At the Frankfurt Book Fair, where thousands of publishers from around the world annually squeeze inside huge exhibition halls to buy and sell foreign rights, a typical booth contains several tiny tables at which agents and editors cut their deals. Stopping to peruse what books are on offer, you wouldn’t have to strain to get wind of potential translations into some language or other. I recently attended the event and overheard a conversation that was startling and not a little worrying.

At the booth of an Italian trade publisher, a British representative was handed a new publication and smiling said, “The cover looks smart.” The Italian rights manager asked, “Do you mean ‘clever’?” Her prospective client paused, a quizzical expression on his face, before responding, “Sure,” as if to quash any doubt that they were on the same… page? I myself wasn’t sure they were speaking the same language, even though it was English. “Smart” can mean “clever” and more, of course, but in this context the British speaker was probably referring to appearance, not intelligence. What would happen, I wondered, when these two started discussing the book’s actual content? Could it be the same book in their words? Their minds? Given the substantial amounts of money that normally change hands in Frankfurt, you might expect a greater sense of mutual understanding to accompany any financial dealings. The conversation hardly inspired confidence about the current state of literary translation.

As an English-language translator who has worked for a wide range of publishers over the past thirty years, but who is also inclined to take a theoretical approach to his work, I have inevitably become a critical observer of the ways in which foreign texts are chosen for translation. Publishing practices that date back to the beginning of the twentieth century have decisively shaped our intercultural exchanges and have contributed, I believe, to the appallingly low volume of Anglophone translation since the Second World Warónow just over 2% of total annual book output according to industry statistics. These practices are questionable not merely because they have admitted relatively few foreign texts from a narrow range of foreign cultures, but because they have formed aggressively monolingual readerships in the US and the UK, generally uninterested in translations. Ironically, then, publishers seem not to have acted in their own best interests, whether those interests are cultural or commercial or both.

The main problem is that long-standing practices reveal a conceptual naïveté, a limited understanding of translation, of the cultural issues that any translation project must confront and somehow resolve if it is to be successful both critically and commercially. Hence I want to suggest that, where translation is concerned, these practices need to be rethought, if not simply abandoned, and replaced by a more savvy approach that is truly concerned with cultural as well as commercial factors. Otherwise Anglophone publishersóand no doubt publishers in other languages as wellówill remain complicit in our present predicament: the absence of what I shall call a translation culture, that is to say, a culture that can sustain the study and practice of translation, that can foster a sophisticated and appreciative discourse about translation in its many aspects, and that can create an informed readership to support and encourage the publication of translated texts.

Early in the twentieth century, a largely unwritten policy came to prevail among Anglophone publishers. Buy the translation rights to a single book by a foreign author. If soon after publication the translation suffers a substantial loss or fails to earn back its production costs or to realize a modest profit, then stop publishing translations of the author’s books. If, however, the first translation manages to break even or to approach a break-even point, then continue to publish translations of that particular author in the hope that more translations will create a readership and add profitable titles to the backlist.

What I am describing as a prevalent policy dictated a print run that turned into an industry standard by the 1940s. In 1949 the translator F.H. Lyon proposed that Stanley Unwin publish an English version of Finnish writer Jarl Hemmer’s novel A Fool of Faith, but Unwin declined on the basis of a sales projection. “We need to be sure of selling the greater part of 5,000 copies,” he wrote back, “if we are not to lose money on a novel today, and we have no confidence that we could achieve this result with Hemmer’s work.” Sales in the range of 5,000 copies became a benchmark for a successful translation of a foreign novel. Yet the figure also came to reflect the sad reality of publishing translations in English. In 2002 Christopher MacLehose, formerly director of the Harvill Press, observed that “for the most part now the majority of even the finest books that are translated find their way to sales between 1,500 and 6,000.”

Over the past hundred years few English-language translations have managed to reach that upper limit. As a result, most foreign authors who have had a book translated into English have not been translated again, either by the initial publisher or by others, who were scared off by the poor market performance of the first translation. A recent case in point is the Italian novelist Paola Capriolo, a prolific, award-winning author of more than fifteen books since 1988. Although her inventive narratives have been translated into some twelve languages, US and UK publishers have avoided them because the first translations have not sold well. Serpent’s Tail brought three of her novels into English: Floria Tosca (1997), The Woman Watching (1998), and A Man of Character (2000). But diminishing sales led publisher Pete Ayrton to abandon her. She has yet to find another trade publisher in English.

Ayrton, to his credit, was willing to support Capriolo’s work through three translations. Although he was then running a mid-sized press, small compared to the multinational conglomerates, like them he was implementing a practice that is a century old. Focusing on a single book by a foreign author or on a single author from a foreign language forces publishers to decide whether to continue translating that author in order to build a readershipówhich is only a readership for that author and may carry no implications for other books translated from that foreign language or others. Given this practice, it is more than probable that a publisher, whether large or small, will abandon an author that has done poorly on the book market.

The exceptional cases are remarkable because they involve the great works of modern literature. In translation these works were commercial failures initially, according to the standards in place then and now, and it is only because some of the publishers involved were willing to add the titles to their backlists or to sell off reprint rights that the translations achieved canonical status in the US and the UK. In 1922 Chatto and Windus published C.K. Scott Moncrieff’s version of Proust’s Swann’s Way in two volumes, and within a year 3000 copies were in print. Yet five years later volume one had sold only 1773 copies and volume two only 1663. In 1928 Martin Secker published his first translation of a novel by Thomas Mann, Helen Lowe-Porter’s version of The Magic Mountain, but it took seven years to sell 4,641 copies, helped no doubt by the translations of seven other books by Mann that Secker had issued in the interval. In 1929 the Hogarth Press published Beryl de Zoete’s version of Italo Svevo’s novella The Hoax,, but after selling 500 copies in the first year the book showed a loss, and publisher Leonard Woolf was soon looking to remainder 300 copies. In 1930 Woolf also published Svevo’s collection of stories, The Nice Old Man and the Pretty Girl, which met the same fate. He attempted to sell the translation of the stories to Alfred Knopf, who had published Svevo’s Confessions of Zeno in 1930. But the editor at Knopf declined. “I am afraid there is no question,” he replied, “but that he has been a failure, although we made immense efforts to put him across.”

I take these “immense efforts” to be promotion and marketing schemes like running advertisements, sending out review copies to newspapers and magazines, and contacting editors and reviewers in order to generate some favorable publicity. But any such efforts cannot address a crucial fact about the translation process: it so radically decontextualizes the foreign text that a translation can be hard for a reader to appreciate on its own, even a group of translations from the same foreign author.

The structural differences between languages, including those that bear significant lexical and syntactical resemblances, require the translator variously to dismantle, rearrange, and finally displace the linguistic features of the source text. Three contexts are lost. The first is intratextual and therefore constitutive of the source text, its verbal texture. The second is intertextual yet equally constitutive, since it comprises the network of relations and allusions that endows the source text with significance for readers who have read widely in the source language. The third, which is also constitutive but both intertextual and intersemiotic, is the context of reception, the various print and electronic media through which the source text continues to accrue significance when it begins to circulate in its own culture, ranging from book jackets and advertisements to periodical reviews and academic criticism to television interviews and internet forums. By “constitutive” I mean that the three contexts are necessary for the signifying process of the source text, for its capacity to support meanings, values, and functions in the source culture. These contexts just don’t travel with a translation, especially of a single book or author. And so no text ever survives intact the transition to a different language and culture.

Consequently, a reader of a translation is unable to experience it with a response that is equivalent or even comparable to the response with which the foreign reader experiences the foreign text. Entire literary traditions, even entire literary canons are never translated into a particular language, certainly not into English. And rarely is a substantial and diverse selection of contemporary works in print at any one time, regardless of how many publishers invest in translations from a globally dominant language like English. No wonder, then, that when confronted with a translation readers automatically fall back on what they do know and prefer: they read and evaluate the translation mainly against linguistic patterns, literary traditions, and cultural values in the receiving situation, which is usually their own culture.

Yet this sort of reception is risky. It can invite a dismissal of a foreign text that may not be recognizable or intelligible in relation to texts and traditions in the translating language. It can invite a complacent reaffirmation of the reader’s cultural values and an ethnocentric rejection of a foreign culture merely because the foreign text cannot be understood in its own terms.

This can happen even with elite readerships who are capable of dealing with challenging texts. While Proust and Svevo initially failed to appeal to English-language readers, Virginia Woolf did appeal and in spades: most of her books met with immediate success upon publication. In 1925 the Hogarth Press published Mrs. Dalloway, and it sold 2136 copies in its first year alone, earning a profit of £150, which is quite impressive for a modernist experiment that attracted a fairly specialized audience. (The profit may seem small today, but it needs to be situated in its moment: in 1932 the Kafka translators Willa and Edwin Muir rented a house in the trendy London neighborhood of Hampstead for an annual cost of £120.) Mrs. Dalloway was so successful, I suggest, because enough readers could understand and appreciate its experimental narrative. By 1923 Dorothy Richardson had published seven novels distinguished by a similar stream-of-consciousness technique. In 1919 excerpts from James Joyce’s Ulysses had appeared in the magazine The Egoist, with the full text published in Paris three years later. English-language readers keen on modernist fiction had a cultural context in which to read Virginia Woolf’s novels, as well as those of Richardson and Joyce. But these readers weren’t crossing over to foreign modernisms.

What context could have enabled the first translations of Svevo’s writing to find an audience? Joyce’s admiration of it obviously did little to improve sales. The Italian fiction then available in translation was dominated by varieties of naturalism, most of which were worlds apart from Svevo’s wry, psychoanalytically inflected tales of anti-heroes. The most popular Italian writer was Matilde Serao, who explored themes like illicit romance among aristocratic and middle-class women, the uneven cultural and economic development between North and South, and the provinciality and poverty of her native Naples. In 1901 Henry James wrote admiringly about the “remarkable spontaneity” of her early fiction, particularly The Conquest of Rome, in which a Neapolitan deputy takes up his seat in parliament only to fall helplessly in love with a sophisticated Roman woman and ultimately lose his office. A succession of translations followed James’s article, so that by 1928 ten of Serao’s books had been Englished, even though her writing grew increasingly melodramatic. In a review for the Times Literary Supplement Lacy Collison-Morely, the literary critic and historian who would translate Svevo’s collection of stories for the Hogarth Press, called Serao’s novel Souls Divided (1919) “a positive orgy of feeling.”

Nonetheless Serao had captured the imaginations of Anglophone readers. Would British and American publishers have challenged the taste for her fiction if they had paid greater attention to her contemporaries who had revised or abandoned naturalism, writers like Luigi Pirandello and Federigo Tozzi? Both wrote realistic narratives but devised innovative strategies to probe character psychology, Pirandello resorting to moments of epiphany and irony, Tozzi crafting discontinuous plots shot through with ominous tension. Readers immersed in the work of these writers might well have gravitated to Svevo’s experiments too. They would at least have been given the opportunity to sample different kinds of Italian fiction, glimpsing what the Italian reader had available. Several of Pirandello’s novels were translated, but they received mixed and in some cases dismissive reviews. Tozzi’s novel Three Crosses appeared in English in 1921, but it was assimilated to the prevailing taste for naturalism. The New York-based Literary Review found that it “grips the attention with the force of destiny.” Neither Pirandello’s nor Tozzi’s stories had yet been translated, but it was in their stories that they developed their most intriguing narrative innovations.

To be sure, various factors play into the reception of any book, in addition to imponderables which guarantee that any prediction of success or failure can never be certain. In the case of translations, however, past practices show quite clearly that publishers have not sufficiently taken into account the decontextualizing process of translating and its adverse impact on the reception of foreign texts. Focusing on a single foreign text or a single foreign author winds up exacerbating this process: it mystifies the loss or sheer destruction of the foreign linguistic and cultural contexts and therefore gives the false and misleading impression that any literary work can be understood on its own. This encourages an essentially romantic notion of original genius that militates against the contextualized reading, the implicit comparisons among texts, which informed readers always do. To enable English-language readers to understand and appreciate a translation, publishers must restore in English at least part of the context in which the foreign text was written. With individual publishers each pursuing their own single-minded focus, this context is unlikely to emerge.

Hence where translations are concerned publishers must rethink their identities as publishers. “Book publishing,” Stanley Unwin wrote, “is a very personal business” in the sense that “a publisher’s own inclinations determine the selection of the manuscripts chosen for publication.” He admitted that these inclinations could be cultural or commercial, but he nonetheless considered them personal.

I am suggesting that with translations publishers must take an approach that is much more critically detached, more theoretically astute as well as aesthetically sensitive. They must publish not only translations of foreign texts and authors that conform to their own tastes, but more than one foreign text and more than one foreign author, and they must make strategic choices so as to sketch the cultural situations and traditions that enable a particular text to be significant in its own culture. Translators too need to participate in these choices, since their expertise is invaluable in assessing the losses and gains in the translation process. But they must regard translation in more self-critical ways than is generally the rule today, when translators tend to take a belletristic approach to their work, making impressionistic comments which show that they, like publishers, find writing to be primarily personal, a form of self-expression or a testimony of their aesthetic kinship to the foreign author. Publishers and translators alike need to depersonalize translation and to become aware of the ethical responsibility involved in representing foreign texts and cultures. What a sad time it is for intercultural exchange when publishers and translators look abroad and see mainly opportunities to imprint their own values.

The initiative I am recommending cannot be pursued by one publisher alone without a significant outlay of capital and probably not without the funding and advice of a cultural ministry or institute in a foreign country. But publishers can coordinate their efforts, banding together to select a range of texts from a foreign culture and to publish translations of them. This sort of investment cannot insure critical and commercial success. But in the long run chances are that it will pay off handsomely by laying the foundations for an informed readership that will not feel inadequate before translations from a particular foreign language and will actually be eager to sample new texts from it. Readers as well as publishers have much to gain from a translation policy that is based on an incisive understanding of the translation process.

This essay was prepared for the panel “To Be Translated or Not To Be” organized at the 2007 Frankfurt Book Fair by the Institut Ramon Llull in connection with the PEN/IRL report on the international situation of literary translation. Figures for publications by the Hogarth Press, Chatto and Windus, and Martin Secker Ltd are drawn from the archives at the University of Reading and used with the kind permission of the Society of Authors and the Random House Group Ltd.