Navigation – Plan du site
Les apports paradoxaux

When loanwords are loaned back: the conundrum of translating French into French

Robert Byrnes
p. 249-259


Dans Ulysse, le roman de James Joyce, Léopold Bloom est le narrateur de l’épisode « Eumée ». Il y emploie trente-sept mots empruntés au français, et indiqués en italique dans le texte anglais. Le but de Bloom est de donner à son discours un style précis et cérémonieux. Il souhaite éblouir, par de tels effets de style, son jeune interlocuteur, Stephen Dedalus, récemment diplômé de University College. Comment les deux traductions françaises d’Ulysse représentent-elles ces emprunts ? Dans la première traduction française (Gallimard, 1929), le traducteur, Auguste Morel, supprime généralement l’italique, laissant ainsi les emprunts se fondre dans le texte français et perdant « l’effet d’emprunt » de l’original. Pascal Bataillard, qui a retraduit « Eumée » pour la version française dirigée par Jacques Aubert (Gallimard, 2004), s’essaie à diverses stratégies pour préserver cet « effet d’emprunt ». D’une part, il restaure l’italique, d’autre part il crée un « effet miroir » en insérant des emprunts à l’anglais dans le texte français. L’objet de cet article est de comparer, grâce à de nombreux exemples, les stratégies de Morel à celles de Bataillard.

Haut de page

Texte intégral

1Robert Byrnes, département de Rhetoric, Writing and Communication, est responsable du séminaire Joyce à l’université de Winnipeg, dans le Manitoba, Canada, et enseigne également l’histoire des médias, le nouveau journalisme, et l’écriture. Ses recherches actuelles portent sur la traduction, la phraséologie et l’analyse textuelle assistée par ordinateur. Ses publications sur James Joyce comprennent : « A Statistical Analysis of the “Eumaeus” Phrasemes in James Joyce’s Ulysses », JADT 2010: Statistical Analysis of Textual Data, vol. 1, 2010 ; « Weiningerian Sex Comedy: Jewish Sexual Types behind Molly and Leopold Bloom », James Joyce Quarterly, vol. 34, Spring, 1997 ; « Bloom’s Sexual Tropes: Stigmata of the “Degenerate Jew” », James Joyce Quarterly, vol. 27, Winter 1990. Il a également publié « What’s Wrong with WAC Anthologies? » dans le Journal of Teaching Writing (co-écrit avec Brian Turner), vol. 14, 1995.

2The scintillating “Eumaeus” episode of James Joyce’ Ulysses runs on pure verbal fun. The voice of Leopold Bloom, the Odysseus figure, has taken over the narrative, and it speaks almost entirely in clichés, usually dozens per page, fifty-four at a maximum on one page in the Penguin Classics edition, fourteen at a minimum, and this goes on for seventy-two pages, almost a tenth of the whole novel, and to almost two thousand eight hundred clichés―a rhetorical performance, no doubt, in the still oratorically-inclined Dublin of 1904, but an affecting one too, as Bloom is addressing Stephen Daedalus, the Telemachus figure, for whom he has paternal feelings―he will soon invite him home to take up residence there, a replacement for his own long-dead son Rudy. In “Eumaeus” Bloom is auditioning for a paternal role in Stephen’s life, pouring out Polonius-like wisdom to the drunken, stuporous, virtually silent Stephen, his captive auditor, as they sip coffee in a cabman’s shelter.

3The primary comedy of the episode comes from Bloom’s endless reels of cliché. Bloom is an autodidact, a miscellaneous reader, a good man without education who aspires, as he might say for himself, to liberal sympathies, cosmopolitan perspectives, and lofty language. And he hopes to impress the Jesuit-educated Stephen. The trouble is that he has nothing in his discursive repertoire but idées reçues and formula phrases: he speaks from a thesaurus of petty bourgeois platitudes, from a phrase-hoard of the popular, the ready-made, and the passé. And this is only the beginning of his problems, in that more comedy and a certain pathos in “Eumaeus” arise from Bloom’s “incongruities of register”, as Christine O’Neill has called them (O’Neill, 1996: 48)―his tonal mismanagement, his desperate lunges up into formal or ceremonious register, after which he collapses back into the neutral register he has struggled to surmount, or falls further, into slang and occasional vulgarity. Bloom’s essential character emerges from just these sallies, as he reaches for effects beyond his grasp. From phrase to phrase, things are liable to shift to a lesser or greater degree, the whole Joycean effect being one of gentle comedy and pathos at the expense of his generous-hearted, overreaching hero.

4Translators must love clichés, and love registers, to take on “Eumaeus”, and if they do, they will have two thousand eight hundred moments of sheer delight. But there is a specter haunting this hortus deliciarum, the problem of the “French loanword effect”. There are thirty-seven French loanwords in the “Eumaeus” episode of James Joyce’s Ulysses, all of them italicized, and the very idea of translating them back into French, for a French edition of the work, seems paradoxical and doomed. If a translator simply re-imports them, in italics, into a new French translation, they will be noticeable, and the reader may well surmise they were loanwords, but they will not be loanwords any longer, and they will never produce the effect they do in the English text; and if one imports them without the italics, they will disappear absolutely, having no effect whatsoever. Could a translator substitute loanwords from some other language, to at least preserve a “loanword effect” in the French text? No, because Joyce, and his narrator Leopold Bloom, were exploiting the prestige of French in 1904, long before the 1919 Treaty of Versailles, when it was still the global diplomatic language, when Balzac, Baudelaire and Flaubert were preeminent in world literature, and when every liberally educated person in Britain aspired to a smattering of French. French attested, as Bloom would say, to their widened horizons and savoir-vivre. French loanwords were magical tokens, power words from a perhaps superior civilization. No other loanwords would do.

5There are other problems. We must remember that all of Bloom’s French loanwords are clichés in English―they were loaned in the first place because they were so useful, and quickly became commonplace. They participate fully in the tissue of cliché that is “Eumaeus.” But only nine of these thirty-seven French expressions are clichés in French (sur le qui-vive, sur le tapis, le coup d’œil, de rigueur, être au fait, lever la séance, par excellence, procurer ses entrées, profitant de ce contretemps). All the rest, like sang-froid, soi-disant or penchant, are absolutely ordinary, neutral expressions in their native environment. A translator who would simply bring all thirty-seven loanwords back, unchanged, into a French version of “Eumaeus” would lose twenty-eight clichés, somewhat diluting the episode’s primary semantic texture. An ambitious translator, solicitous of Joyce’s artistic intention here, would go to the trouble of finding twenty-eight new French clichés, being careful to preserve the episode’s ceaseless, ruthless banality. But of course this solution, laudable as the effort might be, would also fail fundamentally because the twenty-eight new French clichés, like the original nine, would not be loanwords at all in the French translation. The translator would have lost the “loanword effect” that is so much part of Joyce’s art and Bloom’s narration.

6And then there is the “register problem”. What is Bloom’s project in using French loanwords in the first place? He is addressing Stephen, already fluent in French, a recent graduate of University College’s Modern Languages program, and he is striving for credibility, trying to impress him with his savoir-parler. Bloom is not using French because he wants to use French clichés. To him, these words are conjurings and auguries of a higher life. Readers see them as clichés; Bloom thinks of them as prestigious borrowings. Indeed, he has no business pronouncing most of them as French words at all. Virtually all of them have long since been assimilated to English, the greater part having been borrowed, in fact, during the 17th (tête-à-tête, gentleman) and 18th (séance) centuries, some even in the 16th century (sur le qui-vive, change of venue). In English they all have a loose, flattened, anglicized pronunciation. Bloom is borrowing them again, by restoring their French accent (which is what the italics mean in the English original). Bloom undoubtedly pauses in the syntagm, in his delivery, just before he pronounces them, to highlight his options, making his paradigm suddenly visible. For merely narrative purposes, he might have chosen the English “apropos” (no italics), “about”, “concerning”, “regarding”, “in regard to” or “in reference to,” rather than apropos in a French accent, with a French spelling mistake. The pause is a ceremonious gesture, signaling that the speaker is raising the tone of his discourse―after which he selects a prestigious French locution rather than a humble domestic one, bringing an epiphany of occult Gallic sacrament into his monologue. He has ceremoniously paused, and ceremoniously “placed” the loanword, heightening his language, he believes, but in fact mildly ridiculizing himself. Yes, French is prestigious, but Bloom uses only its English clichés, undoing his intended effect.

7These impediments are inherent and multiple. Are they insuperable? We now have the work of two French translators who have dealt with the “French loanword problem” in Ulysses: Auguste Morel in his 1929 rendering of the whole novel, and Pascal Bataillard in his 2004 translation of the “Eumaeus” episode, for the new Gallimard version of Ulysses, by many hands. Morel translated under the direct supervision of Valery Larbaud, and the looser surveillance of Joyce himself. Bataillard has worked with input from the translating team Jacques Aubert assembled, and with eighty-odd years of Joyce criticism to inform his perceptions. Each translator has a drastically different strategy, so let us examine successes and shortcomings in both, and draw out a taxonomy of practical strategies for returning loanwords to their native tongue.

8Auguste Morel must have looked into the entrails of “Eumaeus”, seen the writhing clichés, and decided, correctly, that they were the sine qua non of Joyce’s art. And looking at the paltry nine clichés among the French expressions (within French itself that is, since of course they are all clichés in English) he must have decided it was crucial to upgrade a number of the twenty-eight other ordinary, neutral expressions into blatant French lieux communs, and preserve the cliché density at least, if nothing else. But let us not speculate unduly, as we have solid data at hand. Of the thirty-seven French loaned words and expressions, Morel simply removes the italics from twenty-seven of them, making no other changes; they lose their status as cliché, and they dissolve quietly into the French text, trailing no ceremony or register incongruities to discomfort Morel’s elegant, pleasant translation. One noun phrase retained its italics, and escaped dissolution, because it was a sample of Stephen’s wit in French: he refers to Ireland as the faubourg Saint Patrice (Um: 570). The nine remaining French phrases, none of them either remarkable or clichéd, were dropped and replaced with solid French commonplaces. Bloom’s En route (U: 706), which is neutral and functional in French, becomes the fairy-tale formula chemin faisant (Um: 539), and it is at the ceremonious level into the bargain. Bloom’s in a religious silence of the strictly entre nous variety (U: 726), with its English loaned-phrase cliché, becomes the equally overused dans un religieux silence de l’espèce stricte intimité (Um: 554), which is ceremonious as well. Bloom’s introducing an atmosphere of drink into the soirée (U: 741) becomes apportant avec lui une atmosphère de cabaret dans cette société choisie (Um: 565). Soirée is unremarkable in French, but in English it is a charming and class-conscious way to upgrade “having friends over for dinner” into an elite reunion of select company. Société choisie, too, is ceremonious, has all the class consciousness of soirée in English, and some nifty snobbery as well. And so on up to nine new commonplace phrases, formules toutes faites, for a total of eighteen (nine original plus nine new).

9One can imagine two kinds of judgment on Morel’s solutions. An unsympathetic critic might judge them harshly. The argument would go something like this: it is easy enough to describe Morel’s primary strategy with Joyce’s loanwords: it is what Antoine Berman has taught us to call effacement. Morel simply removes the italics, keeping the words, and lets twenty-seven loanword effects disappear. He loses the loanword effect, the ceremonious effect that Bloom himself creates by using French, and the incongruities of register that supervene when ceremonious French phrases jostle against contiguous English banalities. If, as Berman says, following Bakhtin, “le principal problème de la traduction de la prose est de respecter la polylogie informe du roman et de l’essai” (Berman, 1999 [1985]: 52), then Morel has shirked his duty somewhat in letting so many French loanwords in the original slip away into oblivion in his translation. Has he not, a critic might ask, colluded in “la destruction, non moins systématique, de la lettre des originaux, au seul profit du ‘sens’ et de la ‘belle forme’” (ibid.)? Indeed, it might be argued, this is Berman’s homogenization as well, the whole texture of the French “Eumaeus” being significantly less various.

10But this applies an anachronistic standard, one influenced by a great deal of scholarly commentary in subsequent decades, not the least of it provoked by Bakhtin. Morel works in an older tradition, that of traduction naturalisante, though he probably would not have known the phrase. He wished to create a book that sounded as natural in French as it did in English. He was preparing the French incarnation of Ulysses, as it were, rather than a slavish or purely scholarly rendering. And thirty-seven loanwords, in the midst of two thousand eight hundred clichés, probably did not seem all that important. He increased the number of French clichés from nine to eighteen (adding nine new French clichés to the nine among those that Bloom had used), taking great pains to prevent too many loaned expressions from becoming soft spots in the Eumaean texture, and succeeded. All other effects depend, in any case, on the sheer quantity of clichés, and they have rightly absorbed Morel’s attention. Six of the eighteen are even in the ceremonious register. He did all this while under close supervision by Larbaud, a translation authority in his time, as well as with the scrutiny and concurrence of Joyce himself. He had an imprimatur from both of them. And history has accorded him its laurels. Morel’s Ulysse is a classic of French literature by common consent, his traduction naturalisante having recreated Joyce’s great work in a new language, rather than having imported it forcefully into French, blaring its loanwords, Anglicisms, and other rebarbative source-text exotica.

11When Bataillard begins translating “Eumaeus,” he does so in the long shadow of Morel, and in the bright light of eighty-odd years of Joyce criticism. Much more than Morel, he attends to the “French loanword problem,” experimenting as he goes along, and this practice lends his translation an improvisatory feel. In the end, he invests in three main strategies to make the loanword effect more visible, one of them amounting to a genuine innovation, as we shall see. He begins by reversing Morel’s primary strategy: he keeps the italics on twenty of the thirty-seven French expressions in Joyce’s English, adding texture to the visual experience and provoking readers to guess they’re looking at what were loanwords in “Eumeaus”. They do not of course produce the loanword effect in the French translation, but neither do they efface themselves. The reader feels heteroglossia enter the text and understands that Bloom must have been trying to tart up his discourse with exotic Gallicisms. Bataillard proceeds inconsistently, however, removing the italics from seven expressions, and they disappear into his French text as noiselessly as they had into Morel’s. I have looked for some common denominator among these exceptions (expressions like tête-à-tête and entrées), but to this reader they all look as worthy of italics as the twenty others. If these are not oversights or proofreading errors, then Bataillard has introduced some inconsistency into his practice.

12Bataillard’s second strategy consists of four experiments. Early in the episode, Bataillard tries out what one might call, in deference to traditional terminology, a “local colour” effect. In each instance he replaces Bloom’s French cliché with an English borrowing. These borrowed words are not accepted English loanwords in French, or even anglicisms, but simply unknown English words that translate Bloom’s French clichés. Indeed, it is possible Bataillard did not realize at first that all of Bloom’s French expressions were in fact long-ago domesticated loanwords in English. Bloom says, for example, the coup d’œil was exceedingly grand (U: 724), and Bataillard translates la vista était des plus grandioses (Ub: 780). Vista is utterly opaque to most French readers. In fact, it is an Italian loanword borrowed by English in the 1650s, and it neither looks nor sounds particularly English. Only a bilingual reader would understand it, and perhaps enjoy it. Again, when Bloom mentions a religious silence of the strictly entre nous variety (U: 726), Bataillard responds with un silence religieux de la variété strictement private (Ub: 781). Priv- looks familiar to a French reader, but -ate is quite alien, and this word is likely to be incomprehensible to the monolingual reader as well, ruining the intended effect. A third instance looks more promising. Bloom’s introducing an atmosphere of drink into the soirée (U: 741) becomes amenant avec lui une atmosphère du cabaret dans cette party (Ub: 794). “Party” is homonymous with the French adjective parti, meaning “inebriated” in informal French, which fits the context, and French readers may have heard “party,” meaning fête, on American TV or film, or understand it via French parti meaning “group”, or “political party”. This is perhaps one of the more successful of Bataillard’s experiments so far, but many monolingual readers would still stumble over it in confusion. The fourth and last experiment is the least successful, though it was in fact the first that Bataillard tried. Bloom’s The […] route […] was once more on the tapis (U: 723) becomes figurait une fois encore sur l’agenda (Ub: 779). The French reader sees a familiar italicized French word meaning “carnet,” that is, pocket diary, day planner, etc. But agenda is English here, as the italics indicate, meaning something entirely different, a list of topics or items up for discussion, usually at a committee meeting, and figuratively a list of intentions, such as those of a committee chair, or a political party. Readers will make a mistake if they construe the word as French, and they will not understand it if they realize it is English, at least the monolingual ones will not. It is interesting to note that Bataillard tried these four experiments rather early in the episode, three of them on consecutive pages (Ub: 779, 780, 781), and the fourth soon after (Ub: 794). This is a spurt of experiment which Bataillard quickly abandoned, sensing, perhaps, its difficulties for all but the bilingual reader.

13In the midst of these improvisations, however, Bataillard hits upon his third strategy, his most successful handling of the “loanword problem”, and does so perhaps accidentally, as he will not return to it until late in the episode. On page 781, just before an instance of “local colour” borrowing, Bataillard translates Bloom’s After which harrowing dénouement sufficient to appal the stoutest (U: 725) as Après ce suspense insoutenable suffisant à refroidir les plus gaillards (Ub: 781). “Le suspense” is a quite recent borrowing in French, dating from 1955, of the English “suspense”. The English pronunciation was borrowed as well. “La suspense” in French relates only to legal matters. “Le suspense,” with its English pronunciation, refers to a kind of angoisse one suffers while awaiting a dénouement. The French collocation “suspense insoutenable” translates “harrowing”, jumping out of the French text as much as dénouement does from the English, mirroring the loanword effect, if not the original loanword itself.

14This solution has several things to recommend it. Bataillard startles the French reader with an English loanword, as Joyce had startled the English reader with French ones. He recreates the loanword effect in the new French text, even if it is an English loanword rather than a French one. He translates the signifier rather than the signified. This is a truly innovative strategy, symmetrically recreating the heteroglossia of Joyce’s text. I call it the “Mirror Effect”. It simply reverses the languages on loan, and uses well-known loanwords. Bataillard returns to this invention thirty pages later, and produces it twice more. His second of the three instances is undoubtedly his best. Bloom’s The vicinity of the young man he certainly relished, educated, distingué, and impulsive into the bargain (U: 760) becomes La proximité du jeune homme il l’appréciait nul doute, instruit, gentleman, et impulsif par-dessus le marché (Ub: 811). Gentleman is quite familiar in French, having been imported into the language in 1698, in the current spelling, and in 1558 in an earlier one, gentilleman, since the English word came from the French gentilhomme (Grand Robert, 2001). And gentleman is ceremonious in French, so it maintains the register of distingué. Alas, gentleman is not quite a French cliché. The ideal solution would have been a ceremonious English-loanword cliché. Bataillard gets two out of three, which is not bad for a nearly intractable problem. He tries the same thing again when he turns Joyce’s Come, he counselled, to close the séance (U: 768) into Venez, dit-il, pour clore le show (Ub: 818). “Show” came into French in 1930, and it is an established loanword. It is not a cliché, and it is not ceremonious, but for some reason it has a mysterious “loanword pizzazz”, and works well to mirror the mocking effect of séance in English. This mirror effect seems to me a bold and pleasing strategy; it is neither harsh in its effects, nor does it drag too much strangeness into the text. And it probably works as well for the monolingual as the bilingual reader. One imagines that thirty-seven such mirror effects would have given the French “Eumaeus” all the fun-and-familiar heteroglossia that the English-speaking reader enjoys. If no one else has anticipated him, Bataillard has come up with an inspired solution to a perennial translation conundrum.

15We have seen two dramatically different approaches to the loanword effect, and we can judge them in terms of their success with “Eumeaus,” rather than against any abstract, prescriptive standard. Early on, Morel has made strategic decisions that produced a smooth, elegant translation, one looking very much as though Ulysses had been written in French to begin with. He emphasized clichés, adding nine when he translated French expressions that had been clichés in English but were not in French. As for the loanword effect, he essentially dropped it by removing italics in the French text. One may speculate that any solution he might have considered―italics, strained substitutions, gimmicks of any kind―would have seemed to bring too much abrasive strangeness into the French text, and detracted from its fluidity. Perhaps he did not see the thirty-seven loanwords as terribly important, given the episode’s two thousand eight hundred prominent clichés, and certainly not worth roughening his smooth, naturalized surface. And of course it would be gratuitous to call his solution effacement or homogenization. The loanwords are few for an episode seventy-two pages long, and the pyrotechnics of register incongruity that run through “Eumaeus” are negligibly impoverished by their absence. Morel has translated the rest of the episode’s two thousand eight hundred commonplaces, letting through all of Joyce’s gentle, arch humor as Bloom mismanages his registers—so the French loanwords are not crucial to the comedy of registers, and French readers do not miss them. Everywhere Morel’s competent hand prevails, and his translation of “Eumaeus” reads with the same consistent, elegant comedy in French as it does in English.

16Bataillard proceeds entirely differently. It is undoubtedly too much to claim that Bataillard has treated these thirty-seven French loanwords in an “exoticizing” fashion, but it is surely safe to say he has been less naturalisant than Morel. His translation is a dynamic one, with Bataillard searching for solutions, trying them out on the fly, abandoning them if they do not work, and leaving evidence of his experiments in the text, just as he made them. With Bataillard, at least with regard to the French loanwords, one sees translation as experimentation. He takes the trouble to mark, with italics, at least some of the French loanwords in the original, and to recreate the loanword effect when he is able. He leaves himself open to the ad hoc, the trouvaille, the inspiration of a moment. He italicizes a little, throws in English mystery words a little, and stumbles, virtually, onto his mirror effect. Instead of effacement, the reader sees mise en relief; instead of homogeneity, diversity; instead of smooth, rough. Bataillard took fewer strategic decisions early on, experimented greatly, and produced a more prickly, less predictable, less consistent rendering of the loanword effect than Morel did—but he may well have invented the perfect solution to our conundrum, what to do when loanwords are loaned back. In any case, his mirror effect is the crown jewel in this empirical taxonomy of Eumaean solutions.

Haut de page


Versions of Ulysses

Joyce, James, 2000 [1922], Ulysses, Toronto, Penguin Group, coll. Penguin Classics.

—, 1929, Ulysse, trans. Auguste Morel, Paris, Gallimard.

—, 2004, Ulysse, « Eumaeus », trans. Pascal Bataillard, Paris, Gallimard.

Works cited

Arjonilla, Emilio Ortega, 2008, “Luis Alonso Schökel, pionero de los estudios de traducción en lengua española. Semblanza de un maestro,” in Hernández Guerrero, Maria José and Peña Martin, Salvador (eds), La traducción, Factor de cambio, Bern, Peter Lang, p. 235-286.

Berman, Antoine, 1999 [1985], “L’analytique de la traduction et la systématique de la déformation,” in La Traduction et la lettre ou l’auberge du lointain, Paris, Éditions du Seuil, coll. L’Ordre philosophique, p. 49-68.

Lawrence, Karen, 1981, The Odyssey of Style in Ulysses, Princeton, New Jersey, Princeton University Press.

O’Neill, Christine, 1996, Too Fine A Point: A Stylistic Analysis of the Eumaeus Episode in James Joyce’s Ulysses, Trier, WVT Wissenschaftlicher Verlag Trier.

Wales, Katie, 1992, The Language of James Joyce, New York, St. Martin’s Press.

Haut de page

Pour citer cet article

Référence papier

Robert Byrnes, « When loanwords are loaned back: the conundrum of translating French into French », Palimpsestes, 25 | 2012, 249-259.

Référence électronique

Robert Byrnes, « When loanwords are loaned back: the conundrum of translating French into French », Palimpsestes [En ligne], 25 | 2012, mis en ligne le 01 octobre 2014, consulté le 28 mars 2017. URL : ; DOI : 10.4000/palimpsestes.1883

Haut de page

Droits d’auteur

Tous droits réservés

Haut de page
  • Logo Presses Sorbonne Nouvelle