The place between word (in one language) and text (in another) which I examine in this paper has been occupied most often by a human being. (Non-human interpreters include the gods Hermes and Esu-Elegba, and computers.) Hence, depictions of the act of translation fall under the rubric of Aristotelian mimesis, the portrayal of an action. Attempts at exploring the actual site, the "black box" where translation occurs are inevitably mimetic. Few plastic mimeses of translation are more revealing than a 1795 painting by the Peruvian master Bernardo Rodriguez. Saint Jerome sits at work on his Latin translation of the Bible, when he is interrupted by the word of God, presented in the topos of fanfare from the trumpet in the upper right. For Jochen Hörisch, this painting belongs to a series of illustrations of scribal transcriptions of God’s word, which eliminate the subject in favor of the only Being entitled to make the tautological statement which is subjectivity’s first and last resort: "I am who I am" (88). While the theological pretext for this painting may emphasize the gradient which drives meaning from divine word (logos) isnto human text, we notice that Rodriguez’s painterly emphasis is on the human figure of Jerome, the go-between. It is Jerome who seems to be who he is, possessing as he does one more dimension than his contexts, the book and God’s trombone. The elements which make him stand out from the background of the painting, the lines on his face, connote fear and surprise. Jerome’s humanity draws our attention more than does either the inhuman divine, or inhuman language.

Silent, the painting reminds us of the paradox in which the translator’s ability to substitute the author’s discourse with his own signals what Lawrence Venuti has called the translator’s "invisibility." Increasingly banished from the domain of co-authorship, constrained to write "I am the Other," the translator can be represented only indirectly, as essayist, theorist, or literary character. Like Rodriguez’s painting, fictional treatments of translation, such as Jorge Luis Borges’s "Pierre Menard, Author of the Quixote" and "Averroes’s Search," Nicole Brossard’s Mauve Desert, and Abdelkebir Khatibi’s Love in Two Languages, place the translator center stage, and show translation in the very processual and emotive aspects which theory must inevitably flatten out. Rodriguez’s Jerome stands as the archetype for fictional translators Pierre Menard, Averroes, Maude Laures, and Khatibi’s unnamed narrator. None of these modern texts emphasizes the transitional place between human and divine (though Borges alludes to it). Instead, the action imitated is in each case carried out by a character caught between cultures, at a place Brossard alludes to as the "Angststelle" or place of anxiety, or what Khatibi refers to as the "bilangue." For each of these characters, translation functions as a form of error necessary for life. A comparative treatment of these texts responds to Meir Sternberg’s observation that "Literary art . . . finds itself confronted by a formidable mimetic challenge: how to present the reality of polylingual discourse through a communicative medium which is normally unilingual?" (222).

One of Jorge Luis Borges’s most famous short stories, "Pierre Menard, Author of the Quixote," has frequently been read as an allegory of translation. Menard, wishes to write the Quixote, to reproduce exactly a text already in existence. It is pointed out that his task of recreation cannot simply be one of entering the mind of the text’s original author: "El método inicial que imaginó era relativamente sencillo. Conocer bien el español, recuperar la fe católica, guerrear contra los moros o contra el turco, olvidar la historia de Europa entre los años de 1602 y de 1918, ser Miguel Cervantes. . . . Pierre Menard estudió ese procedimiento . . . pero lo descartó por fácil. . . . Ser, de alguna manera, Cervantes y llegar al Quijote le pareció menos arduo . . . que seguir siendo Pierre Menard y leegar al Quijote, a través de la experiencias de Pierre Menard" (52-53; "The first method he conceived was relatively simple. Know Spanish well, recover the Catholic faith, fight against the Moors or the Turk, forget the history of Europe between the years 1602 and 1918, be Miguel de Cervantes. Pierre Menard studied this procedure . . . but discarded it as too easy. . . . To be, in some way, Cervantes and reach the Quixote seemed less arduous to him . . . than to go on being Pierre Menard and reach the Quixote through the experiences of Pierre Menard" [40]). As George Steiner points out, Menard’s "first approach to the task of total translation or . . . transubstantiation, was one of utter mimesis" (71). He ends up opting instead for the Angststelle. In truly heroic fashion, he resists contamination by the author, choosing instead the hard and arduous path of reconstruction without identification. Instead, Menard will write the Quixote by remaining Menard. Menard’s efforts cannot overcome the force of history, which alters the Quixote itself, and not just its recreation. The narrator humorously quotes a paragraph of Cervantes’s Quixote and the identically worded one of Menard’s, and explains in depth why they mean different things, given the historical gulf between the "author-functions" which guide our interpretations. That a text is, like Anaxagoras’s river, never self-identical stands as one of Borges’s main tenets. Translation provides an x-ray of the otherwise invisible process of interpretive drift, or what Walter Benjamin calls the "Fortleben" (living on) of a text, as Borges demonstrates in his relatively straightforward analysis of "Some Versions of Homer": "Since Spanish is my native language, the Quixote is to me an unchanging monument, with no possible variations except those furnished by the editor, the bookbinder, and the compositor. But the Odyssey, thanks to my opportune ignorance of Greek, is a library of works in prose and verse, from Chapman’s couplets to Andrew Lang’s ‘authorized version’" (1136). In mentioning the Quixote in his fairly straightforward comparison of English versions of Homer, Borges gives the game away for his later Menard text, where he will explicitly attempt to make the Quixote live on in Spanish. For Borges, literature is translation, as Santaella and others have pointed out.

Borge’s fictional character of Averroes also alludes to "Fortleben" in his long speech near the end of the story "Averroes’s Search." Averroes comments on a verse of Zuhair, comparing destiny to a blind camel trampling men in the dust: "Zuhair’s verse, when he composed it in Arabia, served to confront two images, the old camel and destiny; when we repeat it now, it serves to evoke the memory of Zuhair and to fuse our misfortune with that dead Arab’s. The figure had two terms then and now it has four" (154). The anecdote is a short version of "Menard": once again, the paradox of a text which changes even though it stays the same. The figure of doubling, however, alerts us to the central theme of this story which seems composed of many disparate elements: all are united as forms of doubling, and belong to what Jaime Alazraki calls the motif of the "espejo" (mirror) which one encounters throughout Borges's fiction. Both mimesis and translation produce a mirroring effect, a phantom other which Sigmund Freud has identified as the essence of the uncanny.

As in "Menard," the central figure represents a poetically impossible goal, "which is not forbidden to others, but is to him" (155). The goal in Averroes’s case is the proper translation into Arabic of the Aristotelian terms "tragedy" and "comedy." Borges follows the true story of Averroes’s mistranslation of Aristotle’s terms: Near the end of the story, after listening to his friends and enemies and contemplating the reality of medieval Muslim Spain, in which there is no theater. Averroes finally writes in his commentary on the Greek philosopher: "‘Aristu (Aristotle) gives the name of tragedy to panegyrics and that of comedy to satires and anathemas. Admirable tragedies and comedies abound in the pages of the Koran and in the mohalacas of the sanctuary" (155). Throughout the story, Borges has furnished a scenario for understanding Averroes’s failure in terms of his surroundings and upbringing: "In the depths of the siesta amorous doves called huskily, from some unseen patio arose the murmur of a fountain; something in Averroes, whose ancestors came from the Arabian deserts, was thankful for the constancy of the water. Down below were the gardens, the orchard; down below, the busy Guadalquivir and then the beloved city of Cordova, no less eminent than Baghdad or Cairo, like a complex and delicate instrument, and all round (this Averroes felt also) stretched out to the limits of the earth the Spanish land, where there are few things, but where earth seems to exist in a substantive and eternal way" (148). The relevance of this long description to the punchline of the story, in which Averroes’s failure is revealed, may not be readily apparent at first reading, but it sets up Averroes’s interpretation of Zuhair’s verse. Averroes’s inability to understand the aridity of Arabian life even though he is an Arab is analogous to his inability to understand theatrical representation. The memory of Zuhair and of camels appear as the inverse of Averroes’s here and now, the experience of life and culture which shapes understanding and prepares the failure of translation.

There is an irony in my repeated invocation of Averroes’s failure: we condemn Averroes’s choice of panegyric and anathema as generic translations of tragedy and comedy, respectively, without knowing what the Arabic words for "panegyric" and "anathema" were, nor how those genres functioned within Arabic culture. It is extraordinary to see comentators make unfounded statements, such as that this "is one of the few cases in Borges in which readers have some way of judging revealed knowledge, since readers have grounds on which to criticize the distinction Averroes has attributed to Aristotle" (Lindstrom 78). Or, in the opposite direction, to note the quest's "successful outcome" (Sturrock 86). In fact, Averroes's translation may have been the very best one available to him, given the entire lack of fit between Arabic and Greek literary practice and vocabulary. Borges's real aim, Floyd Merrell claims, was to show both the incommensurability of radically different cultural paradigms, and our pragmatic ability to do so. In such situations, "an increase of information transfer by way of the core constituents can yield at least the most adequate, though always incomplete, interpretation under the circumstances" (228).

Averroes's error is merely a more sophisticated and historically important version of what translators between English and European language face with the pronoun "you." Its all-purpose nature, from the most intimate to the most formal situations, escapes the informal-formal distinctions of "tu" vs. "usted," or the Japanese pronouns "kimi" vs. "anata." But perhaps Averroes's choice is more like the romanization of a non-Latin script, such as cyrillic. Everything is there, presumably, but in distorted fashion. These and other cases happen every day in translation, and are not considered failures or misunderstandings. J. C. Catford's classic text, A Linguistic Theory of Translation is full of these and other examples. As Catford explains, "Translation fails ­ or untranslatability occurs ­ when it is impossible to build functionally relevant features of the situation into the contextual meaning of the [target language] text" (94). Linguistically relevant features, of course, such as the examples given above, are very frequently lost, but are only accounted a translation failure if they are tied to functional relevance. Undoubtedly some functional relevance has been altered, but enough to indicate failure? More failure than in European attempts to understan the Greek theatrical experience? Only someone sharing Averroes's cultural situation, but also familiar with Greek theater could make this judgment. Indeed, at precisely the moment where the narrator writes Averroes's translation of Greek into Arabic in Spanish (what are the Arabic words for "panegyric," "satires," and "anathemas"?), reproducing Averroes's own error, the story breaks off. This is of course the moment at which Borges interrupts his story, and documents his own failure, the mirror image of his subject's.

Borges’s translation of Averroes’s failure itself fails. Hence, at the end of the story Averroes dissolves, and Borges enters the text to explain that he felt "that Averroes, wanting to imagine what a drama is without ever having suspected what a theater is, was no more absurd than I, wanting to imagine Averroes with no other sources than a few fragments from Renan, Lane, and Asín Palacios. I felt, on the last page, that my narration was a symbol of the man I was as I wrote it and that, in order to compose that narration, I had to be that man and, in order to be that man, I had to compose that narration, and so on to infinity" (155). This recursive aspect of the story is contained already in the Spanish title, "La busca de Averroes," which can be translated either as "Averroes’s Search" or as "In Search of Averroes." Averroes the translator is, then, another image of Borges the author, and mimesis becomes a kind of translation. Although commentators have frequently cited this last paragraph of the story, they have not applied it retrospectively to the scene in which Averroes supposedly fails in his translation. "El cuento es una toma de conciencia del riesgo y del margen de error admisible en la interpretación del pasado histórico cuando la única fuente son los libros escritos por otros" (Cédola 256). Averroes relies too much on the Quran for his translation, Borges on Renan, Lane, and Asín Palacios for his mimesis of Averroes's reliance.

In Brossard’s novel, Mauve Desert, as in Borges’s "Menard," a work is translated from one language into the same language, with considerable attention paid to the figure of the translator. However, if Borges used intralingual translation as an x-ray of the process of interpretive drift, Brossard uses it as an x-ray of the translation process itself. Paradoxically, we are blinded to what is happening in most translations by the fact that they take place between two languages. Linguistic difference per se occupies our almost total attention, obscuring the fusion of horizons which take place in the translated text. Brossard herself defines translation as "a complex act of passage which inflame[s] the mind," that we translate "our emotions, our sensations, our values, that the words of others could only make sense if we translated them through our own experience, our referents, that part of life escaped us because we did not have the words to name, to translate, that is to embed into our experience the sensations, the feelings, the concepts which seemed so obvious in another languge" ("Nicole Brossard" 53). These gaps in experience and expression form the thematics of Mauve Desert, where, as Susan Kutson has put it, translation functions as a "figure" for Brossard's more central thematic of "making sense" (faire sens), which involves "the interaction of information-saturated wave-fronts in the brain" (206). The figure stretches from Brossard's 1985 L'Aviva, in which Brossard provides French to French translations for ten poems, to the 1989 theory/novel, Picture Theory. By producing a French-French translation, Brossard is able to foreground that fusion between lesbian and heterosexuality, Quebecois and Arizona landscapes, and adolescent vs. middle-aged mentalities. The book is divided into three parts, the first being Laure Angestelle’s novel, Le désert mauve, and the last Maude Laures’s translation of the same, Mauve l’horizon. The middle section is a kind of translator’s notebook, where Laure records notations on characters, settings, concepts, and her interview with the author. The book to be translated is a partial autobiography by Kathy’s daughter Mélanie. It concerns Kathy Kerouac and her lover Lorna Mynher. Mélanie’s passion for driving through the desert finally cathects on the "geometer" Angela Parkins, who is murdered by l’homme long (Longman), the only male figure in the work.

The fictional translational project is presented as a mise en abyme of the policial pastiche of the source text. Thus, the only question Laures finds worthy of asking the author in their imaginary interview is the basic one underlying detective fiction, spoken from the viewpoint of Angestelle’s victim Parkins: "Pourquoi m’as-tu mise à mort?" (141; "Why did you kill me?" [132]). The subsequent exchange between author and translator implies (but does not prove) that Longman assassinated Parkins. As a representative of the patriachal, geometrical reality principle, he prevents Angela from crossing the threshold from desire to praxis, from fiction into reality. Like a detective, Maude Laures works on the principle of identification. As the detective begins with the crime and works backwards to reconstruct its genesis, the translator begins with the fait accompli of the book, and attempts to recapture its "trajectory." Detective and translator alike are contaminated by this principle of identification, as it draws them closer and closer to the thoughts and language of the originator of crime and text: "Trajectoire, pensait Maude Laures, trajectoire. Et elle se faisait de plus en plus à l’idée de devenir une voix autre et ressemblante dans l’univers dérivé de Laure Angestelle. Les personnages allaient bientôt se défiler les uns après les autres, devenir de petites transparences au loin, se cristalliser. Elle serait seule dans sa langue. Alors, il y aurait substitution" (176; "Trajectory, thought Maude Laures, trajectory. And she progressively got accustomed to the idea of becoming a voice both other and alike in the world derived from Laure Angestelle. The characters would soon slip away one after the other, become little transparencies in the distance, crystallize. She would be alone in her language. Then would come substitution" [160]). Interestingly, Borges’s Menard rejects such identification as too easy.

The middle section of the book presents us with various stages of the translator’s quest for identification with the author. Maude Laures visits the places of Désert mauve and embodies her experiences in prose poems on ideas such as "desert," "dawn," and "light." These entries reflect her realization that Désert mauve is itself a translation of space and landscape into language, movement, and emotion. Mélanie’s "translation project" extends not only Laure’s, but also Brossard’s own repeated attempts at destructuring formal space "by the transformation back and forth of energy and matter through pulsing acceleration, as opposed to space which has been mapped by tradition and convention" (Forsyth 334).

Laures entitles one section of her translation "Laure Angestelle," and imagines the author as several people: she progressively reduces Angestelle’s imagined age from fifty to thirteen; leaves undecided whether she grew up in the Western desert or moved there from the East; and outlines numerous possible circumstances and reasons for the publication of this author’s lone work. But, "tout cela qui pouvait être fantaisie n’invalidait pas la pensée que Laure Angestelle ait sans doute été une femme fière, au corps agile, aux yeux pleins de tourment, vulnérable devant la beauté et le silence" (90; "all of this which could be fantasy in no way invalidated the thought that Laure Angestelle had no doubt been a proud woman with a supple body, eyes filled with torment, vulnerable in the face of beauty and silence" [83]). Silence is the unchanging constant in Laure Angestelle seen from all perspectives: "avant que la pensée ne s’exerce à distinguer entre les paroles, les rires, le discours, le bouillon de culture, il lui fallait du silence,mettre du silence devantles êtres comme un écran car elle savait que la beauté était au prix du silence qui accordait toutes les musiques" (88; "lest thinking start distinguishing between words, laughter, discourse, the culture medium, she needed silence, to put silence in front of beings like a screen for she knew the price of beauty to be the silence attuning all spheres of sound" [82]).

Most readers would agree with Laures that silence or reticence is essential to Angestelle’s novel. Can Laures’s translation be faithful to that silence? The torrent of language, like the steady stream of sounds and images from Kathy Kerouac’s television set, isolates rather than unites, creates rather than removes ambiguity. The novel’s language builds mystery: who, for example, is Longman, the novel’s only male character? Is he the assassin of Angela Parkins? The translation chooses not to reveal any more than the novel does. Laures tends to trade one silence for another. This preservation of obscurity moves against the well-known axiom that translation always interprets, replacing connotations with denotations. If Menard’s Quixote changes the original by reproducing it verbatim, Laures’s translation changes the language of Mauve Desert while keeping the text constant. For example, Angestelle’s Mélanie writes: "J’avais roulé toute la nuit. Tucson n’était qu’à quelques kilomètres mais je n’étais pas encore prête à retrouver la peur panique. Je m’arrêterais à ce motel tenu par une amie de ma mère. Je dirais ma fatigue et mon besoin de dormir. Elle m’offrirait une chambre. Je retournerais à la Meteor pour prendre mon sac" (42; "I had been driving all night. Tuscon was just a few kilometers away but I was not yet ready to encounter panic fear again. I would stop at that motel run by a friend of my mother’s. I would tell of my tiredness and need to sleep. She would offer me a room. I would go back to the Meteor for my bag" [38]). On the other hand, Laures’s Mélanie writes, "J’avais roulé toute la nuit. Bientôt je retrouverais Tucson mais je n’étais pas encore prête à affronter la peur panique et le quotidien recommencé du Motel Mauve. Je préférais m’arrêter au Motel Rouge dont la gérante était une amie de ma mère. J’inventerais une histoire, je dirais ma fatigue et mon incapacité à reprendre la route. La gérante m’offrirait une chambre. J’irais chercher mon sac dans l’auto" (212; "I had been driving all night. I would soon be back in Tuscon but I was not yet ready to face the panic fear and the repeating dailiness of the Mauve Motel. I chose to stop at the Red Motel which was run by a friend of my mother’s. I would make up a story, I would tell of my tiredness and my inability to keep driving. The manager would offer me a room. I would get my bag out of the car" [194]). Laures supplies names for the two hotels (no doubt culled from her research) where Angestelle is silent. She also names Mélanie’s strategy of making up a story. However, Laures gives the genus "car" where Angestelle had specified the species "Meteor." This game of relative specificities is played throughout. Where Angestelle is satisfied to say that Lorna "inventait" (2), Laures adds several sentences on her invention of reptile names (182). On the other hand, where Angestelle lets Mélanie write that her first word on seeing Lorna was "salope" (2), Laures prefers the genus: "un mot sale" (182). This trading of genus for species and vice versa is, as Aristotle pointed out, a common device of metaphor, and points to the poetic, emergent relationship of Laures’s text to its original. As in Benjamin’s view, translation here functions as a complement to the original, rather than as its duplicate or substitution.

When author and translator come together for an interview, they share this community of silence: "toutes deux aiment composer avec le silence mais chacune ici cherche à comprendre comment la mort transite entre la fiction et la réalité" (140; "both like dealing with silence but each one here is looking to understand how death transits between fiction and reality" [131]). Falling out of character, Laures tries to assert her rights as translator:

However, the debate about "your" (source) vs. "my" (target) language takes on new meaning, given that the translation is from French into French, recalling Menard’s "translation" of the Quijote. Laure’s reproaches to the author primarily concern her "killing off" Angela Parkins, which however Laures cannot bring herself to undo. A successful intervention, on the other hand, can be seen in a consistent softening of the lesbian scenes. One example among many is the description of Lorna’s schooldays: "Lorna n’avait pas connu d’enfance, seulement des filles après l’école à qui elle donnait rendez-vous avec ostentation à l’heure du midi. Les filles aimaient l’embrasser sur la bouche. Elle aimait les filles qui se laissaient embrasser sur la bouche" (2). The translation of those sentences begins with an almost complete synonymy, but differs radically when it broaches the subject of sexuality: "Lorna n’avait pas connu de jeunesse, seulement des filles de sa classe avec lesquelles elle s’acoquinait à l’heure de la récreation. Les filles l’aimaient. Elle aimait les embrasser sur la bouche à la sortie de l’école" (182). Mélanie’s description of Angela Parkins receives similar treatment: "Comment Angela Parkins faisait-elle l’amour quand elle n’était pas au bord de l’ivresse?" (30) becomes the more elegant and indirect "comment ses gestes témoignaient-ils dans la chaleur et la soif de l’ivresse amoueuse?" (210). Sexuality is not eliminated from these passages, but is presented more ambiguously. While agreeing with commentators like Henri Servin who note that the middle panel of the novel’s tryptich is devoted to the "démarche intellectuelle et surtout affective d’une femme, Maude Laures," whose translation denotes "le passage d’une vision de l’écriture à l’autre, d’une langue à l’autre, notions qui, sur le plan métaphorique, incorporent le thème sexuel et surtout le thème du désir lesbien" (57), I would note the undecidability of this development, of the success or failure of translation, as a parallel to the undecidability of Angela Parkins’s murder. The textual evidence would suggest that Laures is caught between her great desire to translate a text which fascinates her, and an inability to enter the lesbian sexuality which the original text presents, nor into the kinds of description a fifteen-year old would make of that sexuality. Laures as a translator is thus in a situation analogous to those of Averroes and Menard, in which translation’s failure is seen as a descent into the gulf between sets of experiences of author and translator. The particular sets of experiences, of course, are vastly different. For example, the question of women’s vs. men’s experience and of the adequacy of language to capture that experience does not arise in Borges’s texts, as it does explicitly in Mauve Desert and implicitly in Love in Two Languages. Laures is perhaps representative of women drawn to feminism, and trying to invent a new language for themselves through translation.

The influence of landscape, on the other hand, does recur in both Borges and Brossard. As if she had read "Averroes’s Search," Angestelle fears that the landscapes of the translator may erase those of the author. Landscape is not a phenomenal notion here, but a linguistic one. Landscape is expression, and expression is understanding. The "fusion of horizons" between translator and author may be drawn from the perspective of the translator. In the introduction to his book of Imitations, the poet Robert Lowell, updating to John Dryden's centuries-old commonplace, notes that in his translations of Rilke, Montale, and Baudelaire he tried to do "what my authors might have done if they were writing their poems now and in America" (xi). That is, Lowell reimagines these poems as if their authors had grown up interiorizing American rather than European landscapes, history, and politics. Angestelle’s act of translation places the interpretation of desert landscape "en abyme." The contradictory and complex intervention of this landscape in the lives of the novel’s characters escapes summary: it is a discontinuous function. Like Jean Baudrillard, Brossard seems to have found in the desert a questioning of all previous reference points of cultural meaning. In the desert, things stand for nothing but themselves, posing the question: "how far can we go in the extermination of meaning, how far can we go in the non-referential desert form without cracking up?" (10). Longman cracks up, as does the geometer Angela Parkins, as does the translator Laure Angestelle, as does the reader of Angestelle’s translation, who will seek for the similarities between original and reproduction, confined by the iron law of rewriting to a mathematical analysis of similarity and difference whose boundary function will ultimately fail.

Abdelkhebir Khatibi’s Amour bilangue (1983) documents an eternalized moment of indecision between French and Arabic languages and cultures. Khatibi’s fiction may be read as the mimesis of a love affair between a bilingual Maghrebine man and a rather less bilingual French woman. There is no single sentence in this deconstructive text which states this plot plainly, but we are given clues, such as the "gray and gloomy weather where she had spent her childhood" (18; "le climat gris et maussade de son enfance" [12]), references by the narrator to French as his second language, and the occasional Arabic words which faintly punctuate his text like stars. The love affair proceeds as most do, through the ineluctable stages of paradise, lapsus, expulsion, and finally exile.

Translation functions throughout Khatibi’s text as a metaphor or allegory for love: "An extraordinary translation seemed to unite us. I mean . . . that in my own language I myself was transposed, transplanted into a sham so fantastic that she could only conceive of my telling of the unnameable in her as an enormous fiction" (62; "Ce qui semblait nous unir était une extraordinaire traduction. Je veux dire . . . que j’étais moi-même transplanté dans ma parole maternelle en un simulacre si fantastique qu’elle ne le pouvait concevoir, à son tour, que comme une grande fiction" [71]). We notice that this metaphor is reversible: love is also a metaphor for translation. The narrator’s lover, in turn is a metaphor not for the French language, but for the intoxicating flou of meaning and consciousness which arises in a bilingual situation: "The bi-langue! The bi-langue! Herself, a character in this story, on her intercontinental quest, beyond my translations" (98; "La bi-langue! La bi-langue! Elle-même, un personnage de ce récit, poursuivant sa quête intercontinentale, au-delà de mes traductions" [109]).

We witness here the same mise en abyme as in Borges and Brossard. The author’s beloved in not so much depicted in his text as much as she emerges from the bilingual process which continually creates and transforms textuality. Loving the Other becomes possible only in the erasure of difference, literally of the difference between languages. Thus, the bilingual subject of Amour bilingue seeks to make his voice heard by his lover’s "third ear." While standing in front of a simultaneous translation booth, he asks her, "Shall we lock ourselves inside to translate each other?" Translating the Other thus becomes the program of constructing a "bilingue." Khatibi allegorizes this product of translation as the couple’s child, "as if your past had married mine and given birth to a child ­ our love" (18; "comme si ton passé avait épousé le mien, accouchant d’un enfant ­ notre amour" [24]). This metaphor is analogous to Averroes’s anecdote concerning Zuhair: two terms have reproduced themselves and now become four. Meaning is a child of the lover/reader and his beloved text. This new element is also the horizon of Maude Laures’s Mauve l’horizon, something which will exist but which is as yet unknown, the language of the future, transformed through a continual process of translation until it is able to convey women’s experience.

Khatibi's narrator supplies the most explicit links between the act of translation and the situation of the "postcolonial subject," a situation, as Tejaswini Niranjana points out, which translation theory has tended to avoid: "The 'empirical science' of translation comes into being through the repression of the asymmetrical of power that inform the relations between languages" (60). Henri Meschonnic have also called our attention to the fact that translation is always carried out under political conditions, and will reflect the power relationships inherent between any two languages. The relationship between French and Arabic is asymmetrical, (although both could be called "colonial" languages imposed on the Moroccans in different epochs). Khatibi shows this asymmetry both through the "theory" of translation which emerges from the pages of Love in Two Languages, and through its diegesis. I have outlined the former aspect elsewhere (cf. Beebee ). In the context of this paper, I would like to examine two of the text's "events."

While each author’s fictional is thus somewhat different, these three texts raise similar issues of translation, which I summarize here:

1) All three authors focus on intralingual translation. Paradoxically, we are blinded what is happening in most translations by the fact that they take place between two languages. Linguistic difference per se occupies our almost total attention, obscuring the fusion of horizons which take place in the translated text. By examining a Spanish-Spanish translation, Borges is able to reveal this fusion of horizons. By positing French-French translations, Brossard is able to foreground that fusion between lesbian and heterosexuality, Quebecois and Arizona landscapes, and adolescent vs. middle-aged mentalities, while Khatibi is ableto expose the palimpsestic layers of Maghrebine cultures.

2) Fictional translation situations seem to be anti-theoretical, in that they represent the only discursive vehicle for highlighting the presence rather than the absence of the translator.

3) Surprisingly, all three authors, like the painter Rodriguez, associate the translator’s presence with a mental state of angst. This anxiety, which of course has its productive side, proceeds from the instability of the translator’s position between languages. The translator’s job is to saw off the limb she is sitting on, to close off the passageway from one language to another with the perfection of a finished product which substitutes rather than unites.

4) All three authors achieve a mise en abyme which makes their translator-heroes function as images of themselves. This strategy suggests that translation and writing are allied phenomena of reinscription. Though there is undoubtedly a difference between Borgesian neo-platonis tendencies and the emergent qualities of the texts of Brossard and Khatibi, in all events the poetic word belongs to a "reine Sprache" which exists in the interstices of texts as they transcribe themselves through poets. Language and literature flow uphill, against the trumpet’s blast, from text to word. It is perhaps no accident that all my fictional translators inhabit postmodern texts: the author dies, and finds her "Fortleben" in the translator ­ Angstelle’s first name, "Laure," lives on in Maude’s last name. The translator, under erasure for centuries, emerges from the background like St. Jerome.