On Not Reading in Translation
Lawrence Rosenwald
For more than thirty years, I've followed an unusual rule in my reading: not to read in translation any work I can read in the original. My motives for doing that aren't clear to me, and probably they're too deeply rooted for me to figure them out. The consequences, though, are clear, and of some interest; it turns out, in my view at any rate, that following this rule has given me a useful standpoint for thinking about what it means to read translations into one's own language, what it means to read texts in languages that aren't fully one's own, what conversations can take place between people doing the former sort of reading and people doing the latter.
I take my title from Virginia Woolf's beautiful essay, "On Not Knowing Greek." In that essay, Woolf does two things. First, she shows us that we can't know Greek:
it is vain and foolish to talk of Knowing Greek, since in our ignorance we should be at the bottom of any class of schoolboys . . . We can never hope to get the whole fling of a sentence in Greek as we do in English. We cannot hear it, now dissonant, now harmonious, tossing sound from line to line across a page. We cannot pick up infallibly one by one all those minute signals by which a phrase is made to hint, to turn, to live.(39, 55)
Second she shows us something about how she reads Greek nonetheless, in the face of that necessary ignorance, how she reads what she can't fully know, how much she gains from that reading. "Not knowing," we see, helps her to have literary experiences that knowledge would block off.
I adopted my rule before I read Woolf's essay, and I recognize the differences between her sort of not knowing and mine. But I can't think of many essays that throw themselves as boldly as hers does at the question of what we know and don't know as readers, at what we learn both from knowing and from not knowing. I've borrowed her title, not to claim her support, but to do her homage.
First, then, some details of just how the rule works.
1) It's hard to say just when, in my study of a given language, I've chosen to let the rule apply. Probably it's when I've found that there's some literary work, any one literary work, that I can read in the original. The onset of feeling obligated, the moment when I feel the rule's force and moral suasion, has come surprisingly early. For German, the initiating experience may have been reading one of Kafka's parables, or a fairytale, or maybe Rilke's strange fable, The Lay of the Love and Death of Cornet Christoph Rilke – that I'm not sure of. I am, though, pretty sure that I stopped reading German in translation right after my first year of formal study, when I was 21.
The moment of transition is both exhilarating and terrifying. Emerson gets it just right: “When the time comes in the scholar’s progress unawares when he reads pages without recurrence to his dictionary, he shuts up his book with that sort of fearful delight with which the bridegroom sits down in his own house with the bride, saying, ‘I shall now live with you always.’”
2) "Reading in the original" doesn't mean reading in the original without benefit of dictionary, or commentary, or for that matter published translations; I use whatever there is to use. What it means is taking the original as the basic text, obliging myself to read every word of it, regarding all translations as provisional commentaries.
3) It's important to distinguish reading in translation, where the translation takes the place of the original, from reading of translation where the translation is read in relation to the original. I do, in this sense, lots of reading of translation.
4) My actual practice is both more and less strict than the rule might suggest. Certainly when I ask my students to read works in translation, I read the translation they're reading. When I was studying Emerson, and wanting to understand Emerson's mode of reading, I read Goethe's Wilhelm Meister in Carlyle's translation, because that's what Emerson read. On the other hand, the rule would let me read Dostoyevsky in translation, and Cervantes, and The Dream of the Red Chamber. In practice, though, I increasingly tend to read only what I can read in the original. And I'd prefer, even with minimal Spanish, to puzzle over the first page of Don Quijote than to read three chapters of any translation.
One risk of living by such a rule is feeling self-satisfied, so it's important before going any futher to acknowledge two of the rule's bad consequences. First, it's limited my literary experience. I sometimes wish that I would let myself read, say, Sebald in translation, or Hugo, or Calvino, since I'd undoubtedly get to read more of them. I like to think of myself as a cosmopolitan critic, and in some sense I am; in another sense, though, I'm more parochial than I would be if I read more widely in translation.
Second, the reading I do in languages other than English isn't fully competent reading. When I contrast my reading in English with my reading in other languages, I'm acutely aware of my mastery in the former domain, my limitations in the latter. Eva Hoffmann has a wonderful description of what it was like for her to read English with authority for the first time; she was looking at the beginning of Eliot's "Prufrock," and she says,
My eye moves over these lines in its accustomed dry silence; and then – as if an aural door had opened of its own accord – I hear their modulations and their quiet undertones. . . . suddenly I'm attuned, through some mysterious faculty of the mental ear, to their inner sense; I hear the understated melancholy of that refrain, the civilized restraint of the rhythms reining back the more hilly swells of emotion, the self-reflective moody resignation of the melody. . . . Bingo, I think, this is it, the extra, the attribute of language over and above function and criticism. (186)
I know what she means, I know what "the extra" feels like, and I know that I don't have it in any of the non-English languages that I feel the obligation to read
So if I read, say, Natalia Ginzburg's "Inverno in Abruzzi" in Italian, never having been in Italy nor written an Italian essay nor conducted an Italian conversation, occasionally needing a dictionary, reading at about half the speed at which I might read a comparable writer in English, what am I getting from reading that poignant sketch in Italian that I wouldn't get from reading her in Dick Davis' assured translation of it, and from whatever other reading of Ginzburg that I'd have the time to do if I read her in English?
With his sovereign freedom from fear of inconsistency, Emerson also wrote, “I should think it in me as much folly to read all my books in originals when I have them rendered for me in my mother’s speech by men who have given years to that labor, as I should to swim across the Charles River whenever I wished to go to Charlestown." That remark is very important to me. It reminds me of what's quixotic, even perverse, about my enterprise, but also helps me to begin suggesting why I can’t bring myself to give it up.
For one thing, whether you swim the Charles or walk or drive across it, you get to the same place; the bite of Emerson's comment has to do with that fact. Whether you get to the same place by reading Goethe's Faust in German – which is, when one comes to think about it, an odd idiom, since in some sense Goethe's Faust is "Goethe's Faust in German" – and Bayard Taylor's English version of that work, or even Louis Macneice's or Randall Jarrell's, isn't clear.
For another, why the focus on getting somewhere? Emerson's apparent concern is getting to Charlestown. We know from Montaigne, though, that "the journey not the arrival matters," and our own experience confirms that distinction. And Thoreau is with Montaigne here, in his calculation of who gets sooner from Concord to Fitchburg, the man who rides the railroad or the man who goes on foot:
I say to my friend, Suppose we try who will get there first. The distance is thirty miles; the fare ninety cents. That is almost a day's wages. I remember when wages were sixty cents a day for laborers on this very road. Well, I start now on foot, and get there before night; I have travelled at that rate by the week together. You will in the mean while have earned your fare, and arrive there some time to-morrow, or possibly this evening, if you are lucky enough to get a job in season. Instead of going to Fitchburg, you will be working here the greater part of the day. And so, if the railroad reached round the world, I think that I should keep ahead of you; and as for seeing the country and getting experience of that kind, I should have to cut your acquaintance altogether. (NA 1746)
Maybe it's absurd to swim the Charles to get to Charlestown. Is it, though, absurd to walk there or bicycle there rather than drive? (Am I absurd in choosing to climb stairs rather than ride escalators?) Is success in travel, is success in reading, to be defined only by speed? Reuben Brower's term for what we usually call "close reading" was "reading in slow motion," and the slowness of its motion was precisely what Brower was celebrating.
There's a fear at the heart of Emerson's comment, the fear that if you always had to read the original you wouldn't get enough experience - experience of Charlestown, experience of literature. "Read the best books first," said Thoreau, "or you may not get a chance." But how much should that fear be trusted? Don't we know from experience that fears of this sort usually mislead us?
The swimmer of the Charles would arrive in Charlestown dripping wet, and in no state to join in ordinary conversation; the swimmer is in fact a little like unsociable Thoreau, or like the person who says, "so what translation did you read?" Maybe, in fact, that's what Emerson is thinking about - not that it's inefficient to get to Charlestown by swimming the Charles, but that it somehow puts the swimmer at odds with Charlestown society. What kind of conversations, after all, can the swimmer and the driver take equal part in? What kind of rooms would they both be welcome in? And what’s the value of the decorum that keeps swimmer and driver apart?
Whatever my motives for refusing to read in translation, they're not affected by whether a given translation is good or bad; they have to do with what happens whenever we read in translation, whenever we read in the original. Consider, as an example, John H. P. Marks's fluent translation of the beginning of Céline's Voyage au bout de la nuit:
It all began just like that. I hadn't said anything. I hadn't said a word. It was Arthur Ganate who started me off. Arthur, who was studying medicine the same as me, a pal of mine. What happened was that we met on the Place Clichy. After lunch. He seemed to want to talk to me. So I listened. "Don't let's stay out here," he said. "let's go inside." So I went along in with him. "It's grim," he said, "out here on the terrasse. Come this way."
(Ça a débuté comme ça. Moi, j'avais jamais rien dit. Rien. C'est Arthur Ganate qui m'a fait parler. Arthur, un étudiant, un carabin lui aussi, un camarade. On se rencontre donc place Clichy. C'était après le déjeuner. Il veut me parler. Je l'écoute. "Restons pas dehors! qu'il me dit. Rentrons!" Je rentre avec lui. Voilà. "Cette terrasse, qu'il commence, c'est pour les oeufs à la coque! Viens par ici!")
I don't, as noted, know enough to hear Céline's French the way Hoffman describes hearing Eliot's English. But I can see something of what happens to that French in Marks's translation. Tenses get changed and regularized; the French, unlike the English, moves from tense to tense, and spends a lot of time with a present tense that's being used to describe past events. The French carabin, pungent slang for a medical student derived from the word for a carrion beetle, gets rendered by the un-pungent explanation of what the term means. The emphatic, quick "rien" of Céline's third sentence gets expanded and slowed into "I hadn't said a word." French capitalization and quotation systems are turned into English ones. "C'est pour les oeufs à la coque" ("it's for boiled eggs"), like carabin, gets translated in such a way as to explain the meaning of the idiom – "it's grim" – rather than to retain its metaphor. Céline's "il veut me parler" becomes "he seemed to want to talk with me," with a gratuitous distinction between seeming to want and actually wanting.
Now probably Marks knew more French than I do, didn't have to look up carabin, as I did, didn't have to puzzle over the idiomatic significance of oeufs à la coque. Probably he had good reasons for making the alterations he made. Certainly many of the those alterations could be defended. Some, moreover, e.g., the almost automatic adaptation of capitalization and quotation conventions, don't at all affect traits of Céline's text that are central to his style or thematic to this work.
Still, two important facts remain, and would remain even if Marks were a translator of genius. First, for readers of the translation who have not read the original, there is simply no way to tell what the translator has done, no way to ascertain what the relation is between a given feature of the translation and the original. The translation might give us everything we need to know – as a translator, I believe that some translations can give that. It might give us nothing, or next to nothing. The point is that there is no way to tell. We do not know whose work we are reading, interpreting, savoring, judging; we expose ourselves to dizzying uncertainty, we are making a profession of faith. (I might describe my own practice by saying that I can't bear that sort of uncertainty, do not have that sort of faith.)
Moreover, by reading translations into our own language we expose ourselves to the bad consequences of our own linguistic mastery. When I read Céline's French or Ginzburg's Italian, I'm painfully aware of what I do not know - the denotations and tones and histories of particular words, the implicit meanings of tenses and syntactic modes, the norms against which the artist creates significant deformations, the gamut of possibilities within which the artist makes choices. Expert readers of a translation into their own language, on the other hand, cannot know what they do not know; they cannot escape, we cannot escape, I cannot escape, from expertise. Reading Marks's translation, I know why and how "it all began just like that" feels different from "it all started just like that," where in ordinary English discourse "pal" gets used and where not. I cannot not know these things. And I know them so vividly that I cannot keep constantly in mind that they may have nothing, literally nothing, to do with Céline. Or, in Virginia Woolf's words: "[translators'] language is necessarily full of echoes and associations. Professor Mackail says 'wan,' and the age of Burne-Jones and Morris is at once evoked" (56).
Imperfect reading of the original has some positive virtues, too. Woolf's essay evokes some of them, though without being quite explicit about how they work. She writes beautifully of "the words themselves which, in so many instances, we have made expressive to us of our own emotions, thalassa, thanatos, anthos, aster – to take the first that come to hand; so clear, so hard, so intense, that to speak plainly yet fittingly without blurring the outline or clouding the depths Greek is the only expression" (56).
"Clear," "hard," "intense" – true enough, and clarity and hardness and intensity are surely things we seek as readers. But the words are those things partly because we don't know enough to soften them, to complicate them, to veil them with habit. Did thalassa ("sea") have clarity and hardness and intensity for the ancient Greek who used the word every day to refer to the water she lived next to, bright one day, muddy the next, always present, as little noticed as a sunset? The intensity Woolf is finding in these words is in part a consequence of their not being her words. Yet no reader of Woolf's essay can miss seeing how alive her response is to these words that she does not own.
Reading well means, among other things, renewing our capacity to find language strange and full of surprise. That can happen in our own language, of course; it's what good writers are trying to let happen. But the weight of habit is enormous, the task of cleansing the doors of perception is enormous. If we can stop noticing sunsets, we can stop responding to words and sentences. What helps Woolf to respond so fully is her ignorance. I think something comparable is happening when I read in languages that are only partly my own, even if that fact is accidental for me and necessary for Woolf. (Accordingly, I sometimes read very familiar English-language texts, e.g., Moby-Dick or Lord of the Rings, in non-English translation, simply to make the texts strange again.) I think that noticing, while not fully understanding, Homer's clanging hexameter and bright formulae, or Flaubert's intimidatingly exact vocabulary, or Sebald's acrobatically balanced extended adjective modifiers, humbles and reanimates me as a reader, allows me a chastened and fresh sense of meter and diction and syntax.
There's a wonderful passage in Elias Canetti's autobiography about his conversations with Isaac Babel:
I thought I would hear a lot from him about the great Russian [writers]; probably, though, they were too obvious to him, and perhaps it seemed to him boasting to expatiate on the literature of his own compatriots. Or perhaps there was more at issue than that, perhaps he shied away from the inevitable superficiality of such a conversation; he himself moved about in the language in which the great works of that literature had been written, and I at most could know them from various translations. We would not have been talking about the same thing. (274)
Is Canetti right? Maybe; Babel refused to have the conversation, or at least Canetti thought he refused. But numerous conversations of this sort do take place. What should they, what can they, be about?
Theorists of translation like to explore the distinction between what translation can retain and what it must inevitably lose. Ezra Pound says that phanopeia, the play of images, makes it across; that melopeia, the play of sound, does not; that logopoeia, "the dance of intellect among the words," is sometimes successfully rendered and sometimes not. Our own ordinary experiences suggest that translations can retain what's glimpsed when reading fast, must lose what's seen when reading slow. Plot and character are retained more easily than tone or rhythm. Hector dies, and Priam goes to Achilles' tent to recover his body, whatever the translator decides about how to render Homer's recurring epithets. Or, as Umberto Eco writes, “If, in the Red and the Black, Julien directs two pistol shots at Madame de Rênal, and only hits her with the second, then in every translation he must shoot twice and not once. Otherwise the reader couldn't wonder, as certain realists do, about what happened to that goddamned first bullet.”
Now in theory, participants in conversations between original-readers and translation-readers would distinguish between aspects of the original that translation can retain and aspects it can't. Rejecting Canetti’s sense that there’s nothing to discuss, that original-readers and translation-readers are simply “not talking about the same thing,” we would figure out where the two sorts of readers are talking about the same thing: the number of bullets, images like “petals on a wet black bough.” Other matters – style, rhythm in all its senses, diction, tone, excellence, mediocrity, failed and successful humor, allusion – would be off the table.
In practice, though, that's not going to happen. We want, after all, to talk to one another. If we've read works in translation, and have had strong responses to them, surging ideas about them, we want to articulate those responses and ideas, we want to make a case for them if they matter to us, even if we're talking to someone who's read the "same" works in the original. That is, after all, what Canetti wanted to do with Babel, and absent the intimidating presence of a great Russian storyteller, it's what we're in fact going to do with our friends and colleagues.
And that's a good thing, even as regards theory, because the distinctions evoked above are imperfect, and are at odds with one another. We simply don't know, that is, not in any systematic way, what gets lost in translation and what gets retained or found. And since we don't know these things, why should we defer to what other people say about them? The fact is, the only way we're ever going to find out when original-readers and translation-readers are talking about "the same thing," and when they're not, is by way of a thousand painstaking, candid, undefensive conversations between them.
These reflections lead me to offer two recommendations. First, I think that scholars and critics and common readers would all benefit from having more experiences of reading with an awareness of not fully knowing, if only because the experience of reading and not knowing puts the experience of knowing in high relief, makes us delight in our usual grace and mastery.
Second, I dream of some reforms in how we talk and write, reforms to let us explore questions of translation that we now consistently suppress. I imagine, for example, what would happen if in all discussion of authors and texts, speakers and writers had to make clear whether they had done their reading in translation or in the original, and if in translation then in which translation. I imagine what would happen if in all reviews of translated work, and not just in reviews of translated poetry, a significant portion of the review had to be devoted to assessing the translation's successes and failures.
I'll conclude by making these abstract recommendations more concrete. In his introduction to Bruno Schulz's Sanatorium under the Sign of the Hourglass, translated from the Polish by Celina Wieniewska, John Updike writes: "Bruno Schulz was one of the great writers." Updike might be right. I wouldn't know, I haven't read Schulz. But for me Updike's judgment rings hollow. It's not that it's based, as I conjecture, on having read Schulz's work in translation. It's rather that it's based on suppressing that fact, on concealing the process of reading and judgment. Maybe Schulz's virtues come across even in ordinary translation. Maybe Celina Wieniewska's translation is a work of genius. (Updike describes Schulz's work as being "finely" translated by her, but how does he know?) Maybe Updike is intuiting something that isn't quite there. Maybe Updike's account of Schulz would seem a travesty of Schulz to any reader of Schulz's Polish.
The point is that we lose something, lose a lot in fact, by the conventions that let Updike presume that it doesn't matter whether he read Schulz in Polish or only in English. (One thing we lose, by the way, is a just appreciation of the translator's role; Updike's name is on the book's cover, Wieniewska's is not.) We also lose any chance of thinking about Canetti's presumption that he and Babel wouldn't have been talking about the same thing, of wondering whether Schulz's poetry was lost or retained or found in translation. We lose any sense of how chancy a thing it is to read in translation in the first place, of how contingent and individual a thing our reading of an author is likely to be.
Contrast Updike's sentence about Schulz with some sentences of W. H. Auden's about C. P. Cavafy:
. . . I can think of poems which, if Cavafy were unknown to me, I should have written quite differently or perhaps not written at all. Yet I do not know a word of Modern Greek, so that my only access to Cavafy's poetry has been through English and French translations.
This perplexes and a little disturbs me. Like everybody else, I think, who writes poetry, I have always believed that the essential differences between prose and poetry to be that prose that can be translated into another tongue but poetry cannot.
But if it is possible to be poetically influenced by work which one can read only in translation, this belief must be qualified.
There must be some elements in poetry which are separable from their original verbal expression and some which are inseparable . . . . (333)
This seems to me exemplary. The goal in the two passages is the same, the assessment of a writer not read in the writer's language. But Auden makes the process of arriving at that assessment explicit and interesting, and he does that by reasoning from his own experience, his mix of conscious ignorance and stubborn knowledge. He opens up the question that Updike closes down: what can we know, what experience can we have, what judgments can we make when we're passionately reading in translation?
In a world where globalization is both a threat and an opportunity, literary globalization included, we have to figure out how to be genuinely cosmopolitan. And we'll do that better, I think, if we take Auden's mode of reading and judging as our model rather than Updike's, place a higher value on candid narratives of judgment than on Olympian declarations. Or, to return to an earlier scene: we don't want Canetti and Babel to talk about Dostoyevsky without acknowledging what separates them, what makes reading Dostoyevsky in Russian different from reading him in German translation; better not to talk at all, as Canetti believed Babel chose to do. But better still to have the conversation in full cognizance of the barrier, and see what makes it across the barrier, what doesn't, what never could, what might.