RSS feed

On the Pragmatics of Translating Multilingual Texts[i]

Anthony Pym
Universitat Rovira i Virgili, Intercultural Studies Group, Tarragona, Spain


Translators increasingly have to work on texts written in two or more languages. Such texts might typically be working documents, minutes of meetings or similar interim reports on the activities of scientific research teams, international bureaucracies or multinational companies. Indeed, they are likely to emanate from any institutional framework where more than one language is used. The result is that many technical translators are now called upon to work from a multilingual source texts, and do so quite successfully. Yet their success is at the same time a failure for many traditional and not-so-traditional ways of looking at translation. The rendering of these texts requires a mode of pragmatics that adopts an economic-probabilistic approach to the genealogy and authority of texts, ultimately accepting that the place of source-text production may be more intercultural, indeed more hybrid, than that of translations. I hope to illustrate this through three playful examples, including a fragment of my own translating in this field. I will finally argue that this kind of work should make us think seriously about alternatives to translational discourse itself.

Initial Problems for Theories

Multilingual source texts obviously present problems for any theory that assumes a text to be embedded in a particular language or culture. Indeed, the phenomenon should radically question the relation between particular languages and particular cultures. If I am translating the multilingual minutes of a European-Union meeting, for instance, the cultural context able to solve most of my problems is usually that of the meeting itself, of the things being discussed and the history of EU regulations, than the national sociologies of all the countries represented. To that extent, the multilingual texts correspond to specifically intercultural situations, if not to socially developed intercultures. Meanings are not so much in the languages or national cultures as in the intercultural situations in which languages are being used.

Recognition of this problem is by no means new. A certain linguistic awareness of textual plurality can be dated from the beginnings of poststructuralism. Greimas, for example, criticised van Dijk's normative "text in L" as depending on "a rationalist conception of natural-language discourses (...) inherited from positivism" (1972: 9). Fuller critiques can be found in Derrida, particularly in his essay on translation (1985) and the punning exclamation "Plus d'une langue!," insisting that we should no longer talk about texts being in just one language. On a more practical level, Pöchhacker (1994) has emphasised that simultaneous interpreting, as an acute case of translation, never entirely blocks out the input text, which remains present not only in the form of calques and mandatory textual rhythms but is also physically present to the extent that receivers of the output have visual and auditory access to the input speaker. In such cases, it is impossible to assume that translation simply goes from a "text in L1" to a "text in L2." Two or more languages are linked throughout the entire process.

Yet the multilingual situation need not present problems for translational competence as such. If you can translate from one language, why not do so from two? And the intercultural situation is indeed there as a context for translational decisions. The real theoretical problems begin more on the level of determining the purpose and viability of a translation project, particularly as regards readership. If the people producing the text could understand several languages, what do they need translations for? Who might actually require a "translation in L"? And at what price can we afford to give them one?

My first example thus concerns a case where the "translation in L" option has been refused. 

First Example: Economics

Here I have a text that names its intertextual status: The Official Politically Correct Dictionary and Handbook (Beard and Cerf 1993) is an extensively documented satire of politically correct English in the United States. As a dictionary or pseudo-dictionary, the work constantly refers to texts outside of itself, in this case to citations that are all ostensibly within the one culture. Of little import that the texts might be invented, or common sayings, or hearsay. The principle remains the same. The dictionary is highly embedded in its language and culture; it's the kind of text that keeps translation theorists happy. So what happens if we try to translate it into a different culture? For instance, the culture of Catalan.

Any translation theorist will tell us that at least two things can happen. Either we produce a text on political correctness in the United States (and thus a commentary on certain usages of English) or we adapt the text to the target culture, talking about forms of political correctness that the reader of the translation is likely to experience (in this case referring to the Catalan language). One intertextuality or the other. As Schleiermacher (1813) put it, the translator must either move the reader towards the author (teach the Catalan reader about English) or the author towards the reader (make the English dictionary talk about Catalan culture). One movement or the other. This would seem to be what translation can do with cultural embeddedness.

The problem is that no one, to my knowledge, has made either of these moves in this case. There would be little economic logic behind a Catalan pseudo-dictionary of English phrases, quite simply because anyone able to understand the English examples could read the book in English in the first place. Similarly, it would not make financial sense to translate the work by systematically replacing the English phrases with Catalan examples, since the result would be so Catalan that there would be no reason to call it a translation and even less reason to pay the corresponding royalties. Despite the either/or theories (either source-language or target-language intertextuality), the economics of translation are far from binary.

I have nevertheless come across a Catalan commentary on the Politically Correct Dictionary: "La correcció", by Quim Monzó, is a very short piece that explains the general idea of the American dictionary, gives a few examples, and basically makes fun of political correctness in both the American and Catalan contexts. Monzó's text obeys a certain economic logic by mixing translation and nontranslation, source-language and target-language intertextuality, in a way that few translation theorists would want to touch. This strategic mix is worth analysing.

The two pages of Monzó's text reveal at least four different strategies that could be described as translational or peri-translational: 

1. Source-language phrases are rendered at word level: "Treballar com un negre" is entirely recoverable as "To work like a nigger".

2. Source-language cultural items are replaced with functionally equivalent target-language items: "Hi ha moros a la costa" ("There are Moors in sight", i.e. "Watch out!") might render some similarly innocent phrase like "Don't be an Indian giver"; "Jesús Puente és calb" legitimately corresponds to "Yul Brunner is bald" (i.e. "follicularly different").

3. Source-language phrases are explained: "En anglès, la cosa arriba fins al punt de qüestionar history. History significa història; però, alerta, his significa «d'ell»." ("In English, things have gone so far that even the word history is being questioned. History means 'history,' but watch out, his means 'of him'.")

4. Ostensibly Catalan material is added: For example, a certain woman teacher at the Universitat Autònoma de Barcelona is supposed to have refused to give seminaris because the term comes from semen. This, however, could be considered a domesticating translation of the rather tired joke invested in the term ovular, glossed in the American text as "a nonphallogeneric term for seminar" (Beard and Cerf 1993: 53). And so on.

The important point is that all four strategies work in unison, producing a text that is restricted to neither one language nor the other. The result is a commentary on political correctness in at least two cultures. A monolingual text has become a multilingual text. That is what interests me. The phenomenon exists. And although Monzó's text is not a translation in any strict sense of the term, it does incorporate several translational strategies. In fact, a kind of translational discourse would seem to persist here even in an intercultural situation where translation itself is not economically viable. This is an interculturality that might break with the Schleiermacherian binarism of "either/or" movements.

What happens when we try to apply this kind of interculturality to translation in a slightly stricter sense? Is such mixing possible? How can it be perceived and interpreted? 

Second Example: Probability

I once received a letter from Italy inviting me to write up a paper for publication in some conference proceedings. The letter is in understandable English with just a few traces of Italian: I am addressed as You, with the capital, which I could interpret either as the sender shouting at me or as a calque of the Italian Lei. The latter reading seems more charitable. So I was led to read the text as referring back to an unseen Italian text, probably a text that existed only in the mind of the sender. In fact, I was forced to interpret the letter as being partly translational, as a mixture of English and Italian within the context of an Italian-English conference. And this is more or less how I read it, despite the severe limitations of my Italian. This is like the Catalan text on American political correctness except that here the hermeneutics have to discover a pre-textual separation of languages: I am required to believe that the writer was thinking in Italian prior to writing in English, and that the second language was supposed to cover the semantic space of the first. To that extent, the letter is slightly more translational than is Catalan political correctness.

Consider, now, the fact that this letter, sent in October, asked me to send in my paper by Easter. Yes, Easter, some seven months after October. Italians obviously have very generous notions of time. No need for me to hurry on this one! But I was worried about those seven long months. I questioned the word Easter as a translation of the Italian Pasqua; I checked a pocket-sized Italian-French dictionary (the only one at hand) which gave not only Pasqua as Pâcques(Easter, so all was well) but also Pasqua di Natale as Noël (Christmas!). I didn't even bother checking with the writer of the letter: I sent off my paper before Christmas, not Easter. 

What was happening in the interpretation of this letter? I had to project my mind back and formulate hypotheses about the way the letter was produced; I had to attempt what I would like to call basic textual archaeology. But I certainly didn't reach any moment of revelation. In fact, I have since been told that my interpretation was quite wrong: Easter meant "Easter", and the bit about Pasqua di Natale is outright over-interpretation. Yet I insist that the outcome of my interpretation—the turning of the letter into the action of sending in my article before Christmas—was more than correct. Consider the pragmatics. There is a high probability than the letter-writer's Easter means "Easter" and a low probability that it means "Christmas", but the action resulting from the first reading (sending in the paper after Christmas) involves a far higher risk of being wrong than does an action based on the second reading (sending in the paper before Christmas). So I had to adopt the less probable interpretation in order to produce the action with the less risk. The only negative consequence was likely to be reinforcement of a few Latinate presuppositions about Anglo-Saxon punctuality. But I was going to be condemned for that anyway.

My point here is that the hermeneutic archaeology of the multilingual text allows outcomes that are probabilistically correct even in the absence of firm knowledge of language systems. That is, I can read a text in terms of a language I don't know, I can be philologically wrong, but I can still be right in pragmatic terms. The one condition is that I assess the pertinent probabilities and risks in terms of the intercultural situation in question.

Can this principle of interpretation be seen in the way a translator works with multilingual texts?

Third Example: Authority

I would now like to venture into the field of fully declared professional translation. My example (Text 1) is fairly representative of the kind of document I have been translating for the past seven years or so. It comes from a sociological research centre based in Barcelona. The centre has several French personnel and works for clients like the European Commission. The documents I have to translate can be in Spanish, Catalan, French, English or any mixture of these, often with defective usages as writers switch to non-native tongues. The text I receive is a document that would probably never be presented to anyone outside the production group. In fact, my source text is no more than a pre-text, a multilingual sequence of notes and fragments that function as instructions about what kind of target text I should write. As a translator, I am called upon to give the source not only a certain uniformity (I translate it into English) but also cohesion and coherence, often involving a reorganisation of the contents or a degree of summarising. The translations are usually for publication in specialised journals, conference proceedings, or applications for EU subsidies, where success is ultimately determined by whether or not we get the subsidy. These are situations in which translations are considered economically viable but high-quality source texts are not. My position is thus the internal/external divide that separates the actual workings of the research from its presentation to outsiders. My position also involves concrete pragmatic action across this interface: for instance, if the group receives a subsidy for a large project, I get quite a few well-paid translations to do. So as long as I want to remain a translator, I have a financial interest in making whatever changes are necessary for a successful outcome.

My example has some fairly complex textual features. It is basically readable as a series of fragments with quite different lingual and discursive qualities, each fragment referring to or drawing directly on a previous parent text and being reworked along the way. The following fragments—"language-discourses" if you like—are readily identifiable: 

1. Defective English: Basic errors like "preparted", "the paper have" and "elaborted" (all this before line 2!) indicate either a very non-English typist or a non-native rewriting of a previous English-language text.

2. Correct technical English: Some passages are in surprisingly good English, not just in the sense of having no errors but also in that they use structures one would not expect of native speakers of Romance languages. Note especially the passives in paragraph 2 ("the bulk of their technology is physically closed off...") and the possessive+gerund structures in the final paragraph ("...their housing in the processor casing...").

3. Non-defective Spanish: The Spanish sections are by no means good technical language but they are at least recognisably Spanish. The odd syntactic botch can be attributed to inattention and not much else (in paragraph 3, I invite you to sort out "la investigación completadas con una fina observación..."). 

4. A French subtext? The one proper name, Monique Dupont (yes, I have changed it to protect the innocent) is neither English nor Hispanic. It seems unlikely that this author actually produced all the English-Spanish discourses. But if not, what is going on here?

When I started translating this text, I obviously had to ask myself what was going on. I had to read it in much the same way as I read the Italian-English letter. Of course, in an ideal world, my client-author should have explained the task to me beforehand; they should have specified what kind of translation was required, for whom, in order to achieve what particular purpose. But they didn't. Translation in Spain simply doesn't work like it does in the textbooks. My one instruction was, "It's urgent." But after years of translating this kind of text you become an adequate archaeologist: experience and a bit of common sense can usually sort out what is going on.

In this case the situation was not terribly difficult to ascertain. Since the name Monique Dupont is mentioned later in the text, in the third person and attached to a present tense, it seems likely that what we are reading is a report on a research paper presented by this French name (later specified as one research paper, thus solving the ambiguity of "The paper have" in line 1). The source text thus involves a kind of reported speech. The reporting discourse is presumably mostly in Spanish, given the more synthetic nature of the Spanish passages and the fact that this language would have little other reason to be in the text. The reported discourse would then show through in the correct technical English, possibly written by a native speaker of French, although phrases like "their housing in the processor casing" would appear to belong to some previous paper written by a very native speaker of English. Between these two language-discourses we find considerable mixing, patches where the reporting discourse merely transforms the reported discourse, producing an English with frequent Romance-language interference. These mixed areas might be described as intercultural, interlingual, or even, in the terminology of second-language-acquisition theory, interlanguage. Any translator who transforms these defective passages into acceptable technical English is partly working as a language teacher would. In fact, the notes and lines I have scribbled on the text are rather like those I use when correcting my students.

Yet a translator is not quite a teacher. The power relations are different. In the language class, the teacher can assume a certain authority. In the translation situation, the producer of the source text retains expert status; the translator merely provides a service designed to extend that authoritative status across language boundaries. Indeed, since the terminological content of this multilingual source contains much of its own translation, the authoritative author, in a position of relative power, is telling the translator what terms to use. Comparing the passages in different languages we find that forma is to be rendered as shape rather than form; the pair interna/externa is inside/outside and not internal/external; the term household appliance is admitted as a legitimate synonym of domestic appliance, and so on. Many translation problems are solved just by comparing the discourses and trusting that the author or authors, whoever they might be and whatever the quality of their English, actually know what they are talking about.

All this can more or less be assumed by a probabilistic interpretation of the source text. Problems can be solved; high risks can be avoided. But what happens when the interpretation is likely to be wrong or tenuous? How can we assess the risks involved?

To take just one problem, consider the couple innovator/innovador, which appears in English in the second paragraph and in Spanish in the third. The instruction is clear enough: all future occurrences of the term innovador are to be rendered as innovator. This is exactly what I did the first time round. During the revision process, however, something sounded wrong. Surely "innovators" invent things, or promote new techniques, as in art history or poetry? It would be strange, but not impossible, to walk into an appliance factory and find an "innovation department", with a salaried worker occupying the position of "chief innovator". I would be less surprised to find a "Research and Development Department" or a "Design Department". Yes, that's it, "industrial design"! Surely the people in question are "industrial designers", or just "designers"? But if I put designer, how should I explain the term innovator in the source text? What is to be done with the authority of my authors?

The authority of the source text, even in the case of multilingual pre-texts, is not easily subverted. The writers are, after all, the experts in these particular intercultural situations. Yet the probabilities and risks of the multilingual text may perhaps allow problems to be solved on the basis of something more than projected authority.

Consider a possible French connection. The English word designer is often used in French as designer, a clear Anglicism (registered as such in the Robert, with an example from 1971). Standard bilingual dictionaries insist that the equivalent is dessinateur. But this latter term could create confusion, since it also means "sketcher" or "drawer". Hence, no doubt, the French designer. If only the politically correct French liked Anglicisms. The political French even have a law, "la loi Toubon" (sometimes called the "All-Good Law") that aims to get rid of these foreign intrusions. Thus, perhaps, the Anglicism designer was replaced by the acceptably French innovateur.

If we can presume some such switch, it is possible to create a genealogical narrative for my source text: In the dim past a text in good technical English talked about the designers of household appliances; a French text in the same field avoided the Anglicism designer and used the term innovateur; a French author writing in good English saw no reason not to render innovateur as innovator, which is passably correct English; a Spanish writer reporting on the resulting text consequently accepted the authority of innovator and rendered it into Spanish as innovador. Of course, this last step is rather difficult to justify: Spanish also offers the quite acceptable term diseñador, which is not likely to be confused with someone who sketches or draws (a dibujante). In fact, the entire story is a little hard to believe. It could be pure invention, like my over-interpretation of the Italian Pasqua. Perhaps I am simply out of touch with the English language; perhaps industrial designers are indeed now called innovators. But correctness is not the point; there is rarely any absolute revelation in such matters, and my client usually checks the translation anyway. What counts is that the narrative was fleetingly invented, in the mind of a working translator, to solve a problem, to explain away the source-text instructions, to subvert the authority of the author and perhaps even to protect the Englishness of English by calling an industrial designer a low-risk designer. Only by analysing multilingual textual relations could a translator effectively dispense with the same relations.

Against Translation

I attach my English translation of the multilingual source (Text 2). By no means a model, it is the best I could do within the time frame allowed. It is probably quite representative of my work on this kind of text. The translation is entirely in English; the multilingual features of the source have been entirely concealed; there is little trace of mixed strategies or any extensive genealogy. Sentences have been cut; some ideas have been reordered; one whole sentence has been added to the first paragraph, warning the reader that "The paper focuses on the design and use of a food processor" (nice to know!). Quantitative analysis could probably reveal that the translation bears certain linguistic features specific to translations, that it is more explicative that nontranslational texts. But my translation is not half as much a hybrid as its multilingual source! The function of translation is to cover over the mixes. The result is a clean text, with no history.

There is a lot wrong with this kind of translation. As Lawrence Venuti (1995) has argued, such domestication masks the foreignness of the source; it creates a falsely homogeneous target language and thereby makes the role of the translator invisible. Since I, as a translator, am discursively accorded no subject position in these translations (I cannot say "I"), since I am technically invisible, should I really be surprised to find myself receiving ridiculous time constraints and little social prestige? Venuti and others, from Schleiermacher to Berman, believe these evils can be overcome by a more literalist kind of translation, by sticking close to the source text and producing translations that show as much foreignness as possible. These theorists would perhaps have translated the technical text term for term, innovador for innovator, strangeness for strangeness. They might even have been paid for their work. But they wouldn't have been paid as well as me. In fact, they probably wouldn't have heard from the client again. Nor would their literary snobbery have attracted many nonliterary readers.

In the technical fields where I live and work, the role of the translator is unavoidably to make multilingual texts vanish. It is a difficult, expensive and ideologically nefarious process. Yet it is a task now inscribed in the very nature of the profession. It is moreover a process that might currently consumes about 40% of the European Commission's administrative budget, which is the estimate given by Coulmas (1991: 8) for the cost of the EU's policy of separate languages. The ideal of a domestic translation for everyone is perhaps Europe's own form of political correctness. People are paying for that ideal; they will continue to pay for it as long as subsidised culture pushes massive translation programmes into the realm of economically thinkable action. But as globalisation makes text production increasingly multilingual, as its participants learn to work and read multilingually, this viability should not last for long. Something better than translation will have to be developed. 

Why is it that I, as the designer—not innovator—of a nice clean translation, should determine the way readers will use my product? Like the designers of household appliances (read the content of my translation!), I show only smooth surfaces with a few relatively unambiguous buttons to push. The real work is hidden away, beneath the translation, in the languages of the source and in the unexpressed workings of the translator's mind, like the inner machinery that no user of modern household appliances is ever allowed to touch (it's a condition of the guarantees). Despite the theorists, I believe that all translation domesticates to one degree or another, like the machines that are so correctly labelled domestic. But why not show the inner workings? Why not let our readers tinker around in intercultural spaces? Why not go beyond this internal/external, either/or discourse we recognise as the most correct form of translation?

In truth, I much prefer modes of discourse that are able to convey a plurality of possible sources. I would like to subordinate translation the multilingual exchanges that are now a feature of globalisation. I would prefer to receive a letter in Italian, to read it with a little dictionary, and to be able to answer it in English. Or even a letter half in English and half in Italian. We should be learning and using languages as we find them and to the extent that we can make pragmatic calculations sometimes based on limited linguistic competence (this is what we have to do with machine-generated translations anyway). Further, if my Spanish client can report on an English-language text (with possible French-English antecedents), why shouldn't I, as a non-machine, similarly report on the text the client sends me, discursively naming my intervention? Let translation become a mode of subjectively reported speech. Let me be paid for that. And let completely domestic translation become a rare and expensive ideal. Quim Monzó did not translate the American handbook of political correctness; he made it multilingual. I sincerely hope translators, perhaps under another name, will learn to follow similar paths.


Beard, Henry, and Christopher Cerf. The Official Politically Correct Dictionary and Handbook. New York: Vintage Books, 1993.

Coulmas, Florian. "European Integration and the Idea of a National Language". Florian Coulmas ed. A Language Policy for the European Community. Prospects and Quandaries. Berlin and New York: Mouton de Gruyter, 1991. 1-37.

Derrida, Jacques. "Des Tours de Babel." Joseph F. Graham ed. Difference in Translation. Ithica and London: Cornell University Press, 1985. 165-207.

Greimas, A.-J. "Pour une théorie du discours poétique." A. J. Greimas ed. Essais de sémiotique poétique. Paris: Larousse, 1972. 5-24.

Monzó, Quim. "La correcció". No plantasé cap arbre. Barcelona: Quaderns Crema, 1994. 164-165.

Pöchhacker, Franz. "Simultaneous Interpretation: 'Cultural Transfer' or 'Voice-Over Text'?" Mary Snell-Hornby et al. eds Translation Studies. An Interdiscipline. Amsterdam and Phila­delphia: John Benjamins, 1994. 169-178.

Schleiermacher, Friedrich. "Über die verschiedenen Methoden des Übersetzens" (1813). Sämtliche Werke, Dritte Abteilung: Zur Philosophie. Vol. 2. Berlin: Reimer, 1838. 207-245.

Venuti, Lawrence. The Translator's Invisibility. A History of Translation. London and New York: Routledge, 1995.

03-OCT-1995 15:27 A PAG.02



Paper preparted by Monique Dupont

The paper have been elaborted from the data collected between 1991 and 1995 during research conducted in a French company producing small household appliances.

Research in the sociology of technology has revealed how domestic appliances come in the form of black boxes which keep the user ignorant of their funcioning. In fact, the bulk of their technology is physically closed off and inaccessible and only the "outside" parts of the artefacts, those bearing the control buttons, are made available to the user.

Monique Dupont, a través de un minucioso análisis de las entrevistas realizadas durante la investigación completadas con una fina observación nos señala como se realiza esta dicotomía entre el "inside" and the "outside" of the artefact. How the innovators establish the boundary between inside and outside, and determine what must be hidden from the users and what they must have access to and how. It also try to show that this conception of the artefacts is connected to the development of standards and guarantees. Finally, it addresses the question of what the activities are that the users engage in with such artefacts.

Respecto al proceso de cómo se establece esta dicotomía demuestra a través de los objeticos y criterios elegidos por el innovador como se configura el objeto para que su utilización sea fácil para el usario sin que sea preciso que disponga de algún conocimiento sobre su funcionamiento técnico ni sus características técnicas.

El diseño de las formas de la parte interna tendrá como objetivos distintos al de la parte externa. The shapes concerns the adjusting of the parts and subunits of the internal system to one another, their housing in the processor casing and the configuration of the internal environment. Finally shapes are worked on for production and particularly for assembly. The technicians in the method study department seek to achieve simplicity and ergonomics in the assembly sequences with the aim to reduce the time required for assembly and reduce los errores de montaje.

Text 1: Source text received from the client and marked by the translator.



Paper prepared by Monique Dupont

The paper is based on data collected between 1991 and 1995 as part of research carried out in a French company that produces small household appliances. The paper focuses on the design and use of a food processor.

Research in the sociology of technology has revealed how household appliances become black boxes that keep the user unaware of their actual mode of operation. Most of the technology involved in such appliances is physically closed off and inaccessible for the user. Only the outside parts, those bearing the control buttons, are really made available.

Monique Dupont has carried out a detailed analysis of interviews made as part of the research, showing how the dichotomy between the "inside" and the "outside" of the appliance works in practice. The analysis indicates how industrial designers set up the boundary between the inside and the outside, how they determine what must be hidden from the users, what must be made available to the users, and how this relative accessibility should operate. The research also tries to show how the conception of domestic appliances is connected with the development of standards and guarantees, and what actual activities users can engage in with such appliances.

The basic dichotomy is established in terms of the aims and criteria selected by the designer. In general, the object is configured in such a way that the user will find it easy to use and will not require any special technical knowledge about the appliance's features or mode of operation.

The internal shapes are thus designed to fulfil aims that are quite unlike those of the outside. The parts and subunits of the internal system must be adjusted to one another and housed in a processor casing that establishes the overall configuration of the internal environment. The shapes are developed with an eye to production processes, particularly the assembly procedures. Technicians in the method-study department aim for simplicity and ergonomics in the assembly sequences, their main goal being to reduce assembly time and the risk of error.

Text 2: Translation of Text 1, as produced and paid for.

Anthony Pym works on sociological approaches to cross-cultural communication. His publications include Translation and Text Transfer (1992), Epistemological Problems in Translation and its Teaching (1993), Pour une éthique du traducteur (1997) Method in Translation History (1998), Negotiating the Frontier (2000) and The Moving Text: Localization, Translation, and Distribution (forthcoming). He currently teaches at the Universitat Rovira i Virgili, Tarragona, Spain. He can be reached at

[i] A shorter version of this paper has been published as "Multilingual Intertextuality in Translation". Beatriz Penas Ibáñez, ed. The Intertextual Dimension of Discourse. Universidad de Zara­goza, 1996. 207-218.