase ought to read "without having any clearer idea of what the great Case had been about than they did about" the vanished luminaries. Note, too, the reference to "the Case" rather than "the Affair." Was this current in English in Moncrieff’s day, or was he rather pedantically, and foolishly, avoiding a false cognate that has since left its impress upon English?

Or consider this nearby passage:

D’autres, épris d’aviation, tiennent à être bien vus du vieux garçon du bar vitré perché au haut de l’aérodrome; à l’abri du vent, comme dans la cage en verre d’un phare, il pourra suivre, en compagnie d’un aviateur qui ne vole pas en ce moment, les évolutions d’un pilote exécutant des loopings, tandis qu’un autre, invisible l’instant d’avant, vient atterrir brusquement, s’abattre avec le grand bruit d’ailes de l’oiseau Rock.


Others, keen upon flying, seek to stand well with the old waiter in the glazed bar perched on top of the aerodrome; sheltered from the wind as in the glass cage of a lighthouse, they can follow in the company of an airman who is not going up that day the evolutions of a pilot practicing loops, while another, invisible a moment ago, comes suddenly swooping down to land with the great wingèd roar of an Arabian roc.

Well, the roc, of course, is a bird of fable from The Arabian Nights and reputed to carry off elephants, though I’m not sure that the precision "Arabian" roc really helps matters if one doesn’t know this, and "great wingèd roar," though one might stand for it in a pinch, isn’t quite the same thing as a plane landing with a huge din like that of the roc’s wings. But the word that bothers me most in this translation is "evolutions." Because in France, of course, on regarde les évolutions des danseurs, or, as Jean Renoir wrote in his memoirs, "[Le Guignol des Tuileries] m’attirait par ses décors grisâtres dans lesquels évoluaient des personnages aux couleurs ternes." The point is that évolutions in French is a perfectly ordinary word, whereas "evolutions" in English is just too special, and what Moncrieff really wants to say is that the spectators followed the "progress" of a pilot practicing loops; sometimes a translator’s skill at manipulating his own language lets him get away with too much, with putting in too many words, or affecting a needless Latinate turn, where a less skilled writer of English would simply stumble over himself. So the good writer must be even more careful of his translating discipline than the less good one.

I hope you will bear with me if I stick for a while longer to Proust’s text. It is precisely because his means as a writer were so varied that he provides such a rich lode for examining the problems of translation. And Moncrieff’s errors are different from those of the first translator we looked at. Moncrieff is not naïve; things are not usually lost on him, although occasionally he too fails in that simple way:

Il avait épousé quelques mois après la ravissante fille d’une autre princesse de Luxembourg, excessivement riche parce qu’elle était la fille unique d’un prince à qui appartenait une immense affaire de farines.

He had married, some months later, the charming daughter of another Luxembourg Princess, extremely rich, because she was the only daughter of a Prince who was the proprietor of an immense flour-mill.

Now, this "flour mill" was more like General Mills, I should think, than the mill on the Floss, and one might prefer a different choice of words. And "une autre princesse de Luxembourg," applied to the "excessively rich" daughter of an industrialist, carries in French a clear anti-Semitic connotation that the translation loses with its flat "another Luxembourg Princess," where the capitalization signals that Moncrieff, unlike the bridegroom, has mistaken the bride in question for the genuine article rather than the devious counterfeit. But in general such straightforward matters of fact are not lost on Moncrieff, a man of the world. His problem is not naïveté but limitation. He works at the level of the sentence, whereas Proust writes on a vast scale. Proust keeps many balls in the air at once, invokes more levels of language than most of us know exist, and Moncrieff, intent on one or two of them, frequently misses others. The duchess, for example, issues a dinner invitation to Marcel:

Vendredi vous ne seriez pas libre, en petit comité?

That little echo of parliamentary language, en petit comité, evidence of Oriane’s gift of mimickry, of the vaunted esprit de Guermantes, disappears from the translation:

On Friday, now, couldn’t you? There are just a few people coming.

Intent on getting the colloquial rhythm of the first phrase, Moncrieff drops the mimetic charge of the second. Or take this sentence:

Comme dans tous les mariages de ce genre, l’origine de la fortune est l’obstacle, comme elle est aussi la cause efficiente.

Moncrieff’s translation,

As with all marriages of this nature, the origin of the bride’s fortune was the obstacle as it was also the deciding factor,

turns the passage from sly narratorial rumination, the slyness connoted by the Aristotelian overtones of cause efficiente with its implicit cosmic hierarchy of teloi and entelechies, of material, formal, efficient, and final causes, into a clunky authorial obiter dictum that would have embarrassed Balzac.

More seriously, consider this sentence, in which Marcel remarks on the pain he suffers when the elaborate erotic fantasy he has constructed around the person of Mme de Stermaria is punctured by a note informing him that the lady has been called away on a family matter. Remember that the note has arrived only a matter of moments before, followed hard on the heels by the arrival of Saint-Loup, whose presence stirs this new line of thought:

Le sillage que laissait en moi le regret de Mme de Stermaria ne voulait pas être effacé si vite...

Moncrieff renders this as:

The scar left by my disappointment with Mme. de Stermaria was too recent still to be so easily healed ...

But a scar does not heal; it is the enduring mark of an old wound and takes time to form. The English thus entirely misses the force of the French metaphor, sillage, from the wake that a passing ship leaves but briefly on the surface of the water. And sillage itself has etymological-metaphorical affinity with sillon, the furrow left by a plow as it harrows the earth. So the word sillage evokes at once deep, searing, wrenching, cutting, harrowing and above all recent pain as well as brevity and superficiality: it is the mot juste for a just-punctured fantasy, whereas "scar" is but a cliché.

I don’t mean to make a whipping boy of Moncrieff. Someone had to be first to translate Proust, and Richard Howard will no doubt improve on Kilmartin as Kilmartin corrected Moncrieff. Moncrieff’s translation has some admirable strengths. Many of these have been remarked on by others. His English is supple enough to handle Proust’s extremely long sentences, certainly a primary requirement. Shattuck observes that Moncrieff passes the "reading aloud" test. But I would like to comment today on one of Moncrieff’s qualities that has not, to my knowledge, attracted attention previously: he had the courage to reproduce his author’s errors, which in Proust’s case are usually not errors at all but intentional uses of error for expressive purposes. Perhaps he had read Proust’s essay on Flaubert, where the author of A la Recherche remarks on the deft use of pronouns in L’Education sentimentale but also notes that, when not deliberately availing himself of certain resources inherent in the French pronoun, Flaubert is occasionally prone to mistakes of antecedence. Now, it happens that the kind of pronoun play that Proust is admiring here is inevitably lost in translation, because it depends on the availability of gender. But Proust’s virtuoso play with pronouns, more dazzling even than Flaubert’s, can be rendered into English in at least one marvelous passage in his text, and Moncrieff, perhaps unwittingly, did not bungle the opportunity. The occasion I have in mind occurs when Marcel is invited for the first time to visit the duchess at home. It is a rather lengthy passage—indeed, its effect depends, as we shall see, on its length—and I will not read it in full, but I will pause long enough for you to look it over while I read Moncrieff’s English. I’ve marked the pronoun-referent pairs to bring out the structure of the passage:

Je vis déboucher, majestueuse, ample et haute dans une longue robe de satin jaune à laquelle étaient attachés en relief d’énormes pavots noirs, la duchesse. Sa vue ne me causait plus aucun trouble. Un certain jour, m’imposant les mains sur le front (comme c’était son habitude quand elle avait peur de me faire de la peine), en me disant: "Ne continue pas tes sorties pour rencontrer Mme de Guermantes, tu es la fable de la maison. D’ailleurs, vois comme ta grand’mère est souffrante, tu as vraiment des choses plus sérieuses que de te poster sur le chemin d’une femme qui se moque de toi," d’un seul coup, comme un hypnotiseur qui vous fait revenir du lointain pays où vous vous imaginiez être, et vous rouvre les yeux, ou comme le médecin qui, vous rappelant au sentiment du devoir et de la réalité, vous guérit d’un mal imaginaire dans lequel vous vous complaisiez, ma mère m’avait réveillé d’un trop long songe.

I saw emerging, majestic, ample and tall in a flowing gown of yellow satin upon which stood out in relief huge black poppies, the Duchess herself. The sight of her no longer disturbed me in the least. There had been a day when, laying her hands on my forehead (as was her habit when she was afraid of hurting my feelings) and saying: "You really must stop hanging about trying to meet Mme. de Guermantes. All the neighbors are talking about you. Besides, look how ill your grandmother is, you really have something more serious to think about than waylaying a woman who only laughs at you," in a moment, like a hypnotist who brings one back from the distant country in which one imagined oneself to be, and opens one’s eyes for one, or like the doctor who, by recalling one to a sense of duty and reality, cures one of an imaginary disease in which one has been indulging one’s fancy, my mother had awakened me from an unduly protracted dream.

Now, never mind what’s wrong with this translation, perhaps most obviously the incessant "one, one, one," that drones on like a bagpipe from Moncrieff’s native Scotland. Look at what is remarkably right: Proust has created a fascinating chiasmus, as the literary jargonists like to say, a crossing over from la duchesse and sa vue to elle avait peur and ma mère. By committing the solecism of having the pronoun elle precede its referent, ma mère, he sustains, for the time it takes to read this very long, indeed artificially extended, sentence, a doubt in the reader’s mind about the identity of the woman in question. That doubt has been most ably, most shrewdly, most cunningly planted by the previous sentence, which so cannily uses the possessive pronoun sa (sa vue ne me causait plus aucun trouble) to refer to la duchesse. Indeed, we are still under the illusion that Mme de Guermantes is speaking when we come upon her name in the direct quote. Until this discovery plunges us into confusion, we are thus primed to think that an antecedent of the subsequent female pronoun elle has been established, and thus to believe, at the ground level of grammar, however astounding it may seem to our intellect, that the duchess has been in the habit of laying hands on Marcel’s forehead. And so, for a moment, we feel, in the queasy confusion of the syntax, what it was like for Marcel to confuse, not to say fuse, his love for his mother with his love for Oriane, or perhaps more provocatively, his unavowable hatred of his mother with his judgment of the duchess’s failings, a judgment of which this entrée into her salon will begin to make him capable. This chiasmus by pronoun is an astonishing literary device, and a translator bent on "smoothing out" the rough spots in the original could make it disappear in the blink of an eye.

Perhaps it is time to move on from Proust. But before I do—and then à contre-coeur, since Proust is the womb of my French soul—I want to consider one more passage, because it has been commented on by others and because it brings together a number of themes implicit or explicit in what has gone before: the link between error and style, the literary uses of grammar, the danger of favoring certain levels of language over others in the translation of complex texts. The dialogue I have in mind occurs when Françoise interrupts Albertine and Marcel as they are about to do whatever it is they do do on Marcel’s bed. All we know for certain is that they kiss, and it is perhaps the most memorably described kiss in all literature, drawn out over many pages that end thus:

Mais hélas!--car pour le baiser, nos narines et nos yeux sont aussi mal placés que nos lèvres mal faites—tout d’un coup, mes yeux cessèrent de voir, à son tour mon nez, s’écrasant, ne perçut plus aucune odeur, et sans connaître pour cela davantage le goût du rose désiré, j’appris, à ces détestables signes, qu’enfin j’étais en train d’embrasser la joue d’Albertine.

But alas—for in this matter of kissing our nostrils and eyes are as ill placed as our lips are shaped—suddenly my eyes ceased to see; next, my nose, crushed by the collision, no longer perceived any fragrance, and without thereby gaining any clearer idea of the taste of the rose of my desire, I learned, from these unpleasant signs, that at last I was in the act of kissing Albertine’s cheek.

We might quibble about this: "collision" is perhaps de trop and "unpleasant" is rather weak for détestable, but Moncrieff has by and large done a good job of catching the moment in which blindness literally becomes the price of Marcel’s insight that he is "in the act of kissing Albertine." But before this happens, the maid Françoise intrudes on the young couple, ostensibly because, the afternoon drawing to a close, she thinks they may need a light. Marcel, who has been working himself up to this devoutly-wished consummation for some pages, is taken aback by the interruption and protests that the light is awfully bright. Whereupon Françoise, mangling her subjunctive, asks:

Faut-il que j’éteinde?

Teigne? glissa à mon oreille Albertine, me laissant charmé par la vivacité familière avec laquelle, me prenant à la fois pour maître et pour complice, elle insinua cette affirmation psychologique dans le ton interrogatif d’une question grammaticale.

Moncrieff translates:

"Do you want me to extinglish it?"

"—guish!" Albertine slipped into my ear, leaving me charmed by the familiar vivacity with which, taking me at once for teacher and accomplice, she insinuated this psychological affirmation as though asking a grammatical question.

Shattuck is rather more deft:

"Do you want me to sniff it out?"

"...snuff?" Albertine murmured in my ear, leaving me charmed by the lively familiarity with which, taking me at once for master and accomplice, she insinuated this psychological affirmation in the form of a grammatical question.

Perhaps I should explain, for any non-French speakers among you, that the French verb falloir, to be necessary, requires the subjunctive in any subordinate clause it governs, and the correct conjugation of the verb éteindre would imply that FranÇoise should have asked, "Faut-il que j’éteigne?" This is the grammatical question to which Proust alludes: what Albertine murmured (or slipped?) in his ear was the correct final syllable of the subjunctive. But it seems to me that both translators have missed an important point. Marcel is delighted not only by the grammatical question on which he is consulted as teacher (or master?) but by the "psychological affirmation." What is this mysterious psychological affirmation? I confess, I didn’t get it either on first or second reading, and perhaps both Moncrieff and Shattuck did get it but couldn’t translate it, as I confess I can’t. For the point is that teigne in French is not only the morphologically correct ending for the verb, it is also a noun, une teigne, not a common word, which means "shrew" or "bitch" (it also means "tinea, ringworm"). And so the mystery is dispelled, but we still do not have a translation. Can the psychological point be made perceptible without resorting to periphrase or, horror of horrors, a translator’s note? The solution of this problem is left as an exercise.


Earlier I compared translating to playing the piano, but perhaps it would be more apt to say that translating is adapting to the piano a work originally scored for another instrument or, in the case of a writer as resourceful as Proust, for a whole symphony. A Liszt may be able to pull off a piano transcription of Beethoven without descending to travesty; a lesser musician may not. But for the moment let’s stay with the metaphor of an instrument-to-instrument transcription. One of the first problems to be dealt with is the fact that each instrument has its own set of standard patterns, devices, and techniques, some of which may not be available on the other instrument. A guitarist can "bend" a note by stretching a string, but a pianist can only "fake" it, as jazz musicians say, by clever substitution of an appoggiatura or trill. Much of the daily business of translation thus requires a different kind of training from what I have been discussing so far. Until now I’ve been focusing on training the ear. Now I want to talk about training the hand, about perfecting English equivalents for the devices, ornaments, grace notes, and trills that come naturally to players of the French horn.

I translate a great deal of scholarly French, and I am constantly struck by the change that has taken place in French scholarly style over the past century, a change far greater, it seems to me, than anything discernable in English. Consider, for example, these two passages:

Ce fut le 5 janvier 1757, à six heures du soir, que le Roi, montant en carrosse au château de Versailles pour aller souper à Trianon, fut frappé au flanc d’un coup de couteau. La blessure était légère, mais on craignit que l’arme ne fût empoisonnée; Louis XV se mit au lit et se confessa.

Au premier rang des obligations de la veuve, la prière, on l’a vu dans la Tradition Apostolique, héritière en cela de 1 Tim. 5,5.

I don’t think I have to tell you which of these is from Lavisse, published circa 1900, and which from a contemporary work of historical scholarship. The presence of the passé simple and imperfect of the subjunctive, the symmetrical periods, the stately calm of the prose in the face of the violence of the event, date the first passage; the jerky colloquialism, incomplete sentence structure, breathless compression, and quirky use of héritière in the second suggest that the writer is emulating the hustle-bustle, métro-boulot-dodo style of the newsweeklies, Le Nouvel Obs and L’Express. Indeed, many scholars in France—et non des moindres—write for these newsweeklies and have adapted their styles accordingly. The best use this new scholarly colloquial dialect as one more arrow in their quiver. With it they can hit targets otherwise unattainable. For example, one historian can begin an essay on Napoleon with a sentence of classical balance worthy of Lavisse:

La Révolution en France n’a pas aimé la vieillesse, et cette règle n’épargne pas le plus grand, peut-être le seul de ses héros, Napoléon Bonaparte.

But a few sentences later he slips into the more modern idiom:

Si bien qu’on peut partir, pour le comprendre ou pour le cerner, de ce qui l’enracine si profond dans l’histoire de France, lui le Corse, l’Italien, l’étranger, le "Buonaparte" des douairières de la Restauration; à savoir, de son élection par la Révolution française, dont il a reçu l’étrange pouvoir non seulement d’incarner la nation nouvelle—d’autres l’avaient eu, Mirabeau, Robespierre—mais de l’accomplir enfin.

This second excerpt—again an incomplete sentence—meanders on in the manner of a lecturer thinking, rectifying, amplifying as he goes.

What is the translator to do with such prose? Recall Pannwitz’s dictum, which I cited earlier: "The basic error of the translator is that he preserves the state in which his own language happens to be instead of allowing his language to be powerfully affected by the foreign language." When it comes to scholarly translation, I think Pannwitz is wrong, or at any rate not entirely correct. For to succumb entirely to the mode for the scholarly colloquial would result in disastrous, unreadable English. There are respectable authorities who hold that scholarly translation and literary translation have nothing in common, that the scholarly translator is obliged, in order to preserve intact his author’s all-important ideas, always to be literal. I could not disagree more, though I do concede that there is a difference between a work of history, say, and a novel. The distinction I make is this: literature is for those who have the eyes, ears, and wits to savor its beauties, and the translator is obliged to preserve as many of them as he can; scholarly prose is written for consumption primarily by a scholarly community, and the translator must respect the linguistic norms of the target community even when that means sacrificing some of the flavor of the original. Consider the example quoted above:

Au premier rang des obligations de la veuve, la prière, on l’a vu dans la Tradition Apostolique, héritière en cela de 1 Tim. 5,5.

I would translate this as follows:

Foremost among the obligations of the widow was prayer; 1 Timothy 5:5 had established this principle, and Hippolytus had followed it in his Apostolic Tradition.

Note what has happened here. The English is written in complete sentences. The "on l’a vu" has been dropped entirely as a needless tic, a cliché of the French horn that has no place on the piano. Instead, I’ve added the name Hippolytus, the author of the Apostolic Tradition, last mentioned more than two pages earlier in the text. This backward reference is really the only point of on l’a vu—to remind the reader that the work was mentioned some while back—so why not simply spell it out? The comma splice has been replaced by a semicolon. The second clause has been rephrased to obviate the awkward construction with héritière, but the inverted structure of the first clause, beginning with "foremost," has been retained as the only really important stylistic feature of the French: the emphasizing of the widow’s obligations, the topic of the paragraph. Foremost was prayer, leading the reader to suspect a next, a second and perhaps a third element in a series. To have meddled with this, to have placed "prayer" first, would have disrupted the structure of the paragraph.

Now, consider: you’re being paid by the word, and there aren’t many words in this sentence. All the small decisions implicit in the foregoing paragraph must be made rapidly, or you’ll starve. You’ve got to develop a feel for this kind of thing, an easy way of scooping up the French sentence and tossing it onto the screen of the word processor in English. It’s like fielding a ground ball: it’s almost impossible to explain in words how you do it, and anyone who tried to follow such a description would look pretty foolish on the ballfield. But any spectator can see when the job is being done well.


I know, I’m mixing my metaphors: first the piano, then baseball. Let’s go back to music for a moment. Suppose that our Liszt, our piano transcriber, is faced with the problem of rendering the horn call at the beginning of a scherzo or an antiphon between viola and oboe. The piano lacks the right sonorities. Now, to complicate the problem even further, suppose that the antiphon in question is actually a musical quote, Mozart in the middle of a symphony harking back to the theme of a Bach fugue. A solution is nearly impossible. Some would say it doesn’t matter, that the listener won’t get the allusion anyway and the music can be enjoyed without it. Probably true. But in translation such matters can sometimes be of crucial importance. Consider this sentence from a book on collaboration during World War II:

L’antisémitisme sûr de lui et dominateur a bien les traits d’un délire général qui ne suppose aucun terrain hors d’atteinte du "cancer juif."

Now, one may translate these words in all innocence and without a doubt convey the gist of the meaning to the American reader. But any French speaker will tell you that there is an important "nuance," as they say in French, that tells you a great deal about the author’s attitude toward the subject he is describing. That nuance is in the form of a buried quote, for it happens that after the 1967 war in the Middle East General de Gaulle referred to the Jews as "un peuple d’élite, sûr de lui et dominateur." La petite phrase du général provoked a furor as it was endlessly repeated and widely commented on, to the point where it has left a mark on the language, a scar far more obvious than, say, "grassy knoll" or "Where’s the beef?" in English but perhaps comparable with "the light at the end of the tunnel" or Henry Kissinger’s "peace is at hand." In such cases a note is, however undesirable, all but inevitable.

General de Gaulle was actually quite profligate in dispensing these petites phrases. Another that crops up constantly is a phrase from the opening sentence of his Memoirs: "Toute ma vie je me suis fait une certaine idée de la France." Richard Howard translated this quite nicely as, "I have always thought of France in a certain way." The problem is that the authors who have picked up on the general’s superb subsumption of the nation by his cogito are often playing on the word "idea," which has vanished from the translation.

In passing, by the way, let me remark on the structural similarity of that opening sentence of de Gaulle’s Memoirs with the opening sentence of Proust’s novel:

Toute ma vie je me suis fait une certaine idée de la France.

Longtemps je me suis couché de bonne heure.

When Richard Howard lectured at B.U. some years ago, he had a number of interesting things to say about the translation of this opening sentence of A la recherche, and many of you will have read Roger Shattuck’s brilliant analysis of it. It is remarkable, I think, that twentieth-century France has produced two great works whose subject is the domination by an individual of the vicissitudes of time and events, and that both begin with the designation of an expanse of time followed by a first-person pronoun followed by the passé composé of a reflexive verb that bridges the gap between the past of experience and the present of enunciation. Contrast this with the opening sentences of two of the century’s other great works, Les Mots and L’Histoire de la folie à l’âge classique:

En Alsace, aux environs de 1850, un instituteur accablé d’enfants consentit à se faire épicier.

A la fin du Moyen Age, la lèpre disparaît du monde occidental.

Sartre’s passé simple establishes the mock-heroic tone of his memoir from the opening chord. It is as if he had written: "In Rome, around 75 B.C., a captain weary of the army agreed to serve as emperor." Availing himself of his freedom to negate, he severs with his choice of tense the thread that ties his forebears to the inauthentic self he claims to have shed. Foucault’s historical use of the present is more radical still. The time is out of joint: by a violent feat of mind the writer places himself in the past in order to speak its otherness. We must remember, in reading sentences such as these, how the men of the French Revolution saw themselves as beginning time anew. The old calendar was abolished, and the new, revolutionary calendar was made to coincide with both the laws of nature (it began with the equinox) and the rights of man (it commemorated the founding of the First Republic). Time thus enters the French consciousness in a way we, of gentler tradition, can scarcely appreciate, whence the uneasiness I often feel when forced by certain iron laws of language to obliterate the sheer audacity of the French verb and its tenses.


Words thus wear themselves into the mind, not only the memorable words of statesmen but the fixed phrases that recur in daily speech and do so much to mark the level of language, the tone of style, the character of the language user. Sartre’s characterizations in Le Sursis, which depend largely on such linguistic tagging, have the economy and some of the flatness of film characterizations that rely solely on visual cues:

N’ayez donc pas peur. Nous ferons toutes les concessions qu’il faudra, il n’y a pas de limites. Et puis l’Allemagne peut s’offrir le luxe de reculer, qui oserait appeler Ça une reculade? On dira que c’est de la générosité.

Je me disais: c’est pas ordinaire, aussi, qu’ils me réveillent si matin, dit père Croulard.

Je suis passé prendre mon écharpe, dit le maire. ... Ils ne parlent pas de réquisition.

Il y’a une autre affiche, dit le lieutenant.

Bon Dieu! dit le maire. Bon Dieu de bon Dieu! Et voilà que Ça recommence!

J’ai fait la guerre, moi, dit le père Croulard. Cinquante-deux mois sans blessures.

Ca va, dit le maire. Voux avez fait l’autre, vous ne ferez pas celle-ci. Et puis vous vous en foutez, vous, des réquisitions.

Il faut faire quelque chose, dit le lieutenant. Il faut marquer le coup.

Le tambourinaire est malade, expliqua le maire.

Je peux le remplacer, dit le père Croulard.

Le tambourinaire, dit le lieutenant. Vous allez me faire sonner le tocsin. Voilà ce que vous allez faire!

In the first passage we have the bourgeois intellectual in bad faith, whose argument mimics logical sequence and inevitability: et puis has the force of what mathematicians call an if-then statement, although here it expresses nothing more than a pious wish. This intellectual professes an ability to read between the lines, but of a text that is only a fairy tale. The translator’s aim must be to catch this nervous logic, this false front woven of a blend of the esprit de géométrie with the esprit de finesse. In the second passage it is from the bare bones of desiccated cliché that we infer the attitudes toward the impending war of the hoary veteran, the shocked français moyen sensuel and homme républicain, and the gruff officer whose traditionalist imagination of war extends no further than the alarm bell. Hence the success of any translation depends on catching the scarcely articulate emotion in phrases like si matin, Bon Dieu de bon Dieu! Et voilà que ça recommence!, and Voilà ce que vous allez faire!--expressions figées whose meaning is apparent but whose translation is the summum of the art.


One way, of course, that language directs itself repeatedly down certain channels has to do with its physical nature. For language is, in one of its embodiments, sound, and sound, by imposing rules of its own in ignorance of the rules of language, is sometimes the basis for a particular choice of words. Take this sentence from a lecture on Voltaire:

Il y a un Voltaire latent et un Voltaire patent.

Latent-patent: the pair tumbles trippingly off the tongue, and the translator must be very careful. Latent-patent in English strikes me as overkill, because patent is such a very rare, very prissy word. Latent-manifest has unwelcome Freudian connotations. Overt-covert evokes too vividly unwanted thoughts of war. Perhaps: a visible Voltaire and a not-so-visible Voltaire.

Of course this property of language as sound to suggest associations of words is the basis of the pun, perhaps the most vexing of all intralinguistic phenomena for the translator. It is interesting in this connection to note that the critic Serge Doubrovsky has succeeded in writing the untranslatable novel, Le Livre brisé, a text that makes the pun its principal stylistic device:

Nouveau-nés, mioches, là où les ébats me blessent.

This is actually more than a pun, because it plays not simply on a word but on a cliché: là où le bât me blesse is a common idiom meaning roughly "where the shoe pinches." Ebat, however, is a literary word for frolicking, reveling, playing about, with well-established sexual connotations. Mioches, by the way, means tots, brats. Perhaps a rough equivalent to the condensation of all these intralinguistic phenomena that is achieved in the French would be:

Babies, brats, that’s where the screw pinches.

This Englishing is slightly more indecent than the French, but not bad. Because such language play so delights this author, who seems to have set himself quite literally Radiguet’s goal to déniaiser les lieux communs, Le Livre brisé is an excellent whetstone for sharpening the translatorial eye. I say it is untranslatable because the requisite trouvailles are not always available, and, worse, they cannot be extended at will, whereas a play on words in the source text may extend over many pages. A case in point is Oriane’s pun Taquin le Superbe for Tarquin le Superbe in Proust. Moncrieff hits on the absolutely marvelous Teaser Augustus, and I doubt that this can be bettered. But consider its effects. At one point in the text, after this example of the esprit de Guermantes is repeated for the hundredth time, we are given to understand something about the cultural ken of the princesse de Parme by the fact that she must have the pun explained to her. She is not quite familiar with the name Tarquin. Now—to American eyes at least—it is one thing not to be familiar with the legendary king Tarquinius Superbus, another not to know Caesar Augustus. So the poor princess is made to look even simpler than she really is. Nevertheless, I concede to Moncrieff the irresistibility of his translation.

I once had such luck myself: for vade mecum de va-nu-pieds in a text by Michel Tournier I found "beggar’s Baedeker," which I thought was absolutely wonderful until, years later, I came upon the more precise "barefoot Baedeker" in the travel section of The New York Times. Now I’m not sure that the whole alliterative business didn’t originate in English only to be translated by Tournier into French.


To close, I shall avail myself of my nasty, brutish, and short service career. In the army every lecturer was required to begin instruction with these words: "Gentlemen, in the next hour you will learn X, Y, and Z." After precisely fifty-five minutes of teaching, the instructor would come to attention behind the lectern and announce, "Gentlemen, in the past hour you have learned X, Y, and Z." This always struck me as an unduly optimistic view of educational intercourse, and perhaps it is the source of the military’s irrepressible and, when you think of it, necessary optimism. Because I subscribe to Gramsci’s well-known dichotomous view of the matter, I won’t presume to tell you that you have learned anything, but I do hope at least that I will not have wasted your time. I thank you for your attention.