World literature: What gets lost in translation?

Abstract

In What is World Literature? (2003) and other influential works David Damrosch suggests repeatedly that world literature “gains in translation”. This article begins by showing that Damrosch gives no convincing account of what this phrase means. It then develops a wider argument that, even if translations may be accomplished literary works in their own right, the very notion of literature — or at least, one important notion of literature — is associated with untranslatability, or what is lost in translation. The losses, it is argued, may be felt or imagined in various dimensions, and reach into the institutional foundations of the study of literature and of foreign languages.

David Damrosch concludes his book What is World Literature? with some answers to that question. World literature, he asserts, “is not a set canon of texts but a mode of reading”, and “is writing that gains in translation” (2003: 281). ¹ If you put those assertions together, the first way to define world literature, or its “mode of reading”, must be: reading in translation. Translation, we might say, is the craft of world literature. But not, of course, in the sense that working in world literature means working as a translator of literature, or having the linguistic competence that would make that possible. Rather, it means studying literature through published translations, or at least, accepting that your students do so.

This essay will reconsider how far this commitment to studying literary texts in translation is compatible with a commitment to literature’s “techniques and crafts of writing”, as advocated by the editors of the present special issue (for good reasons, in my view). When Damrosch reads or teaches a text written in a language he doesn’t know, how can he tell that it has gained in translation? And how can his students tell? Are they relying on translators to tell them? And how easy is it to find translators who would agree that their translations are superior to the texts they have translated? With such questions in mind I will begin by looking sceptically at what Damrosch means by “gains in translation”, and throughout the essay will be trying to offer clarity about notions of translatability and untranslatability. There are, of course, many other definitions of “world literature” besides those offered by Damrosch, and I am not alone in finding flaws in his; ² and if I focus on his definitions, it is because they, along with his work as an anthologist, appear to be among the most influential on — and/or, more importantly for my purposes, most attuned to — “world literature” as an academic practice. (It will be clear already, I hope, that I will use the phrase “world literature” as it is currently most often used, to discuss academic practice: my subject is a “mode of studying” rather than just a “mode of reading” or a vast menu of texts). In exploring literature’s association with untranslatability, I will, among other things, draw on Jacques Derrida’s notion of the institution of literature, which underscores the importance of a particular relation between the reader and the literary text, and on Robert von Hallberg’s descriptions of literary musicality, and I will touch on questions of paraphrasability, polysemy, and the subjectivity of readers’ responses. My argument will be this: paradoxical though it may seem, in order to valorize both literature and the craft of translation one must accept that by definition literature loses in translation. And conversely, to downplay or deny the disadvantages of reading in translation tends to undermine at once the grounds for studying literature in any language, and the grounds for studying foreign languages. The last issue is of particular concern to those of us who work in “language” departments (French, in my case); but I will be arguing too that anyone who is serious about world literature should also be serious about making their students learn foreign languages.

Gains in translation

In debates around translation the notion is frequently metaphorized (“all reading is translation”, and so on), and the boundaries between the metaphorical and the literal — for example, if we talk about the translation of a play into a screenplay — frequently unclear. By contrast, in this essay I am concerned primarily with the down-to-earth meaning that metaphorization sometimes obscures: that is, translation from one language (in the usual sense of language: French, for instance) into another language. That is the primary sense of “translation” at stake for Damrosch too. Although, as we shall see, Damrosch sometimes nuances, in ways that undermine its programmatic quality, the assertion that world literature is writing that gains in translation, the catchphrase crops up elsewhere, notably in the preface to the 6-volume Longman Anthology of World Literature, where Damrosch and his co-editor, David L. Pike, are unremittingly positive about working with translated texts. Translation functions, they say, “not so much by mirroring and reflecting an unchanged meaning, as by refracting it, in a prismatic process that can add new highlights and reveal new facets in a classic text” (Damrosch and Pike, 2009: xxv). Fair enough, but the obvious objection is that one may equally well say that old facets and highlights may be lost or obscured. To tip the balance in favour of gains in translation, Damrosch has two main lines of argument.

The first concerns the usual meaning of the phrase “to gain in translation”, that is, broadly, that the translation of a given text is somehow preferable to the original. On this front, Damrosch’s specific examples of gains are intermittent and anecdotal and less than convincing. I will discuss two of them. In What is World Literature? Damrosch quotes Goethe’s remark that the Latin translation of one of his works, when compared to the original, “seems […] nobler, and as if it had returned to its original form” (2003: 12). ³ But the general precedence of the “original”, here with a possible Platonic resonance, is still taken for granted in Goethe’s remark. Moreover, the “noble” character of Latin for Goethe is linked to the supreme status of the classics in his culture, a status resting on a certain conceptual hierarchy of cultures, classical and national, that sits uncomfortably with the project of world literature as we now understand it. Damrosch also claims that Goethe “read […] translations as readily as originals, even in the case of his own works” (6–7). Goethe once remarked that he no longer liked to read Faust in German, and that through a new French translation he found his own work “‘again fresh, new, and spirited’”. But this does not persuade me that Goethe believed his own work gained in translation, in the relevant sense. If a reader who spoke both German and French had asked Goethe whether to read Faust in German or in French, I doubt he would have recommended the French; more to the point, I doubt anyone, including Damrosch, would recommend the translated French over the original German to any reader with mastery of both languages. I doubt too that anyone has ever thought it would be good to translate the fresh French version into a fresher German version. One could respond that the cult of the author and prejudice against translation are influencing my assumptions here. Perhaps; for now, I am simply inviting readers to consider how far they share those assumptions, and whether Goethe’s remarks constitute an argument against them.

In How to Read World Literature there is a chapter called “Reading in Translation”, in which Damrosch turns to Voltaire’s Candide to draw the same lesson about the gains to be made by reading in translation. The title page of the first edition, published in 1759, reads: Candide, ou l’optimisme, traduit de l’allemand de Mr. le Docteur Ralph: the original French text is presented, in other words, as a translation from a preceding German text by “Dr Ralph”. In the revised, slightly expanded edition of 1761, the title page adds: “With the additions found in the Doctor’s pockets when he died in Minden, in the year of our Lord 1759”. Voltaire was not named as author in an authorized edition of the book until 1768, and the use of the fictional author figure of Dr Ralph, and indeed the fiction the text was translated, should be linked, as Damrosch indicates, to practices of censorship at the time. ⁴ According to Damrosch, through the subtitle, “Voltaire suggests that his ‘translation’ marks an improvement over the original” (2009: 69). But drawing any such suggestion from this subtitle is highly dubious. This is not only because we now know that Voltaire was the author, and that it was a fiction that the French text was translated. In his own time too, neither censors nor other readers necessarily believed Candide was translated, or were necessarily expected to do so by Voltaire. Moreover, even if the translation and the additional material in the doctor’s pockets had been real, the question of whether such “additions” were an improvement is wholly separate from the question about translation.

The other point Damrosch makes about Candide is that the meaning of the word “Galiléenne” in the original French is obscure to most modern readers of French, whereas modern translations give something clearer. (Damrosch notes that it means “low-class Christian”.) In this part of his discussion Damrosch certainly shows the benefits of working with more than one translation. If, however, we are talking about studying texts (which we are), and we agree that students gain by studying several versions of a text, it still has to be better if you can read the original too, if only in the minimal sense that it is one more version. And with regard to “Galiléenne”, Damrosch himself notes that “the crucial shift has occurred over time rather than across languages” (2009: 71), which suggests that in this instance too the issue is not really about the inherent advantages or disadvantages of reading translations. Some translators might feel they should choose an equally obscure word in English; the discussion would then be one about footnoting rather than about translation. One of the wider points here — whether it is desirable to dissipate obscurity of the sort “Galiléenne” now represents — is something I shall discuss further in the final section.

Damrosch’s other line of argument for “gains in translation” is pursued more extensively but not, to my mind, more compellingly. He writes: “works become world literature when they gain on balance in translation, stylistic losses offset by an expansion in depth as they increase their range” (289). So here it is accepted in passing that there may be “stylistic losses”, but these disadvantages are outweighed, it is asserted, by “an expansion in depth” as texts “increase their range”. What this comes down to, as far as I can see, is that translated texts gain readers, and get interpreted in new ways, outside their own linguistic sphere. This is indeed a good thing, if the text is worth reading, but it is an odd way to use the phrase “gains in translation”; to me it seems a bit like describing a book’s argument as wide-ranging on the basis that you had the book in your backpack when you went on holiday. In any case, any writing would “gain in translation” in that broad sense, which therefore cannot help define and delimit “world literature”. Clearly there are interesting things to say about certain books’ afterlives, in the original and in translation, across different cultures, as Damrosch’s own work demonstrates. But I do not see how that amounts to a claim for what Damrosch calls an “expansion in depth” when texts are translated. And none of this addresses the concerns of those who think that literary texts tend to lose in translation, or to be “untranslatable”.

Literature and/as the “untranslatable”

Those concerns, I now want to suggest, are deeply associated with a certain conception of literature. Damrosch, who insists that the texts of world literature must be “read as literature” (6), says he does not have much interest in venturing a firm definition of literature, but in his conclusion he writes:

There is a significant difference between literary language and the various forms of ordinary, denotative language, whose meaning we take to be expressed largely as information. A text is read as literature if we dwell on the beauties of its language, its form, and its themes, and don’t take it as primarily factual in intent. (288)

Here he does offer a definition, then, and one through which we can begin to see why the literary may be associated with “untranslatability”. Damrosch’s definition is close to some well-established conceptions of literature, and of poetry in particular: first, through the emphasis on “form”, which may mean, but is not limited to, demonstrable traits such as those defining a haiku or a sonnet; and second, through the broader notion of a foregrounding of or attention to the signifier. “Attention to the signifier” may refer initially to the efforts of the author, but is simultaneously, and more importantly for the purposes of the definition, a way of describing the reader’s relation to the text. And this relation, prompted and mediated by literary form, creates peculiar effects: on the one hand, effects of meaning; on the other hand, “musical” effects of sound, rhyme and rhythm, effects that are irreducible to meaning. ⁵

Such points are very familiar to all literary critics, including Damrosch. He writes:

[S]ome works are so inextricably connected to their original language and moment that they really cannot be effectively translated at all. Purist views of literary language often take all poetry as “what is lost in translation”, in Robert Frost’s famous phrase, since whatever meaning a new language can convey is irretrievably sundered from the verbal music of the original. “A poem should not mean / But be”, as Archibald Macleish wrote in 1926 in his “Ars Poetica” in lines that convey their own declarative meaning with surprising success. Much poetry, including Frost’s and Macleish’s, has been translated with great effect into many languages. It is more accurate to say that some works are not translatable without substantial loss, and so they remain largely within their local or national context, never achieving an effective life as world literature. (288–9)

This is another occasion where the term “translate” is sliding around. ⁶ If we want to understand what people mean when they say that poetry always loses, or is lost, in translation, the point about the translations of Frost et al. does not help us much. Those who say poetry is untranslatable need not believe that translations of poetry are worthless, or indeed inferior to the original; I am sure many translators of poetry, including the gifted poets who have translated other poets’ work, would agree that poetry is “untranslatable”. The crucial point is the one Damrosch evokes when he talks about the way some works are “connected” to their own language, and, through that language, their readers and their context — which includes their “moment”, and also the particular traditions and conventions of reading and writing with which they are engaged and through which they are formed. Indeed, to talk about the relation of a poem to its language as a “connection” may already misrepresent what poetry is, or why it is valued by those who value it. Many of the specific effects of a poem, in the space or the interaction between form and meaning, really cannot be reproduced in another language. I should reiterate my conviction that the poem a translator creates may itself be beautiful, and a distinguished literary work in its own right; but that translator will inevitably feel that it has been impossible to capture some of the reverberations and other qualities that made it worth trying to translate the poem in the first place. They may add new “facets”; on various levels, things may have been gained; but other things will have been lost. As I implied earlier, perhaps the fundamental problem with Damrosch’s assertion about gains in translation is that it seems to assume arbitrarily that gains outweigh losses. It is hard to imagine translators of poetry ever feeling that their translations are, on balance, and however adept and creative, an improvement on the original; and if a translator ever decides to make an improvement, he or she is no longer translating, but doing something else.

Two further issues arise at this point. The first concerns the relationship between “untranslatability” and translation. We can agree, then, that poetry may stand emblematically for that which is particularly hard to translate, or “impossible” in some sense that is not exactly metaphorical, and also that, nevertheless, tellingly, quite a lot of poetry does, in fact, get translated, and translated well, and published and read in translation. This seems to suggest, contrary to what Damrosch implies at the end of that last quotation, that it is not “translatability” that decides what gets translated. ⁷ Indeed, untranslatability, or the “impossibility” of translation, clearly attracts some translators, and may help make of their translations compelling creative works (that may or may not get published, read, or studied). Besides many translations of poetry, another striking example would be the translation of Perec’s La Disparition, a novel from which the letter e, usually the most common in French, is entirely absent. For a writer such as Perec, the self-imposed lipogrammatic constraint is seen as productive, in the same way that the constraints of rigid verse forms may promote a poet’s creativity and produce innovative effects for the reader. For prospective translators of La Disparition (literally “The Disappearance”), the constraints of translation, on top of the ferocious constraint of the lipogram, must have a similar appeal. ⁸ Besides Gilbert Adair’s acclaimed English translation, whose title, A Void (1994), is already an impressive accomplishment, there are unpublished English translations and, as Wikipedia (n.d.) points out, translations into Dutch, German, Italian, Romanian, Russian, Spanish, Swedish and Turkish. The “untranslatable” is exactly what some people feel the urge to translate; and being untranslatable need not stop a text from travelling, in translation, around the world.

The other issue arising here is how far poetry may be taken as emblematic of literature, rather than as an awkward subcategory that is anomalous because peculiarly resistant to translation. Not all literary translators face the problems posed by rhyme and metre. Nevertheless all of them must, I believe, see literary language as a distinctive form, or use, of language. Although beauty, as Damrosch suggests, is often an important characteristic of literary language, including most verse, it does not define it; we all know that literary language, including language used in poetry, may also be ugly, disconcerting, and so on, and workaday rather than intricate, “poetic”, musical or elevated. At such moments, at the strictly linguistic level, it may be indistinguishable from other language. In Laclos’s epistolary novel Les Liaisons dangereuses (1782/2003), for instance, letters that are poorly written sit alongside others that are elegant examples of linguistic control, and the book generally does a skilful job of “pretending” to consist of a collection of letters, in much the same way that Candide “pretended” to be translated. Again, this is not to say that readers were ever deceived, or were meant to be deceived. The point I am making is that nothing in the language itself can tell you Candide was originally written in French, or that the fictional letters are the work of a single author, and are fictional.

All of this is a reminder that reading a text “as literature” need not mean dwelling on “the beauties of its language, its form, and its themes”, as Damrosch puts it. He is closer to the heart of the matter when he says that a text is read as literature if we don’t take it as primarily factual in intent. And from this point of view, poetry — which readers tend not to turn to for facts — is a particularly clear-cut example of the general rule. That general rule is described by Derrida, in his discussion of what he calls the “institution” of literature, in terms of “a suspended relation to meaning and reference” (1989/1992: 48). His argument is not that the text is cut off from reality — far from it. “Suspension” also suggests threads of attachment; and readers may even have reason to take some literary texts as “factual in intent” to a high degree. The point is that, in principle, the literary text can become detached from reality at any moment (detached, or partially detached, from denotation, reference, information, facts, and from the author’s own identity and views) — and is free to disguise or deny that it is doing so. This suspension of reference not only concerns the text’s relation to the world, then, but has implications for the nature of the author’s responsibility for the text’s meaning. ⁹ When Flaubert was put on trial in 1857 because Madame Bovary (1856/2001) was considered obscene, the prosecutor complained that he could not tell what Flaubert thought (“ce que pense la conscience de l’auteur”), or what he meant; and therein, although the prosecutor hoped, ultimately in vain, to see Flaubert and his amoral sort of literature disgraced, we can see the emergence of the notion of literature invoked by Derrida. ¹⁰ Evidently, this notion is both historically and geographically limited in scope; but it seems to me — and I shall return to this in my final section — that it is fundamental to the sort of academic practice in which people such as Damrosch and I are now engaged as critics and teachers.

The implications of this conception of literature for literature’s relation to untranslatability may become clearer if we make a comparison between novel-writing and historiography, as we could say that historiography is “primarily factual in intent”. (To say this is to simplify, of course, but I think it makes sense for our current purposes.) Historians’ responsibility for what they write is different from that of novelists, and accordingly historians stand in a different relation to the reader. One way to explain this difference is in terms of the possibility of paraphrase. However carefully or beautifully written a particular work of history may be, readers can have a meaningful discussion about whether paraphrasing a particular sentence would make it clearer, that is, would capture the historian’s assumed intended meaning and would clarify it. Such a discussion is possible because readers share a reasonable and necessary assumption that the historian is aiming to say what she or he means. You cannot, or should not, have the same kind of argument about a literary text, because (or insofar as) you cannot or should not make the same assumption about intended meaning. And without such an assumption, one lacks a basis for paraphrase; or at least, for the sort of paraphrase or rewriting that can be assumed to be an improvement.

In the case of Joseph Conrad, for example, I would certainly say that his style could be made to sound more like that of a “native” at moments, more fluent, easier to understand; but because his works are literary, the obscurity, or oddity, or uncertainty of meaning cannot be taken simply as a flaw. When a character in Heart of Darkness says “I did my possible” (Conrad, 1899/2008: 135), anyone reading the text “as literature”, and with confidence in Conrad as a literary author, will be inclined to view the unidiomatic phrase not as a lapse by Conrad from his third language into his second language, but as part of the complex construction of the fictional world, and of Marlow’s idiosyncratic grasp of the world he lives in. In this literary text, the Gallicism (which, incidentally, would be impossible to translate “back” into French) may be interpreted as the character’s own, or as indicating that the character was speaking French (which is to say, one may take the direct speech itself to be translated, and only indirectly referential). The doubt over how it was intended does not damage the text as literature, where it might have damaged a work of history.

So, reading any text “as literature” in the sense invoked by Damrosch entails an idea that all of the text’s stylistic qualities can be made fruitful by the reader. In literature, in other words, the as-it-were demonstrable precision of the language stands in a kind of dialectic with the assumptions about precision that we bring to the text when we read it as literary. All this implies not only that “reading as literature” involves a certain tolerance or valorization of polysemy and openness to interpretation, but also that the text has a certain integrity, a notional unalterability, that poses a fundamental problem for paraphrase, and for translation. The uncertainty of intended meaning, and of imagined reference, and the proliferation of possibilities of interpretation mean there is no firm basis on which to correct, clarify, or improve the text. An aesthetic criterion has started to emerge here: as readers, and more especially as literary critics and teachers, if we have this idea of literature in mind, we assume that every word, at least in the best texts, bears scrutiny, and that every word was chosen for a good reason — or, to be more precise once again about our relation to a projected literary author-figure, could have been chosen for a good reason; and as we read, and as we study the text closely, good literary writing will tend to reinforce that assumption, whereas poor literary or would-be literary writing will undermine it. Prior to any specificities of form and beauties of language, in other words, is a fundamental convention governing “reading as literature”, that means that not just in poetry but in literature in general, every word matters, in all its possible facets. Herein lies another way of explaining why the task of the literary translator is not just difficult but in some meaningful sense impossible. ¹¹

In this respect poetry, particularly verse, again may be the most clear-cut example of the general rule. In his book on lyric in English, von Hallberg writes:

[M]usic and poetry obstruct the circulation of ideas by means of paraphrase, as though language were entirely semantic, and meaning entirely linguistic. Paraphrasable truths are thought to travel well. Musical effects endorse the local: a determinate set of words (not others just like them), in a distinct sequence, at a particular point in a poem seem just. The achievement of musicality paradoxically summons an authority beyond words on behalf of only these words, in this order, at this point. (2008: 229)

Although von Hallberg’s main concerns lie elsewhere, several ideas in this passage resonate with the arguments I have been making about “world literature”: the local (hovering between two meanings here), the unparaphrasable, that which refuses to “travel”, and also the idea that “musicality” is not always achieved, or can be achieved to varying degrees. These ideas, combined with the idea that assumed “precision” becomes an aesthetic criterion, suggest that “the literary” must be thought in two different ways. On the one hand it exists on some kind of sliding scale, such that some texts are “more literary” than others. Some poems are more musical than others (and, through their music, embody and evoke the sense of “justness” that is one of Hallberg’s concerns); and perhaps some historiographical texts are more literary than others in the sense that they are more reliant than others on particular phrasing, metaphors, assonance and so on for their effects. On the other hand, literariness is more like an on/off switch, regulated by the “institution” of literature, outside any author’s or reader’s control. When the switch is on, when a text is read as literature, reference is suspended, and the distinct sequence of words becomes inseparable from the what text means, or is, and does. This second sense, as I have emphasized, is what excludes the historiographical text. None of this implies, I should reiterate, that nothing can be gained in translation, or that literary translations are necessarily inferior to (other) original texts as works of art, and of craft. Rather, it suggests again that our conception and our valorization of literature is tied fundamentally to a certain relationship to the text, a relationship drawing on and necessitating a certain sense of the integrity of the original text (“a determinate set of words (not others just like them), in a distinct sequence”), from which translation must depart.

Von Hallberg’s remarks also point to another crucial aspect of our notion of literature, which may be part of how we would want to define it: its emotive or affective capacities, its ability to act upon us. Those capacities too may be tied to qualities such as beauty, but clearly they can have many other aspects, including effects of “identification” or “alignment”. Perhaps the power and variety of such effects is another measure of the power, and the distinctiveness, of literature compared with other forms of writing, though it may be shared with other forms of art. Damrosch touches on this when he argues that translators who, in order to explain unfamiliar concepts and contexts, weigh a translated poem down with numerous footnotes end up “loading us up with too much information [and making] it hard to experience the poem as literature, turning it instead into an object of study” (158).

Here it is worth thinking a little more about what travels, and what translates. David Bellos (2011), in his recent book on translation, also picks up critically on Robert Frost’s famous idea that “poetry is what is lost in translation”. ¹² And Bellos argues, correctly, that “You could only make a convincing case [for it] if you knew both languages and their poetic traditions sufficiently well to be able to experience the full scope of poetic effects in both of them” (2011: 152). His reaction is the flipside of my first reaction to the idea that world literature gains in translation: how would you know? Bellos’s point is that everything can be translated; that, in Damrosch’s terms, there are good translations of Frost, which themselves are beautiful and moving. Nonetheless, Bellos recognizes his attachment to original foreign texts, in cases where, thanks to his education, they have been accessible to him, and have taken root in his mind. His example is Rilke’s Duino Elegies:

For me, no English translation can have the same weight or familiarity or perfection or mystery — nor can any paraphrase in German […] The emotion that for me and me alone is wrapped up in the opening of Rilke’s Duino Elegies derives from my past, and although I can tell you about it in this roundabout way, you can’t share it directly with me. What can’t be shared can’t be translated, obviously enough. But that doesn’t make the poem untranslatable for anyone else. (2011: 153)

Again, the notion of the translatable is sliding around here. As I have already insisted, when people say that a poem or a sentence or a joke “does not translate”, or “loses in translation” they do not mean no actual translation exists or could exist; they mean that something important about its meaning or effect cannot be recreated in another language. This may happen when, for instance, there is no equivalent pun. But more generally, it means that important connotations, associations and resonances will be lost. In his attachment to Rilke’s German, this is what Bellos feels. What is misleading about his remarks, it seems to me, is the emphasis on, and interpretation of, the singularity of his own reactions. In some respects those reactions are no doubt uniquely personal. But much of what is most intimately personal is not unique, but shared. That insight, applied to and drawn from an infinite variety of experiences, shaped by diverse cultural and historical circumstances, is surely a profound aspect of what literature offers its readers. So, where Bellos says: “For me, no English translation can have the same weight”, we might say: “For pretty much any reader who is deeply familiar with the original German, no English translation can have the same weight”; and where he describes “The emotion that for me and me alone is wrapped up in the opening of Rilke’s Duino Elegies”, we might say: for you and lots of other people. Other people’s emotions will not be identical, but nor will they be utterly dissimilar; it is a particular combination of words, rather than, say, a random bump on the head, that sets off Bellos’s associations; and the power of those words to create associations and mobilize emotion is one of the main reasons why Rilke’s work is still being read.

One of the ways in which a text may lose in translation, then, and may lose as it is distanced in other ways from the author’s own culture, is in a loss of its affective powers, a diminished capacity to prompt and shape a reader’s highly mediated experience of a certain sort of immediacy. I agree with Damrosch, then, that the experience of reading a text where you need lots of footnotes is very different from, and in important ways inferior to, the experience of reading a text that you feel you understand without notes. I also agree with one of the points Bellos is making, that your grasp of a foreign language and its culture needs to be very strong before you gain more than you lose, in terms of immediacy as well as comprehension, by reading in the original. But those studying the foreign language can read and learn from both original and translation; and I think it is slightly odd, in a debate that is all about the objects of study, for a professional critic and teacher to use the idea of an “object of study” as a term of abuse. One of the benefits of reading in the original is that it improves your command of the foreign language; and one of the benefits of studying a language and its literature together, intensively, may be that, as a reader, you can work your way towards greater immediacy (and the paradox is only superficial), as well as greater depth.

Literatures, languages and windows on the world

This leads me towards the points on which I want to conclude, perhaps the most important of which are practical points about language learning. There can be no doubt that Damrosch is familiar with most if not all of the arguments I have been making, and I think he may be sympathetic to many of them. He himself calls for more language learning — more students of a wider variety of foreign languages. Why then is he so insistent on the advantages of translations, and so reluctant to acknowledge fully that there are good reasons as well as bad to insist on the attractions and benefits of reading in the original?

I imagine that his main aim is to counter indirectly the argument that because no one can learn all the world’s languages, world literature is not a feasible project. I think it must be for similar reasons (besides any residual uncertainty he may have about the validity of relying wholly on translations) that Damrosch appears unwilling to reveal the limits of his own linguistic expertise, and the extent of his reliance on fellow scholars and translators. Damrosch gets through How to Read World Literature without a single footnote or endnote (no doubt a policy encouraged by the publisher, but Damrosch is better placed than most of us to challenge such a policy, or to choose to publish in a different series). In What is World Literature? the referencing is not much more extensive, although the primary texts were written in modern German, thirteenth-century German, English, French, Spanish, Chinese, Akkadian (the Mesopotamian language in which The Epic of Gilgamesh was written), Nahuatl (the language of the Aztecs), Serbian, Russian (Damrosch praises Nabokov’s “wonderfully inventive” 1923 translation of Alice in Wonderland), ancient Egyptian (where, discussing a particular inconsistency in the use of hieroglyphic pronouns, he notes that “Egyptian scribes were notoriously casual in their use of pronouns” [155]), and old Norse: he mentions in passing that “Snorri Sturluson’s dynastic saga Heimskringla is a major document in medieval Nordic culture, but it only makes compelling reading if you are fairly knowledgeable about the political history of Norway and Iceland” (289). Are we to assume, from the phrasing and the lack of references, that Damrosch’s knowledge of Old Norse as well as of political history has allowed him to find the text compelling in the original, before making this claim? Perhaps: his range is extremely impressive, and I assume his linguistic abilities are too. Nevertheless, I doubt he (or anyone) has a good enough command of all of these languages to have read all the texts in the original, or to have a strong enough grasp of their original cultural contexts to be able to judge accurately what roles those contexts have played in making them compelling, or not, in English. Vast investments of time and effort are involved in learning even a single foreign language, or studying a particular era of its literature, in real depth. More than once in his work Damrosch calls for literature specialists to work collaboratively; but the most important way most of us in the humanities work collaboratively, especially when writing, is by reading and drawing on each other’s work — which in this instance means works of criticism and history, and must also mean, crucially, translations. It seems only right that students of world literature should acknowledge those whose shoulders they stand on in order to gaze commandingly across centuries and cultures.

Giving due weight, more carefully and more generously, to the translators of literary texts would, in other words, be one obvious and important way to valorize the craft of translation. This is a point that reviewers of translations in para-academic publications such as the Times Literary Supplement (TLS) might also bear in mind more consistently. Within academia, greater formal credit could be given for the work involved in translation — work that involves creativity, and linguistic skill, and also research and scholarship, whether or not it is brought to the surface through footnotes and the like. In practical terms, this would mean, among other things, giving the work of translation greater prestige in promotion and tenure committees, or in research assessment in the UK. All of this would help promote one of the roles that “language” departments should play, for the benefit of students of world literature but also more distant disciplines and general readers: that is, translating and making accessible material from the languages and cultures in which they have expertise.

In all that I have said, then, I have not wanted to imply — though some readers may have assumed the opposite — that no one should study texts in translation, still less that they should not read texts in translation. There is a lot to be said for seeking out texts not originally written in the language(s) one knows, and for studying translations (preferably multiple translations, preferably alongside originals). I have not wanted to deny that texts may gain in translation, in a range of senses, or to create a balance sheet of the losses and gains that texts accrue in translation. Instead, I have tried to show that our very notion of the literary, of what it is to “read a text as literature”, tends to imply that literary texts must by definition be considered to lose in translation, in significant ways (even if they simultaneously gain in other significant ways). It is this sort of notion of literature, I have argued, that is usually at stake when people talk of “untranslatability”; the term is partly, but not wholly, metaphorical, and cannot be countered simply by showing that “untranslatable” texts have been translated with linguistic and aesthetic success.

Those arguments, to recap further, were based on Derrida’s notion of the “suspension of reference”, from which I span out the ideas intended to support that central claim about literature’s association with untranslatability: ideas about what the “suspension of reference” implies for the relationship between the reader and the text; ideas about the perceived integrity, and potential fruitfulness, of whatever specific combination of words constitutes that text; and from there, notions of the unparaphrasability of the literary text, here associated with “untranslatability”. I emphasized too the importance to the definition of literature of its affective capacities, again tied inextricably to a “determinate set of words”, to repeat von Hallberg’s phrase. This is reminiscent of arguments made by Leo Spitzer, and discussed by Emily Apter: Spitzer insisted not only on reading in the original but also on quoting solely in the original, explaining that

since it is my purpose to take the word (and the wording) of the poets seriously, and since the convincingness and rigor of my stylistic conclusions depends entirely upon the minute linguistic detail of the original texts, it was impossible to offer translations. (quoted in Apter, 2004: 105)

As Apter comments: “Spitzer’s practice of non-translation is not an argument against translation per se, but, rather, a bid to make language acquisition a categorical imperative of translatio studii. A profound respect for the foreign-ness of a foreign language […] motivated Spitzer’s plurilingual dogma” (2004: 105).

While some of the arguments I have made about the nature of the literary have been quite abstract, I have wanted to keep in mind all along that, as I indicated right at the start, debates about “world literature” are debates about what we, as critics and teachers, study and teach, and how we do so. With these questions in mind, it matters that “literariness” in the sense of the “suspension of reference” can survive unharmed the process of linguistic translation (which is another reason, alongside the creative merits of many translations, that literary translations can be taught successfully alongside original literary texts), but it also matters that this notion of literature is far from universal. As I indicated when mentioning the Madame Bovary trial, that notion of literature emerged relatively recently, and commands only partial consent even in societies such as post-1960s France where it has become predominant legally (in the sense that any “literary” text is now likely to escape censorship). The Satanic Verses affair, to take another canonical example, provided a sharp reminder of the notion’s fragility: the novel’s qualities as a sophisticated and stimulating piece of fiction were effectively imperceptible, and/or irrelevant, to those approaching it within a certain political framework and/or outside a certain literary culture which, though international and cosmopolitan, remains far from global. Damrosch is right, then, to note that the definition of literature varies from one literary system or culture to another, and in one sense right to be relativistic about what literature is, when he says literature is best “defined pragmatically as whatever texts a given community of readers takes as literature” (14). But at this point there is a risk of inconsistency in relation to different understandings of, or degrees of, relativism. If different communities of readers take different things as literature, that may imply not that one should shy away from definitions, but that, at least for some purposes, you should, and can only, discuss literature as your community of readers takes it. Damrosch’s apparently pragmatic definition is quite precisely a definition arising from a certain community, a definition that would in fact be unacceptable to other communities, including those whose notion of literature (and of, say, the boundary between the literary and the sacred) is less pragmatic. The institution of “world literature”, I have been trying to suggest, and its place within a certain sort of education associated with freedom of thought, is deeply linked to the institution of literature in Derrida’s terms.

On a less abstract level, though not unrelatedly, another prominent part of the attraction of the texts of world literature is that they offer, in Damrosch’s phrase, “windows on the world” (15). ¹³ The phrase captures something important about how those texts are generally approached academically, and also, I think, by most non-academic readers. Readers may be attracted to the unfamiliar, even the disconcerting: which may mean a text from the other side of the world, but may mean texts about subcultures nearer to home, or texts made difficult by their form or style. I touched on this when discussing the translation of “Galiléenne”: a fundamental part of the attraction of the foreign, of that which is unfamiliar by virtue of its geographical or temporal distance from one’s starting point, may be, and perhaps should be, its foreignness. Damrosch argues that to enter world literature, texts must circulate beyond their “linguistic and cultural point of origin” (6); and concedes, when discussing poetry, that “some works are not translatable without substantial loss, and so they remain largely within their local or national context, never achieving an effective life as world literature” (289). (Earlier I argued against the logic of the phrase “and so” in that last sentence.) Not everyone will agree with my argument that literature as such should be understood to suffer loss in translation. But I think it would be harder to disagree that if you are really interested in literature from other cultures, your interest should extend to works that have not been translated (perhaps for economic reasons, as I noted earlier in passing); and you should be interested, and even particularly interested, in that which loses, or has lost, in translation.

All of this returns us to the idea that if you, or we, are serious about world literature, and about the crafts of world literature — about literariness, about translation as the craft of world literature, and about openness to and exploration of cultural difference — then we also need to be serious about language learning. ¹⁴ If part of what you value in the texts of world literature is their ability to offer “windows on the world”, their capacity to capture and communicate worldviews different from your own, then language-learning should be seen as fundamental in allowing you to read and experience the untranslated and the — perhaps — untranslatable. The foreign language should not only be seen as a tool, or a path towards literary and cultural alterity, but also prized as a form of alterity, with its own ability, in connection with literature or outside it, to draw you into different perspectives and ways of making sense. I have tried to sketch out some of the principled reasons for a commitment to reading in the original, but if we have good reasons to teach through translations too, we need more students who can discuss with authority the qualities and contexts of a particular translation, more prospective readers and students who are drawn into the language via the literature and vice versa, and we need more translators and better translators. I want to end, then, simply by calling on my readers to do something very difficult: promote the study of languages in your institution, and in your discipline, if your discipline is, or is becoming, world literature. In practice I think this means introducing a compulsory language-learning element into your programmes, and not, or not just, a series of taster courses, which might look attractively global, but the possibility and indeed the obligation for students to study at least one language to a high standard. Spending time on language learning is bound to take some time away from literary study as such, initially at least, and is hard work, and is likely to be resisted for those reasons by some students as well as by monolingual staff and managers. But without that commitment, world literature not only risks intellectual self-defeat, for the reasons I have tried to elucidate, and risks slipping into what Damrosch rightly condemns as a “high-minded amateurism” lacking “any real engagement with the works’ cultures of origin” (290), but also risks blending smoothly into the other forces undermining the language and literature programmes we already have.