Is That a Fish in Your Ear?

TWENTY-SIX Style and Translation

Translations typically alter numerous features of the source in order to produce matches for those of its dimensions that count in the context it has. But there is one traditionally perceived quality of written and spoken language that is identified not with any particular dimension of an utterance but with the overall relationship between them—its style.

Style is more than genre. Kitchen recipes are typically translated not into something as vague and undifferentiated as “English” but into “kitchen recipese,” the genre constituted by the conventional features that kitchen recipes have in our tongue.

In like manner, you don’t translate French poetry into “English” but into poetry, as the American poet and translator C. K. Williams insists. Poetry is a characteristic social and cultural use of language and can therefore count as a genre in our sense, but it comes in many different forms. Beyond the genre, a poetry translator has to choose the particular style that he is going to use.

Twenty years ago, Eliot Weinberger and Octavio Paz brought out a curious essay-cum-anthology titled Nineteen Ways of Looking at Wang Wei—nineteen different English translations of a poem by a Chinese poet of the eighth century C.E., . Setting aside all their arguments about which of these “ways of Wei” is to be preferred, what is quite obvious is that they represent nineteen different ways of writing poetry in English, nineteen “styles” of fairly recognizable kinds (Eliot-ish, Ashbery-ish, free verse–ish, and so forth). Ten years later, Hiroaki Sato brought out One Hundred Frogs, a compilation of actually rather more than a hundred already published English versions of a famous haiku by Matsuo Bashō:

Furu ike ya

kawazu tobikomu

mizu no oto

I

II

III

If “style” is the term that names the principal means of distinguishing the differences among these three versions of Bashō’s haiku, then it means something that is not an individual property of, say, the poetry of Allen Ginsberg, John Masefield, and Ogden Nash but a collective property of poetry written in that style—in Ginsberg-ish, Masefield-ish, and Nash-ish, so to speak (one of them was written by Ginsberg, in fact). Style in this sense is eminently imitable, and not just for comic effect. Students of musical composition develop their skills by writing in the manner of Mozart or Bach, and writers also practice at writing like Flaubert, [160] or writing like Proust.[161] The following pieces are not by William Wordsworth, T. S. Eliot, or J. D. Salinger—but it does not take much more than vague memories of school to know which among them is Eliot-ish, Salinger-ish, and Lake Poet–ish, respectively:

and

and

These examples could lead us to believe that the translation of style is an exercise in pastiche, the translator’s task being the choice of an existing style in the target culture to serve as a rough match for the “other.” Many literary translators go about their job in just that way. On reading a new work in French, for example, I certainly do run through in my mind the kinds of English style that might fit, and when starting on a new job, I often rifle through the books on my shelf to remind myself of the particularities of the “style match” I have in my head. But this idea of style as a culturally constituted set of linguistic resources characteristic of an author, period, literary genre, or school clashes with another widespread idea of what a “style” is: the irreducible difference of any individual’s unique forms of language. In brief: If style is “inimitable,” how come it can be imitated?

The muddle about what style is began in the gilded halls of the Académie Française, an institution set up by Louis XIV to promote and defend the French language. In 1753, a natural scientist was invited to take his place as one of the forty “immortals,” as members are called. Georges-Louis Leclerc, Comte de Buffon, an eminent botanist, mathematician, and natural historian, gave an extraordinary acceptance speech that has since become known as the “Discourse on Style.” In it he sought to reassure his audience—the thirty-nine academicians who had just elected him—that the promotion of a mere scientist to such elevated rank would not topple rhetoric from its proper place at the pinnacle of French culture. He may even have been sincere—but I wouldn’t count on it. In his much-quoted but mostly misunderstood conclusion, Buffon emphasized that what matter above all are the arts of language. Scientific discoveries, he declared, are really quite easy to make, and will quickly perish unless they are explained with elegance and grace. That is because mere facts are not human achievements—they belong to the natural word and are therefore hors de l’homme, “outside of humankind.” Eloquence, by contrast, is the highest evidence of human agency and genius: le style est l’homme même.

This meaning of style, as a synonym for elegance and distinction, continues to motivate most modern uses of the word and its cognates. Stylish clothes are those considered elegant by some group of people; to ski or to dance or to serve cucumber sandwiches in style is likewise to do these things with fashionable grace. Buffon’s style is a social value. Nobody is free to construct his or her own idea of what is stylish, save by getting other people to agree. Similarly, stylish writing conforms to a shared notion, however vague, of what is fashionable, appropriate, socially elevated, and so on in the way you speak and write.

Matching posh for posh in translating between languages used by cultures with linguistic forms that correspond to hierarchical social structures is no sweat. Where the social structures of the source culture are more elaborate than those of the target, a degree of flattening occurs: the different social implications of Estimado señor and Apreciado señor at the start of a formal letter in Spanish, for example, can’t be represented in English, which can say only “Dear Sir.” To compensate for losses of this kind, which can be far more substantial when translating between cultures as unrelated to each other as Japanese and French, for example, the translator may invent target-language analogues for distinctions that belong to the social world of the original, and be accused variously of quaintness, condescension, or fidelity to the source. But there are even less tractable issues involved when the social register of the language used in the source is low. There is a seemingly inevitable bias against representing forms of language recognized in the source culture as regional, uncouth, ill-educated, or taboo by socially matching forms in the target tongue—presumably because doing so risks identifying the translator as a member of just such a marginal or subordinated class. As a result, translation usually takes the social register of the source up a notch or two. The social dimension of “style” doesn’t flow easily from tongue to tongue.

The novelist Adam Thirlwell has argued that the meaning of the word style changed in 1857.[162] In the convincing story he tells, style flipped over, almost in one go, from being a description of the elegance of a whole manner of expression to being about just one subelement in the composition of prose—the sentence. The culprits for this radical reduction of style were Gustave Flaubert, his novel Madame Bovary, and the many comments Flaubert made about sentences in his partly teasing letters to his girlfriend, Louise Colet. Since 1857 or thereabouts, Thirlwell argues, critics and readers have needlessly restricted their idea of a writer’s style to those low-level features of grammar and prosody that can be exhaustively identified between a capital letter and a period. Henri Godin, writing about “the stylistic resources of French” just after the Second World War, was quite certain that style and syntax are the same thing and reach their point of perfect harmony in the writing of … Flaubert.[163]

Because the grammatical forms, the sounds of individual words, and the characteristic voice rhythms of any two languages do not match (if they did we would call them the same language), the “Flaubert shift” made style instantly untranslatable. Thirlwell’s main aim is to show that this is nonsense—and that the novel is a truly international and translinguistic form of art.

At some point in the course of the nineteenth century, the idea of style as “the aesthetics of the sentence” got thoroughly muddled up with a completely different tradition that came to France and Britain from German universities. Scholars in departments of Romance philology tended to justify the attention they paid to canonical writers on the grounds that their works represented special, innovative uses of language, distinct from the norms of the speech community, and were therefore important factors in the course of linguistic change. Poets, they argued, were not simply users of language but the creators of it; a language was not a smooth and rounded whole but a gnarled old potato marked by bumps and dents that speak the history of its creation. “Style research,” or Stilistik, pursued with fervor for a hundred years, and reaching its brilliant peak in the essays of Leo Spitzer (1887–1960), was an exciting but quite circular pursuit: the language of a “great work” becomes a fine-grained map of the ineffable individuality of some great writer’s “self”; but the “self” or the essence of, let us say, Racine is entirely constituted by what can be mapped through his language, subjected to a particular kind of analysis of his style. Style in this sense is inimitable by definition—that’s the point of it. And if it can’t be imitated in the same language, it’s not even worth trying to translate it.

But it isn’t true. Most of the features of language use that Spitzer identified as significant aspects of Racine’s “self,” for example, can also be found in the language of Racine’s contemporaries writing in the same literary genres. Yet the remarkable tenacity of the philologists’ principle that every great writer has a manner that is unique and inimitable led people to reinvent the very history of the idea of “style.” They went back to Buffon’s famous “Discourse,” took his maxim that le style c’est l’homme même (“style is what makes us human”), lopped off the last word, and recycled the remainder—le style, c’est l’homme—so as to prove that “the style is the man.” As the noted Oxford scholar R. A. Sayce put it in his 1953 study Style in French Prose, “details of style … reveal the deeper intentions and characteristics of a writer, and they must be dictated by some inner reason.”[164]

“Style” thus has a very curious history. A sentence uttered in 1753 as a defense of literary eloquence came to be touted around as a pithy formulation of the idea that no two people speak or write in exactly the same way because no two speakers are the same person.

It’s indisputable that every speaker of any language has an idiolect, a characteristic set of (ir)regularities that is not identical to the usage of any other person. Why this should be so is discussed see here of this book, but it should be obvious that there are no intellectual, psychological, or practical obstacles to speaking in the same way as some other person (impersonators and pasticheurs do it all the time). But the fact of linguistic variation at the individual level has some very practical applications—such as catching out forgers. Among the early applications of computers to the humanities were statistical programs for identifying the authorship of suspect documents. The programs themselves rested on rival theories about what “style” was: typical patterns in individuals’ use of verbs, or vocabulary, or other parts of speech, that were unfalsifiable by anyone else; or else that “rare pairs” (two words occurring typically together) could be used to identify and distinguish different authors; or that the position in the sentence of common words was what gives the identity of the writer away. This last guess was called “positional stylometry” and was developed in the 1970s by A. Q. Morton and Sidney Michaelson at Edinburgh University. Results of their computer program were admitted as evidence in court in many cases and also used to make scholarly hypotheses about the provenance of different parts of the Hebrew Bible.

“Style” in this individual sense cannot possibly be the object of translation. It would make no sense to try to simulate in English the statistically irregular positioning of, say, the negative particle pas in some French original.

Two interesting consequences ensue. If “style” is such an individual attribute that it cannot even be controlled by the writer (thus allowing sleuths to catch forgers out), then every translator has a “style” of that kind in his target language, and the style of all his translations must be more like itself than it can ever be like the style of the authors translated. I often wonder, in fact, whether my English versions of Georges Perec, Ismail Kadare, Fred Vargas, Romain Gary, and Hélène Berr—whose characteristic uses of French are manifestly quite different—are all, stylistically speaking, just examples of Bellos. By some accounts, they have to be: computational stylistics gives no quarter on that score. Secretly, though, I am quite happy that it should be so. After all, those translations are my work. But it will be known for sure only by some large computer program.

All the same, style can’t be swept away just like that. Admittedly, we do not mean “elegance,” as Buffon did, when we talk about literature and translation, even if we still do when we talk about clothes or cucumber sandwiches. We do not mean statistical regularities in the way we place the indefinite article, though we do when we gratefully accept a court ruling on the incompatibility of the style of our uncle’s alleged will with its claimed authorship.

We mean something else, not so difficult to express: “style” is the reason a novel by Dickens is just Dickens’s, why a piece of P. G. Wodehouse—even if it were written by somebody else—is still in its essence a piece of Wodehouse. Style is, if not the man, then the thing! It is what makes any work uniquely itself.

I also know a Dickens when I see one. But that’s trivial. The question is: At what level is the Dickensianity of any text by Dickens located? In the words, the sentences, the paragraphs, the digressions, the anecdotes, the construction of character, or the plot? Because I, translator, can give you the plot, the characters, the anecdotes, and the digressions; I can even give you the paragraphs, and most of the time I can give you a fair approximation to the sentences, too. But I cannot give you the words. For that, you have to learn English.

For Thirlwell, novelistic “style” is the name of a holistic entity that comes somewhere between “a writer’s special way of looking at the world” and “a writer’s own way of writing novels.” Characteristic uses of sentence structures and sound patterns are certainly a part of the latter, and maybe of the former, too—but only a part. Style in Thirlwell’s sense—the most usable and purposeful sense—is something much larger. If it were not, it would disappear in translation. The circulation of novels among all the vehicular languages of the world and their incontestable conversations with one another demonstrate without a shadow of doubt that style does survive translation. The means that translators use to ensure this are no more than the common skills used in all translation tasks.

In sum, the widespread notion that style is untranslatable is just a variant of the folkish nostrum that a translation is no substitute for the original. There is no more truth to it than there is in the idea that humor can’t be preserved by rephrasing in the same or another tongue.

There is a difference between translating jokes and translating style, however. The first is typically done by concentrated effort; the second is better done by taking a slight distance from the text and allowing its underlying patterns to emerge by their own force in the process of rewriting in a second tongue. What they have in common is this: finding a match for a joke and a match for a style are both instances of a more general ability that may best be called a pattern-matching skill.

We’re still short of an answer to the question of what we mean by “match,” but we’re getting closer.