[This is a revised version, written in 1990, of an article originally written in 1987 and published, unrevised, in META in 1990.]


Brian Mossop


This article considers the role of translating institutions (companies, governments, newspapers, churches, literary publishers) in determining how a translation is done--whether it will be relatively 'literal' or relatively 'free', whether the language will be idiomatic or innovative, whether there will be a change in level of language, and so forth. Attending to the role of the institution in whose service the translator works can I think cast fresh light on certain questions and certain common assumptions about translation: What is a mistranslation? Is the defining characteristic of translation that it preserves meaning during a change of language? Is the function of translation to promote communication?


I will be particularly concerned with the notion that translations should be idiomatic. By 'idiomatic', I mean that the wordings in the target language are statistically normal for the genre of text. Thus if one is translating a recipe, one uses the terms, phraseology, syntactic structures, level of language and layout that are typical of recipes in the target culture: language that will be familiar and immediately comprehensible to the readers.0 I will suggest why the idiomatic approach to translation has come to be used in an institutional setting with which I am familiar--the Canadian Government's translation service--and I will look at some theoretical implications of an 'institutional' understanding of the translation process.


1. The Institutional Nature of Translation


1.1 'Bad Translation' in Newspapers


Unidiomatic language, arising either from translation or from failure to edit quotations of words spoken in English by people whose first language is not English, is frequently found in daily newspapers in Canada. Here are some examples from the Toronto Globe and Mail:


(1) Yesterday, Mr LaSalle denied accusations from callers that he is an opportunist, had taken part in "tractations" with Quebec Justice MInister Marc-Andr Bdard, and is running just to keep the PQ in power. (14 January 1981)


(2) "The 12 jurors listened to the proof, informed themselves of the facts and withstood a hail of emotion. I am very happy for my clients."

(22 October 1984)


(3) But the Quebec people "might not, for reasons many of which are conjunctional, want to take that step. ... It's not, because Quebeckers would not want to pronounce themselves on the sovereignty issue, that it (Quebec) is not a distinct society, or a people". (story on a speech by former Quebec Justice Minister Pierre-Marc Johnson to a University of Toronto audience, date uncertain, punctuation as printed in the The Globe)


(4) His letter quotes Mrs saying in one interview: "There were persons who had neither the capacity nor the willingness to adapt to change. We therefore had to depart ourselves of them". (19 December 1980)


The items I have italicized include some rather extreme examples of unidiomatic language:

(a) non-existent words and expressions: depart ourselves of in (4); tractations in (1),

(b) words which exist but are nonsensical as used: conjunctional in (3),


as well as some less extreme cases:

(c) 'faux amis': proof in (2), a faux ami of French preuve,

(d) direct renderings of French syntax: it's not because...that in (3), a syntactic calque of French ce n'est pas parce que...que.


Professional translators are extremely fond of citing examples of the former, more extreme type (which obviously can give rise to a complete breakdown of communication), but they will also reject the second type, even if it does not present any great difficulty for readers. Journalists will be charged with 'translating the words' rather than rendering the ideas in idiomatic English-- the approach that is now so established as the correct way to translate, in government and business, and in schools of translation, that it appears to be natural. Other ways of translating are seen as the products of unenlightened minds.


I think much can be learned if, instead of simply labelling such examples 'bad translation', we instead ask: why was the translation done this way? In some cases an adequate answer will be that the translator was careless or unqualified, or was not given enough time, or had inadequate documentation. But often it will be necessary to look deeper, because in fact there is nothing natural about idiomatic translation, even if we are considering only so-called 'pragmatic' texts (those not received in the culture as 'literary'). We tend to see as natural that mode of translation which has been selected by the institution within which we work.


Translation is often described as a form of communication between the source-text author or the translator and the reader of the translation, but in actuality the translation of written texts is not primarily a matter of communication between individuals as such, or even individuals as representatives of cultures. When I translate a text, it is not simply me personally conveying to a reader what someone else wrote in French. Translation, like most other forms of writing, takes place within an institutional setting and can only be understood within that setting (Williams 1981 ch. 2; Mossop 1988).


It may appear that the level of formality or technicality of the language in a translation is decided by the translator, after an examination of the level of language in the source text and a determination of who the readers will be and the use to which the translation will be put. But a case can be made that such decisions are to a great extent predetermined by the goals of the institution within which the translator works.


In journalistic translation, some of what is labelled 'bad translation' is indeed a result of bad technique (failing to consider the larger pattern of meaning when working on a particular word or phrase). Bad translation in newspapers is also in part a manifestation of a larger carelessness in journalistic writing today that involves everything from punctuation, spelling, word choice and sentence structure to background research and the presentation of a coherent narrative or argument.0 But not all 'bad translation' in newspapers arises from carelessness or untrained technique. Some of it merely offends the doctrine of idiomatic translation, a doctrine which may not be suited to the goals of the newspaper.


Consider the following passage from former Quebec Premier Ren Lvesque's resignation letter, and the Toronto Globe and Mail's rendering of it:


Je vous saurais gr de transmettre pour moi au conseil national ce simple message: merci du fond du coeur, merci B vous comme B tous ceux et celles qui se reconnatront, et qui n'ont cess depuis tant d'annes de payer de leur personne et de leur portefeuille pour btir, enraciner, maintenir ce projet si sain et dmocratique que nous avons dessin ensemble pour notre peuple.

I would appreciate it if you could transmit for me to the national council this simple message: thank you from the bottom of my heart, thanks to you and to all those, who will recognize themselves, and who have not stopped for so many years paying with their personal lives and with their pocket-books in order to build, estab-lish, maintain this project which is so healthy and democratic and which we have designed together for our people.


Leaving aside the consideration that many readers actually expect translations--especially translations of quotations--to be unidiomatic, one can argue that this translation achieves a particular journalistic goal better than an idiomatic translation0. Certainly it has its flaws, including the failure to render ceux et celles and enraciner, and the confusing expression who will recognize themselves. However the unidiomatic paying with their personal lives is not a flaw despite the odd implication of people paying with their lives. Neither is the rendering of transmettre pour moi by transmit for me, despite the ungrammatical position of for me and the telecommunications connotation of transmit. These renderings contribute to the overall goal of the translation.


That goal, whether consciously formulated by the Globe or not, might be described as that of identifying the voice 'heard' by the reader of the translation0 as a voice external to English Canada. This is to some extent necessary for purely technical reasons: an idiomatic rendering--a translation which made Lvesque sound, incongruously, like the premier of an English-speaking province--would clash with his television English. But more important is the political effect. The unidiomatic Globe translation creates an awareness of the 'otherness' of the text (its origin in a distinct society that has its own goals), while at the same time reducing the legitimacy of Lvesque, partly because readers (Anglophones at any rate) tend to grant the highest level of credibility to words that appear to emanate from their own culture, and partly because unidiomatic translations sound like they are not addressed to the target-language readers (the reader has the impression, instead, of 'overhearing' words written to others).


1.2 Mis(?)translating Freud and the Bible


The notion that translators may intentionally produce for their readers an effect different from that of the source text is now fairly commonplace in writing on 'literary' translation. Unfortunately the discussion is often not descriptive but normative, with the emphasis on condemning error0. Two fairly well-known examples are Bruno Bettelheim's criticism of the Strachey translation of Freud (Bettelheim 1984) and Henri Meschonnic's criticism of Eugene Nida's approach to Bible translation (Meschonnic 1973).


Bettelheim complains that the standard English translation of Freud errs in substituting Greco-Latin 'scientific' words for Freud's everyday German. Why, he asks (p. 53), is das Es rendered as the Id rather than the It (Es being the singular neuter third-person pronoun)? Why is the same approach not used as in the French rendering le Ca?


Let us assume for the sake of argument that Bettelheim's understanding of Freud as a literary humanist rather than a scientist is correct. Do we then have a case of mistranslation--an error in the level of technicality of the language? The answer depends on how the role of translation is seen. One common understanding is that translation preserves meaning except for certain necessary cultural adaptations and the need to select certain aspects of meaning as pertinent (since not all aspects can be preserved; for example, a metaphor may have to be eliminated in order to preserve cognitive meaning). In this view, the Id is a mistranslation because there is no need to change the level of technicality in this case: it could have been preserved in English just as it has been in French.


However in an 'institutional' understanding of translation, translations are seen as preserving meaning only within the limitations of institutional purpose. From the point of view of the translating institution, the 'unnecessary' changes in meaning are as important as the preserved meaning, perhaps even more important.


The Stracheys very likely perceived that Freud could have selected German words of Greek and Latin origin, and that therefore the selection of everyday German words was significant. (Greek and Latin words were certainly used in scientific German, and in many sciences, everyday words have always been used as technical terms in English.) But Freudians in the English-speaking world were consciously trying to construct a 'science of psychoanalysis' as the foundation for paid professional therapeutic practice. They therefore needed a translation that sounded scientific. If Freud was in fact a literary humanist, that was simply irrelevant to their purposes.


Was the level-of-language change perhaps a necessary adaptation to the English-speaking readership of the time? This is certainly a possibility. Only a historical investigation could determine whether the translation would have been 'unreceivable' in the English-speaking world without the change. However we should not assume this to be the case. In the institutional approach to translation being advocated here, the assumption would be that either a 'literary' or a 'scientific' translation would have been receivable (though perhaps by different audiences), and that a conscious choice was made.


The translation was directed at a particular readership. For other possible readerships the result was 'unreadable'. Transcultural communication did indeed take place, but in a very limited way, as Freud's text was 'communicated' from the German-speaking world into a very restricted segment of the English-speaking world.


What we now perceive to be a terminological equivalence (Es=Id) was thus not a discovery but a motivated invention. For a student translating a psychoanalytic text, it is now just a neutral technical matter of terminological research. A translation teacher will be technically correct in calling the It a mistranslation of le Ca, but a good teacher will point out that it is an error because it does not conform to an original institutional decision.


As for Bettelheim's view that the Id is a mistranslation, I would first of all suggest that another term--'non-translation'--be used for cases where a significant source-text feature which could have been preserved has not been, in order to serve some institutional goal. 'Mistranslation' could then be reserved for cases where a significant source-text feature which could have been preserved has not been, for what might be called technical reasons: the translator has been careless, or lacks language, terminology or subject-matter knowledge, or lacks basic translation training so that information is lost through under-translation, or through translation of units smaller than the units within which meaning is generated in the source text.


This latter problem, doubtless the major source of mistranslation, is described in detail in Folkart (1989). A very simple example would be that of translating at word level when meaning is being generated at phrase level, as when exercice d'vacuation is rendered by evacuation exercise instead of fire drill. This sort of technical defect does play some role in the Globe translation discussed in section 1.1, but the lack of idiomaticity is principally a result of the institutional goal of identifying the text as external to the target culture. Thus the Globe's rendering involves what might be called motivated mistranslation--where technically necessary changes (transmettre: pass on, rather than transmit) have not been made, in the service of some institutional goal.


Second, Bettelheim does refer to the institutional purpose of creating a 'science of psychoanalysis' (p. 32), but appears to deem this irrelevant to the question of what constitutes mistranslation. His argument seems to be that changing the level of language is by definition mistranslation. This is problematic because it suggests that translating could somehow be carried on in some pure state uncontaminated by the goals of translating institutions. But if we agree that it cannot, does it follow that we cannot make a judgment about the translation of Freud? No, it simply means that what has to be judged is the institutional goal itself.


Was it a good thing to create a 'science of psychoanalysis', rather than present Freud as a guide to human self-knowledge, as Bettelheim urges? Whose interest did this serve? These are questions of an ethical and political nature, and translators cannot avoid them on the ground that they are performing a merely technical service. The changes they make in meaning may or may not be a good thing.


Bible translation is a field in which there has been a considerable amount of ethical and political debate. In a critique of Eugene Nida's approach to translating the Bible, Meschonnic (1973 pp. 328ff) suggests that as one strips away the original metaphors and the references to Jewish and early Christian cultures, in order to make the Bible understandable to certain modern readers, the Bible ceases to be a literary, cultural, theological and historical text and is reduced to a set of moral lessons.


The Bible-translating institutions with which Nida is associated have been criticized for their translations into both Third-World and 'First-World' languages. Thus Prickett (1979 p. 263) responds to the preface of the Good News Bible ("every effort has been made to use language that is natural, clear, simple and unambiguous") by saying that "religion is not about things that are natural, clear, simple and unambiguous". Meschonnic (1973 pp. 339, 348) says that Nida's method is suited to exporting0 a certain version of American Protestant morality to the Third World. Given this institutional goal, the poetic language, the historical context and even some of the theological content may appear irrelevant to the translator. In the view being proposed here, there is no question of mistranslation in the sense defined above, but there are certainly ethical and political problems: is the evangelical goal a worthy one , and whose interest does it serve?


If it is granted that there is a politics to the method of translation in the cases discussed so far, the question arises whether the institutional approach has a more general relevance. Surely, it will be said, there is no politics to the translation of an accident report, a study of water pollution, or a letter complaining about the way the writer is being treated by the unemployment insurance authorities. To demonstrate that there is indeed such a politics, I want now to look at the Government of Canada as a translating institution.


2. French-to-English Translation by the Canadian Federal Government


The federal government's 'translation doctrine' states that one should render "not the words or the structures of the source text but rather the message or, in other words, the author's intention" (Translation Bureau 1984 p. 3). This statement, prepared for the government's freelance translators, is based on a 1978 document with the interesting institution-oriented title "La traduction au service de l'tat et du pays". And indeed , as we shall see, not only the fact of translation but also the method of translation are designed to serve state policy.


A further document (Translation Bureau 1985 p. I-4) says that one should render "not just words but ideas, so as to convey the message clearly, without keeping slavishly to the expressions and structures chosen by the author". The formulations "not just words but ideas" and "without keeping slavishly" reflect differences of opinion regarding the degree to which the source-text author's wording must be respected. On balance, though, the statement comes down on the side of the modern doctrine of translating 'ideas', conceived as (in the main) culturally neutral and thus expressible idiomatically in the target language.


The government's approach to translation was developed for English-to-French translation. It has been pointed out (for example by Juhel 1982 pp. 55ff) that English is overwhelmingly the translated language in Canada, and French the translating language. Not only is translation in Canada a form of communication that goes mainly in one direction (90% of the federal government's official-language translation work is into French, 10% into English), but also translation forms a relatively high percentage of what Quebeckers read, as opposed to original French writing. It is therefore argued that if the translations are not idiomatic, the French language will cease to be an instrument of cultural identity and, ultimately, political survival.


Translation is seen by French-Canadian language professionals as a major form of writing in French: the distinction between the traducteur and the rdacteur is blurred. The focus of translation into French is 'therapeutic'--the preservation of an authentic French voice. Messages from English Canada are of course conveyed in the translations, but the goal is not to convey information about English Canada and English Canadians.


Perhaps the most important aspect of the government's translation doctrine, then, is that translations are to be 'authentic': "Authenticity is the impression conveyed by a translation that is is not, in fact, a translation, that it was composed in the target language from the outset, that it is an original piece of writing" (Translation Bureau 1984 p. 6).


Now one could argue that authenticity can serve, just as much as its lack, as a vehicle for introducing the culture of English-speaking North America into Quebec. Language and thought are not completely identical, and it is surely possible to express aspects of Anglophone culture in idiomatic French, and thus make them more 'receivable' precisely because they are disguised as French originals. However that may be, authentic (idiomatic) French translations do doubtless perform useful functions such as making it easy for Francophones0 to read income-tax instructions originally drafted in English. But they also convey the false impression that the federal taxation authorities are somehow 'French'.


Idiomatic French translation by its nature conjures up in the mind of the reader a certain image of the state which does not correspond to reality since, given the demographics of Canada, the federal public service and the federal law-making and regulatory agencies are, and are likely to remain, predominantly in the hands of Anglophones even if Francophones are represented in proportion to their numbers (slightly less than a quarter of the population).

Turning now to French-English translation, a certain amount has been written about the impact of different translating methods on one of the main roles of literary translation into English: conveying (or failing to convey) information about French Canada to English Canadians0. But in this article I want to discuss non-literary translation, and I want to focus on the choices made by an institution rather than by individual translators.


Beginning in the 1960s, large numbers of Francophones were attracted to work in federal institutions (in the federal capital Ottawa, and in the Quebec regional offices of federal departments). The policy of bilingualism gave them the right to work in French. It therefore became necessary to translate what they wrote into English. While some of the translations are published for readers outside the public service (international scientific audiences, and the Anglophone minority within Quebec), a very large part of the demand is for the information of Anglophone public servants who do not have an adequate reading knowledge of the other official language, despite language training programs.


Thus one function of French-to-English translation is to act as a substitute for second-language learning, and the other is to assist in a small way with the preservation of French since translation enables Francophone government researchers, for example, to write their articles in French. This material for an international audience needs to be idiomatic in order to gain the confidence of the readership. But it is not immediately clear why the translations prepared for the information of Anglophone public servants, or even the English-Canadian public, should be idiomatic, since there is no need, either inside or outside Quebec, for any special effort to preserve English.


Translation into English would certainly be much easier if one could use innovative, unidiomatic language. A very interesting example of the work entailed by the requirement of idiomaticity is described in an article entitled "Animation and animateur: a translator's nightmare" (Hutcheson & Adshead 1983). The article demonstrates the need to use some fifteen different English words or expressions to convey common meanings of animateur (host, DJ, leader, facilitator, moderator, community worker) and another fifteen or so for animation (community development, chairing, motivation, group training, leadership).


Why, we may ask, cannot the verb animer be rendered by its 'faux ami', the English verb animate ? The answer, from the point of view of the doctrine of idiomatic translation, is that it would hinder communication because English animerhas a much more restricted meaning than its French cognate. According to this view, a unilingual Anglophone will always take a sentence like She animates the meetings to mean She makes the meetings lively. Surely this is false. Given the right context, it could be clear that the intended meaning is She chairs the meetings. If the pure idiomatic theory were true, linguistic innovation would always create chaos in a language community, but clearly it does not.


Now some might be prepared to admit animate on the ground that it fills a lexical gap in English (there is no other word with the generality of French animer). But I want to pose my question in a more general way: why cannot the speech habit many Anglo-Quebeckers have of adopting faux amis and calques of French grammatical structures (regardless of whether they fill a gap) simply be taken over and used--with care0--in the written English of federal translators? In other words, why cannot the translating institution (the Translation Bureau, acting on behalf of the federal government) produce translations that contain some of the features of the Globe's rendering of the Lvesque letter?


The answer, I think, lies in the ideology of the federal government's policy of bilingualism. The policy reflects a liberal/cosmopolitan outlook that is often associated with the former prime minister Pierre Trudeau. In one common understanding of his thought, the specificity and separateness of local and national communities are a thing of the past. What is important is that every individual have an equal opportunity to participate in the public worlds of law, business, administration, science and technology. In the Canadian context, this means equality regardless of mother tongue.


The purpose of translation is thus to let scientific-technical, adminstrative, legal and commercial messages pass between French and English individuals, not to create bridges between the two communities. Idiomatic French-English translation in the public service is well suited to this purpose because it erases the 'local colour' of the source text (the marks within the text of its author's French origin), and its natural-sounding language lets the 'universal' administrative or technical content come across more easily to the unilingual anglophone reader.


Unfortunately, since the language of any idiomatic translation is necessarily a local language--not some truly universal translation-language made up of elements of all languages--the 'universality' of the result takes a strange form: it appears to the readers of the translations that their language, English, is not local but universal. French Canadians are not really a distinct people: they sound just like English Canadians!


Idiomatic translation is thus a paradox. On the one hand, it promotes communication in the sense that Francophone public servants can work in French yet still get their administrative or technical messages across to their Anglophone colleagues in a very readable manner. On the other hand, it hinders communication because it makes the Francophone presence vanish: the reader has the impression of being addressed by another Anglophone despite the signature at the end of the document. The reader may be intellectually aware that the originator is a Qubcois, but this knowledge will be somewhat abstract, especially if the reader has never met the author or lived in the author's home milieu. Once again, as so often in Canadian history, the two solitudes fail to touch.


Suppose translators could write unidiomatically, with carefully selected faux amis and syntactic calques (like transmit for me in the case of the Lvesque letter). Readers would then be reminded of the 'otherness' of the person who wrote the document. Or translators could go even further and try to convey something of the nature of that otherness. For example, in administrative texts written in Quebec there is sometimes a certain chummy informality that is combined in varying degrees (and sometimes incongruously) with the elegant (not to say precious) formality associated with texts from France. But in the prevailing doctrine of translation, all such information about the writer is not deemed 'pertinent' for administrative/ technical texts.


I have argued that a slightly gallicized French would be perfectly understandable to unilingual Anglophones. Of course, one could be even more confident about using animation in the French sense if widespread learning of French were underway and translation were functioning as a supplement to rather than a substitute for language learning.0 Certain types of unidiomatic translation would then be receivable as a linguistic manifestation of a broader inter-community dialogue ranging far beyond the present tiny group of proficient bilinguals. But the focus of the federal bilingualism policy is on individual equality rather than inter-community dialogue.


Indeed, in the actually existing situation of ongoing antagonism between Quebec and English Canada, arising from conflicting visions of what Canada is, unidiomatic translation could even be interpreted within English Canada as evidence of French 'infecting' English, and could give rise to suspicions of the sort whose most extreme form is regular manifested in cries about an imminent 'French takeover' of the public service or the federal government.


To sum up: An idiomatic translation, even if it is technically correct and avoids mistranslation, can nevertheless convey an institutional meaning, in this case something like 'English and French Canadians are essentially the same.' For further discussion of problems arising from the idiomatic approach to the translation of pragmatic texts, see Mossop (1989).


3. Translating Institutions and Translating Theory


I would like to conclude by pointing out some implications for translation theory of the institutional nature of translation.


3.1 Equivalence, pertinence, communication


Maurice Pergnier writes that:


Une approche thorique de la traduction doit...faire porter l'accent sur la dfinition de ce qui est B traduire, c'est-B-dire le message. ...Seul, en effet, vise B rester constant dans le changement de langue qui constitue la traduction, le message en tant que contenu d'information. (1980 p. 27)

Il n'est pas douteux que toute traduction vise l'quivalence, au niveau du contenu informatif, d'un texte traduit avec un texte original. (p. 47)


Surely it not the case that every translation aims at equivalence. Translators as individuals may well believe they are seeking equivalence (seen either as equivalence of effect, or as preserving functionally pertinent 'information'), but this is beside the point. Institutions frequently do not seek equivalence, and even if they do, that is as much a choice as the decision not to. It is the institution that decides which aspects of meaning are pertinent; it is the institution that decides whether to disguise the translation as an original; it is the institution that decides on the degree to which the voice of the source text will be allowed to address the readers of the translation, and the degree of adaptation to the target culture or the particular readership.


A translation theorist or teacher may advocate deriving the method of translation from certain features of the source text, or the target culture, or the readership, but the translator can only implement the advocated method to the extent that it does not conflict with the institution's goals.


Pergnier is certainly right to want translation theory to focus on the message being directed at the reader, with its 'information' (in the broad sense of this term that includes cognitive, expressive, social and other aspects). Communicative and sociolinguistic models of translation have certainly been useful in pointing to the fact that the real activity of translation is less a matter of relating languages than relating users of language.


However the usual communicative models, in which information moves along a more or less complex route, from a sender in culture X to a receiver in culture Y, are too abstract. The social characteristics of the various parties involved tend to be overlooked:


- The reader: Non-communication of the original 'information' may occur not for technical reasons (eg from the common habit of undertranslation; from failure to do research; from failure to revise) but because the translating institution wishes only certain information to be passed on to certain readers in the target culture, with other readers excluded. Different readerships constitute different subcultures. Translation theorists often speak of 'the' target culture but perhaps it would be better to say that the translating institution, in its approach to such matters as text selection, language-guardianship and information selection, responds in different ways to the real or perceived needs of each of the different subcultures that make up a society.


- The translator: Communicative models bear the marks of the communications-engineering theory from which they emerged, in that the translator appears as a benign, technical medium (like a relay station in a telecommunications network). In actuality, there is no neutral, self-effacing strategy available to the translator, who must select from among many options in order to meet institutional goals.


Communicative approaches to translation theory also betray their 'engineering' origins in their focus on the effective and efficient transmission of information, on 'getting the message across' to the targetted readers:


- Efficiency: Translations are seen as efficient if there is minimal loss of 'pertinent' information. Pertinence is sometimes described in terms of the reader's needs, and sometimes as if it were objectively determinable (if the text is a memo on acid rain, then only the administrative and scientific content is pertinent). But communicative theories never describe pertinence in terms of "what we want to get across to the reader". In an institutional approach, the quantity of pertinent information which is lost is of no interest in itself. What is of interest is certainspecific losses of information, whether or not the institution defines them as pertinent: the question is whether they are significant from the viewpoint of the source text.


- Effectiveness: In communicative theories, translations are seen as effective if they are idiomatic. In an institutional approach, idiomaticity is seen as a choice. A non-idiomatic translation can be very effective in achieving an institution's goals. A central purpose of an institutional theory of translation is precisely to explain why translations have been translated one way rather than another.


- Getting the message across: Theories of translation as a form of communication focus on the target reader. The translator is seen as seeking the most receivable (ie the most idiomatic) way of expressing the pertinent information which he or she has ferreted out of the source text. Peter Newmark has been making the point for several years that communicative models of translation privilege receivability over accuracy whenever there is a conflict between the two. Thus they can serve as models for only some of the institutional practices that are to be found in the history of translation.


To sum up: Unlike a theory that focusses on the transmission of messages, an institutional theory will focus on the moment of production of the translation--the way in which the translation is designed so that it will later give rise to certain key meanings in the minds of the readers.


3.2 Translation theory, institutional practice


Translating institutions do not just produce translations; sometimes they also produce translation theories. These theories may be contained within statements by institutions of their translation doctrine (instructions for translators), or they may be elaborated, perhaps for a set of similar institutions, in independent writings by translators or by representatives of various academic disciplines.


There is as yet no general theory of translation. What exists are restricted theories that can be seen to be rooted in the activities of particular translating institutions: conference interpretation involving the languages of the European Community; Russian-English machine translation of scientific texts for the U.S. armed forces; translation of the Bible from Ancient Greek and Hebrew into native languages of Latin America, and so on.


It is often said that translation theory and practice need to be brought closer together. But have they ever been apart? Indeed, are they not too closely intertwined? Much of what we call translation theory consists in more or less elaborate and more or less abstract formulations of the practices of the institutions with which the theoreticians are associated. Valuable as such formulations are, they inevitably tend toward advocacy of institutional translation methods. They also by and large lack critical distance from institutional practices, and this inevitably reduces the insight they achieve.


- The theory that translating literary texts and translating pragmatic texts are two fundamentally different operations is perhaps less a reflection of differences between text types than a reflection of the fact that literary and non-literary translators work within starkly different institutional arrangements.


- The limitations of purely linguistic theories of translation can now be seen to lie, at least in part, in the purely linguistic nature of the machine translation experiments alongside which these theories arose.


- Nida's theory that the translation of 'meaning' can be separated from the translation of 'form' can be seen as a counterpart to a particular practice of Bible translation (Simon 1989).


- A case can be made that Danica Seleskovitch's well-known interpretive theory of translation is more applicable to conference interpretation than to the translation of written texts.


- The theory of translation as rdaction (Flamand 1983), or as an exercise in comparative stylistics (Vinay & Darbelnet 1958), can be seen as a reflection of a central goal of the translating institutions of French Canada--creating or preserving an authentic French voice.


- Lastly, the theory that translation is communication is a prevalent one because the biggest translating institutions at present are concerned precisely with 'getting the message across' effectively and efficiently.



Concluding Remark


By making the translating institution itself the central focus, the approach to translation theory suggested here aims to achieve critical distance from any one institution, as well as a certain amount of generality (by encompassing all institutions). However an institutional approach as sketched here could not by itself constitute a satisfactory theory of translation.


First, it does not deal with the success of the institution in achieving its goals. That is, it does not cover the activity of the reader interpreting the translation (readers, unlike writers, are generally not acting on behalf of institutions). And it says nothing about the reception of the translation in the target culture (for example, if the text is known to be a translation, it may be judged against translational norms of the target culture).


Second, the institutional approach is essentially about what I have called non-translation, and so it would have to be supplemented by a general theory of mistranslation, such as is provided (though she does not present it as such) by Barbara Folkart (1989).


The distinction between mistranslation and non-translation I hope to elaborate at a later time. Essentially, avoiding mistranslation means conveying to the reader the pattern of significant meanings borne in the linguistic and rhetorical structures of the source text, with significance defined from the vantage point of that text. This can almost never be accomplished through use of lexical and grammatical material that directly corresponds to the material of the source text. Varying degrees of transformation are called for, including the correction of careless writing in the source text. (But an idiomatic, easily readable translation will constitute mistranslation if the presence of innovative or difficult language in the source text is significant.)


Sometimes, significant meanings are not conveyable, as when a religious concept or a genre is not known to the target culture. It is then necessary to resort to what might be called 'necessary mistranslation': the use of 'equivalents' or explanatory overtranslations (presented as such or not). Beyond this, one finds non-translation: transformations are made that are not necessary to conveying the pattern of significant meanings of the source text. Non-translation may sometimes be unavoidable, given a particular targetted readership, but this does not account for all of what goes on under the heading 'adaptation to the readership'. Institutions may understand their non-translation practices as adaptation for the reader, but what is often happening is that the institution is conveying its own meanings to the reader rather than the meanings of the source text.




BETTELHEIM, B. ([1982], 1984): Freud and Man's Soul, New York, Random.

FLAMAND, J. (1983): Ecrire et traduire, Ottawa, Editions du Vermillon.

FOLKART, B. (1989): "Translation and the Arrow of Time", in TTR vol. 2 nM1, pp. 19-50.

HERMANS, T. (1985): The Manipulation of Literature: Studies in Literary Translation, London, Croom Helm.

HUTCHESON, H. & M. Adshead (1983): "'Animation' and 'animateur' - a translator's nightmare", in Terminology Update vol. 16 nM1, pp. 1-6.

JUHEL, D. (1982): Bilinguisme et traduction au Canada: Rle sociolinguistique du traducteur, Quebec City, Centre int'l du recherche sur le bilinguisme.

MESCHONNIC, H. (1973): Pour la potique II, Paris, Gallimard.

MEZEI, K. (1988): "Speaking White: Literary Translation as a Vehicle of Assimilation in Quebec", in Canadian Literature 117 (Summer), pp. 11-23.

MOSSOP, B. (1987): "Who is Addressing Us When We Read a Translation?", in TextContText vol. 2 nM1, pp. 1-22.

MOSSOP, B. (1988): "Translating institutions - a missing factor in translation theory", in TTR vol. 1 nM 2, pp. 65-71.

MOSSOP, B. (1989): "'Write Idiomatically and Translate Ideas not Words': Three Defects of the Prevailing Doctrine of Translation", in C. Sguinot (ed.), The Translation Process, Toronto, H.G. Publications, School of Translation, York University, pp. 7-20.

NEWMARK, P. (1988): A Textbook of Translation, London, Prentice Hall.

PERGNIER, M. ([1978], 1980): Les fondements sociolinguistiques de la traduction, Paris, Honor Champion.

PRICKETT, S. (1979): "What do the translators think they are up to?" in New Universities Quarterly Spring 1979, pp. 257-268.

SGUINOT, C. (1988): "Translating the ideology of science: the example of the work of Alfred Tomatis", in TTR vol. 1 nM1, pp. 103-112.

SIMON, S. (1987): "Dlivrer la Bible: La Thorie d'EugPne Nida" in META 32(4), pp. 429-437.

TRANSLATION BUREAU (1984): Contractor's Guide - Translation, Ottawa, Department of the Secretary of State.

TRANSLATION BUREAU (1985): Reviser's Handbook, Ottawa, Supply and

Services Canada.

VINAY, J.-P. and J. Darbelnet (1958): Stylistique compare du franais et de l'anglais, Paris, Beauchemin.

WILLIAMS, R. (1981): Culture, Glasgow, Fontana.



0.1. This is roughly what Newmark (1988 p 24) calls 'natural' language. He uses 'idiomatic' to refer to translations that use catchy language, as when fonds qui seraient utiliss specifiquement pour is rendered as funds that would be earmarked for. I am not using the term 'idiomatic' in this narrower sense, or in the narrower sense of Pergnier (1978 ch. VIII), according to which a translation is idiomatic if, out of all the word combinations which are structurally allowed by the grammar of the target language, it used only those which are in fact habitually employed in the language as a whole (eg exercice d'vacuation comes out as fire drill not evacuation exercise). An idiomatic translation, in the sense of this article, is one whose language is idiomatic in the broader sense of being appropriate to a genre. Thus, to take a classic example, nasty dog (for chien mchant) would be perfectly idiomatic English in the narrower sense, but is not appropriate to the genre of lawn signs.

0.2. Here the argument moves into the socio-economics of writing. Daily papers do not demand or provide translation training because even bad writing enables them to carry out their functions. In earlier times, language quality had to be higher because the articles had an important informative and analytical function and were more central than the pictures, headlines and ads. But now, the news-reporting, news-analyzing and language-regulating functions of daily papers (unlike certain more specialized publications) are less important than the advertising function and the ideological function (calling on the reader to 'defend the West' or 'oppose Terrorism and Drugs' or 'support the Family'). Bad writing serves to fill up the space between ads with something vaguely intelligible, gives the reader something to pass the time with on the way to work, and is easy for the journalist to produce quickly.

0.3. Here is a possible idiomatic rendering of the passage: " I would appreciate it if you could pass on this message to the Provincial Council for me. I would like to say thank you from the bottom of my heart. My thanks to you personally and to all those men and women--they know who they are--who have been making personal and financial sacrifices for so many years in order to build and maintain a healthy, democratic path to the future, a path which together we have laid out for the people of Quebec." Note how the syntax has been normalized (build and maintain instead of build, establish, maintain), the emotional impact has been flattened (cf our people, so healthy in the Globe), and the politics has been altered by use of the institutional equivalent Provincial Council. (Non-Canadian readers should realize that French-Canadians regularly use the word 'national' in reference to Quebec institutions, whereas English-Canadians normally use it only to refer to federal or pan-Canadian institutions. National Assembly is used as the English name of the Quebec legislature, however--an interesting example of an unidiomatic translation that has become conventionalized; English-Canadians would normally refer to this is a provincial legislature.)

0.4. In selecting a method of translation, the institution, through its translators, creates a distinctive 'voice', and the readers later construct for themselves an image of the voice which they sense as 'addressing' them (Mossop 1987). The institution can decide whether or not it is going to make the readers aware of the fact that they are reading a translation, by creating an alien-sounding or a familiar voice. Even in a text with an identified author who is known by the targetted readers to be outside their culture, the institution can seek, by its translation method, to play up this outsideness, or play it down (as is illustrated by the two translations of the Lvesque letter given above).

0.5. See Hermans (1985) for a descriptive approach which does not seek to "provide guidelines for the next translation to be made" or "pass judgement on any number of existing ones" but "takes the translated text as it is and tries to determine the various factors that may account for its particular nature", looking at the "constraints and assumptions that may have influenced the method of translating" (p. 12-13)

0.6. A distinction worthy of futher study is that between institutions which export translations and those which import them. The Stracheys were importing Freud to the English-speaking world, whereas in the case of Nida, the Biblical text is being exported.


7.The noun 'Francophone' is used here, following Canadian usage, to mean someone whose first language, or whose sole or main language of everyday life, is French. It does not mean 'someone who can speak Fench' (no matter how well). In this article, 'Francophone', 'French-Canadian', and 'French-speaking' are used interchangeably. 'Anglophone' is to be interpreted in like manner.

0.8. See Mezei (1988). Sguinot (1988) is a rare example of a discussion of the impact of different methods of translating non-literary (in this case scientific) texts.

0.9. Certainly readers of translations need protection against unthinking 'translationese'. English now has more non-native users worldwide than native users, and these non-native users--often acting on behalf of large corporations and governments--are generating large amounts of unidiomatic English that is highly reminiscent of translationese. In this sense, English does indeed need protection.

0.10. It is quite possible that there is a trend underway toward increased English unilingualism outside Quebec, despite English-speaking children going to 'French immersion' schools. According to the 1989 Annual Report of the Commissioner of Official Languages, only 6% of Anglophones outside Quebec (as compared to 59% inside Quebec) reported to the census-takers in 1986 that they could carry on a conversation in French--up from 3.5% in 1971 but still a very small percentage, and there is no way of knowing from the census what level of proficiency respondents are claiming. About 6% of all elementary and secondary students outside Quebec were enrolled in French immersion in 1988/89, up from 0.5% in 77/78. This may lead to a modest increase in the small elite of Anglophones with high proficiency in French. But it is not clear at all whether there is any trend toward more individuals having at least a small degree of proficiency. Participation in French instruction at the elementary level outside Quebec stood at 56.2% in 88/89, up from 41.6% in 77/78, but participation at the secondary level was lower at 46.7%, and up only slightly from 40% in 77/78. Taking these figures, along with the fact that far fewer post-secondary institutions in English Canada require a second language than had such a requirement before the bilingualism policy was instituted in the late 1960s, and there is certainly no clear prospect that translation might become a supplement to language learning.