TRANSLATION RULES ARE ETHICAL DECISIONS

 
Ethics is a professional concern

 
Ethical questions concern translation on two levels. On the one hand, tired repetitions of traduttore traditore presuppose some kind of ideal loyalty to a source text, author or sender, often pitted against similar loyalty to a receiving language, culture or receiver. On the other, codes of ethics are written for the control of translation as a profession, regulating the translator's relations with other translators, with clients and with questions like official secrets. These are two very different levels. In the first case, the ideal translator remains an invisible linguistic figure, corresponding to no I-here-now. In terms of the profession, however, the ideal translator is a juridical and fiscal entity who, according to most contemporary ethical codes, should have paratextual and extra-textual presence as the partly responsible source of translated texts. The implicit anonymity of the first level would seem to be overridden by a call to explicit professional presence on the second.

Historically, this difference in levels can be projected as a long process going from politically enslaved anonymity to independently professional practice, a process that has been accompanied by the progressive development and justification of translational ethics. That is, to the extent that translators have slowly transformed their anonymity into a professional identity, they have been able partly to develop a professional ethics.

But can it simply be assumed that all individual translators have become equally professional? Have they really gained sufficient authority to develop their own ethics? And if so, where did this authority come from, and in whose interests should it be used?

Approaches to translational ethics mostly fail to address such questions because they are almost exclusively focused on the practice of the abstract individual translator. Experts thus set about writing rules on the model of "when in situation A, take action B", hoping that inexpert individuals will conform to an ethically unquestioned and ostensibly unquestionable norm. But if no isolated individual has ever gained enough authority to formulate and apply this kind of rule - truly isolated individuals tend to be called traitors - , how can a realistic code of behaviour possibly be formulated in such terms? The starting point for translational ethics must be the professional group, not the lone hand.

This means that, since the historical development of the profession concerns a collectivity - translators as a social group - , it is misleading to formulate translation rules as simple precepts for individuals who might be morally right or corrigibly wrong. The essential problem of translational ethics is not how to translate in any given situation, but who may decide how to translate.

Partial answers to this question can be gleaned from the long march from slavery to professionalism. Translators became professional, but they did not do so spontaneously or individually. They passed through several intermediary stages, recountable in terms of political models and arabesque arguments concerning inspiration, individualism, divided loyalty and the apparently neutral use of natural languages.

Let us review just a few of these models and arguments.
 

Translators are rarely above suspicion

The first written references to translation did not mention translators at all: Kurz (1985) notes that in sixth-dynasty Egypt (2423 - 2263 BC) one of the official titles of the Princes of Elephantine was "overseers of dragomans", with nothing said of the dragomans or interpreters-guides themselves, who were presumably controlled nobodies. Similarly, the Biblical references listed by Nida mention not the history of translators, but the history of the kings, princes and priests for whom translations were carried out. Or again, although the Rosetta stone has the priests of Memphis conferring divine honours on Ptolemy V, it mentions no honours for its hidden translators. Performatives belong to kings, princes and priests; translations remain as anonymous as the overwhelming majority of those who, for at least four thousand years, have sacrificed their extra-textual identities in the interests of one kind of equivalence or another.

It is not particularly scandalous that few translators have been kings, princes or priests. There is even a certain pride to be taken in the fact that political and moral authorities have had to trust the knowledge conveyed by their translating servants. But how might the prince know that a particular translator is worthy of trust? It would be foolish to suggest that all translators are equally competent, that their fidelity corresponds automatically to what they are paid, or that their loyalty is beyond doubt. Some kind of extra-textual support is ultimately necessary. Perhaps the prince's confidence is based on a diploma from a specialised translation institute, references from previous employers, comparisons with other translators, or even on what the individual translator is able to say about the practice of translating, since theorisation is itself a mode of professional self-defence.

Although these extra-textual phenomena indicate the nominal existence of a translating individual, they should not contradict translational equivalence, since their very function is to provide support for the acceptance of equivalence. In theory, translators can only be accorded the absolute anonymity of equivalence when they can be trusted absolutely. Unfortunately, in practice, the principle of anonymity is mostly relative, since the extra-textual indicators are themselves not equally trustworthy - foreign diplomas and references are easy to forge - and intellectuals tend to have too many personal opinions anyway. The problem remains.

Now, which extra-textual factors are most likely to be trustworthy? Traditional authority mechanisms tend to subordinate meritocratic indicators to factors like birthright and group identification. Moreover, since politics is largely the discursive elaboration of an inclusive and exclusive "we", the person who is to be trusted should ideally be included in the same first-person plural as the king, prince or priest distributing authority. In Spain, sworn translators are authorised by the Ministry of Foreign Affairs and have their certificates issued in the name of the king. The translator should identify with the central authoritative figure. The truly trustworthy translator should ideally be "one of us".

The only problem with this traditional guideline is that, since translation concerns exchange between different cultures, there is always at least one double "we" involved. In theory, the translator's loyalty could always be to the foreign prince or trader, the potential enemy or thief. In practice, ostensible allegiance almost always goes one way or the other, perhaps according to which prince is paying the most for translational services. Yet objectively mercenary behaviour is never a guarantee of absolute loyalty; clever princes require some form of subjective allegiance as well. Nor are translators able to pretend to the strict disinterest of other kinds of mercenaries: a hired gun can fight in any battle whatsoever, independently of subjective identification, but translators, like spies, can only be employed in situations between cultures of which they have substantial personal experience; they can only be employed in places where their loyalty is open to question. The Princes of Elephantine were not just overseeing linguistic vraisemblance; they also had to trust dragomans as guides, as former nomads and border-dwellers who knew the foreign lands to be crossed and thus partly belonged to a foreign "we". Translators are habitually from diglossic border regions, from families of mixed background, from situations where language loyalty contradicts national frontiers (Weinreich 1953, 100ff). Since they have by definition incorporated elements of multiple subjective identification, since they speak the language and know the lands of the foreigner, since they share the cultural references of real or potential enemies, translators will never be able to convince sceptical princes that their inner subjective identification is entirely one-sided. This was indeed the case in Elephantine. Which is probably why sixth-dynasty pharaohs decided to make "overseers of dragomans" a princely title in itself, placing authority in a home-grown hierarchy above anonymous translational work, to ensure the success of their commercial expeditions to the south.

Sovereign hierarchisation was a partial solution to the problem. But not the only one.
 

Inspiration may have come to isolated cells on Pharos

A simple solution to the problem of divided loyalty is for the translator to seek or invent a higher or even transcendent source of authority. In the case of religious texts, this could mean divine inspiration. And for profane texts, there could be direct inspiration from the author, either through literary metempsychosis, as claimed by Chapman (T. R. Steiner 1975, 23), or perhaps through "Sortes Vergilianæ", as W. F. Jackson Knight intimated with respect to his translation of the Aeneid (1956, 23-24). In both the religious and profane domains, the source of authority is an author who is physically absent from competing worldly princes. Authority thus avoids the political "we" by becoming a fact of individual authorship. Inspired authority may then ensue from the assumption of inspired individual authorship.

If inspiration is best left to the inspired, its sociological import can nevertheless be described on the basis of how its authority is attested. The model example here is the miraculous translation of the Greek Bible, usefully recounted by Nida (1964, 26-27). The Septuagint is said to have been translated by seventy-two rabbis - six from each of the twelve tribes of Israel - who, working in groups of two and in isolated cells, translated the Old Testament with such divine inspiration that the resulting TTs were all absolutely identical.

The legend is at least eloquent on the subject of authority. Rather than compare a Greek TT with a Hebrew ST, this exemplary professional association of translators insisted that the only valid test was to compare one Greek TT with another, since only the translators themselves were competent to judge the meaning of the original text. Regardless of whatever tricks or illusions were used to ensure absolute equivalence between TTs, the very fact of such equivalence could then in turn explain away any deviance between the TT and the ST, including outright deletions and additions, as indeed it did for Augustine:

"[The Spirit] with divine authority could say through the translators something different from what he had said through the original prophets [...]. We will conclude, in the case of something in the Hebrew which is missing in the Septuagint, that the Spirit elected to say this by the lips of the original prophets and not by the lips of their translators. Conversely, in the case of something present in the Septuagint and missing in the Hebrew, we will conclude that the Spirit chose to say this particular thing by the seventy rather than by the original prophets, thus [...] all of them were inspired." (City of God 18.43; cited by Nida 1964, 26)

It would seem that, for as long as translators agree amongst themselves, for as long as they have a solid professional association, their divine inspiration can even claim infallibility.

The rabbis merit two further observations. First, the political problem of their "we" was overcome not only by their being drawn from all of the twelve tribes of Israel, but also by their status as something more than mere dragomans: they were all religious teachers, with a claim to extra-translational authority even before their miraculous texts were compared. Their professional association thus drew its initial authority from contiguous social structures. Second, proof of their inspiration does not necessarily depend on belief in their prior isolation. After all, if they had not been isolated, would it have been any less remarkable that they could all agree on one definitive TT? It should be remembered that Quine's indeterminism is based on firm belief that "one translator would reject another's translation" (1969, 296-97), and yet seventy-two rabbis would appear to have reached a common accord. This is indeed miraculous determinism! And if they could all agree on a TT, surely it would have been be far easier for them all to agree to say they had been isolated and inspired.

Historically, the structure of divine inspiration functions as a double step towards professionalisation. First, it enables translators to draw authority by contiguity with authors and their representatives, be they divine or otherwise. Second, it demonstrates the usefulness of burying differences and seeking safety in numbers. But more importantly, it offers a mode of professionalisation that does not depend on established sovereign hierarchies.
 

Nec translatores debent esse soli

A slightly different solution to the problem of the political "we" can be observed in the background of major international summits. When two presidents meet in different languages, there are usually two interpreters on hand for the necessary shadowing. Why two? Since summit-level interpreters are competent in two-way communication, only one should be strictly necessary from the point of view of linguistic skills. Yet neither president wants their words filtered through a foreign mouthpiece. In practice, two interpreters are necessary so that each can function as a check on the equivalence produced by the other: one is presumed to be "ours", the other is "theirs". In this way, the problem of trust is partly solved by making specially selected translators their own mutual overseers.

Heck (1931, 5ff) describes how in the early Middle Ages two translators often worked on the one text, the first producing a literal version, the second then adjusting the literalism to the stylistic requirements of the target language. In principle, this double translation allows for a checking of loyalty in both directions. The "ours" and "theirs" of summit translators becomes an internal fact of the translating process itself, without any recourse to hierarchies or claims of divine inspiration.

Extended forms of intra-professional checking have long been used in Bible translating. When Luther states that "a false Christian or a person with a sectarian spirit cannot faithfully translate the Scriptures" (in Nida 1964, 16, 152), his insistence on divine grace functions not only as a safeguard against potential treason but also as a way of making translators work together. Translators should belong both to a "we" defined by potential inspiration - which should never be rejected out of hand - and to another "we" based on mutual presence. Instead of legendary isolation in cells, Luther insists on obligatory conversation: "Nec translatores debent esse soli, denn eim einigen fallen nicht allzeit gut et propria verba zu... (Tischreden, in Kloepfer 1967, 36); the right words in the right order do not always occur to the solitary translator. Luther thereby condemned his translators to painstaking committees - "We often spent a fortnight or three or four weeks questioning a single word" (Sendbrief, in Störig 1963, 32) - , but he made sure that the word finally found was suited to the intersubjective principles of equivalence for the people.

Authority, in this context, is presumably recognised within the group of translators as they work. It is something that exists in the relation between translators; it depends on neither divine authors nor worldly princes, since it has effectively been transferred to the level of dialogue and mutual recognition. The prohibition of solitude thus marks a further major step in the professionalisation of translators.

But does Luther's "we" entirely avoid the problem of political overseers? After all, he himself assumed enough authority to lay down the principle of group collaboration, which in turn worked to the benefit of his desire to break with literalist tradition. Yet the authoritarian position is in this case not quite pharaonic. Since Luther was himself one of the translators, his "we" was also an extension of the intra-professional model. Group work and authority were contained within the social unit formed by translators as a professional organisation, independently of princes, popes and cardinals.

A similar "we" can be found in introductions to modern Bible translations. The New International Version, for example, is presented as the work of "over a hundred scholars" selected and controlled by a complex system of committees which "helped to safeguard the translation from sectarian bias" (NIV, v). However, as in Luther, the work of the overseeing committees was based on subjective identification of the translators themselves as a group of believers:

"...the translators were united in their commitment to the authority and infallibility of the Bible as God's Word in written form." (NIV, vi)

The structure of divine inspiration still exists in this insistence on faith as a collective professional qualification.
 

Isolated inspiration is also regulated

If Luther's communal precept can be seen as an extension of Septuagint inspiration and traced through to contemporary Bible translation procedures, there is perhaps no less historical continuity to be found in the Hieronymian aesthetic of individual inspiration, against which Luther's committee procedures were partly formulated (Kloepfer 1967, 36-38). Should the ideal translator work alone or in a group? It is no doubt possible to find at least one great individual translator for every great professional group of translators, and no amount of historical or statistical argument will win the day. In the field of Bible translation, solitary scholars like Emile Osty believe that their individuality protects them from "fantasy and lack of consistence, a reproach warranted by so many versions produced by a plurality of authors" (1973, 9). Newmark has typically raised this individualism to the level of an eternal principle: "...a first-rate translation must be written by one person..." (1981, 158). But how serious are such claims to inspired individuality?

Especially in the field of literary texts, preferences for the collectiveness or singularity of translators tend to follow general ideas about ideal authorship and the strength of the corresponding property bond between author and text. If an age believes that all great authors are individuals, its preferred translators will also tend to be individuals. Hegel believed that the greatness of the Iliad proved the individuality of Homer (1835, II, 1048-1098), despite Wolf's early exposition of the Homeric epics as a translational series. Newmark would no doubt have agreed. If authoritarian individualism has its preferred workplaces, so much the worse for the facts.

Examples of relative correspondence between authors' and translators' anonymity:

It seems pragmatically correct that "authorless" genres like information brochures should not name their translators, and that strongly authored texts like poems should always give the translator's name. But there are many genres where authorship bonds are weak and, although ostensible sources are cited, it is very difficult to have a translator's name printed. In Spain, where I nevertheless try to ensure such naming, this difficulty concerns political and economic texts, children's literature, encyclopedia articles, how-to-paint books and almost anything else that, in commercial passages from publisher to publisher, can undergo any number of uncontrolled rewrites. In these cases, authors are often named quite independently of their knowledge or desire, and translators tend to become become as anonymous as the unscrupulous intermediary agents they work for. For this reason alone, it would be difficult to base translational ethics on criteria of authorship. But there are other reasons as well.

The historical projection of the relation between translational practice and authorship ideologies demands more investigation than can be undertaken here. But it should nevertheless be clear that this relation does not imply that individual translators can or should assume the status of individual authors. The relation pertinent to professional ethics exists on the level of two collectivities. If translators are able to use authorship ideology in order to absorb authority from more prestigious social institutions like national literatures, they do so in a way that affects the professional as a whole, in a collective space logically anterior to individual authority. There can be no question of analysing the authority of the translator's profession as the consequence or sum of individually authorised translators, as if there had never been any Princes of Elephantine, as if all translators were always already priests. The historical order of models was in fact quite the reverse: just as primitive communism and exploitive hierarchies preceded capitalist individualism, just as the obligations of fixed verse forms preceded the liberty of literary prose, so collective professionalisation historically preceded the authority of the individual translator. Exceptions to this anteriority are merely exceptional. In general, only once translators had a positive collective status could they set about imitating authorship structures, assuming individual authority.

Radical individualism should thus be taken with a large grain of salt. If works like Luther's Bible and the King James Version can legitimately be criticised for fantasy and inconsistence, it nevertheless seems difficult to classify them as automatically second-rate because of group authorship. Such prejudice should quietly be absorbed by the more global principle that the collective profession provided the conditions necessary for the rise of the authoritative individual.
 

There can be no ethics of linguistic neutrality

If the individual translator's ethical decisions concern cases of loyalty divided between ST and TT cultures, an easy way to solve such dilemmas is presumably to get rid of the figure of the individual translator, thereby getting rid of the subjectivity originally called upon to make such decisions. Indeed, the very existence of authoritative professional groups suggests that there is a certain strategic virtue to be found in the semantic absence of individualism, in tacit retreat from situations of individual choice. By complying with group decisions, the individual's equivalence might be supposed to escape partisan bias and attract substantial social guarantees.

In accordance with this view, an ethics of anonymity would have the translator remain an essentially passive entity with no identity beyond professional unanimity. Translators might perhaps work, but they should not be seen to work. In fact, if one is to believe a standard contemporary ethical code, the translator's only active right should be the capacity to refuse a given text:

A somewhat brutal summary of the code of ethics of the Association des traducteurs littéraires de France (ATLF), 1988:

Let it suffice to say that translators:
1. Must have adequate linguistic competence.
2. Must have knowledge of the pertinent subject matter.
3. Must refuse to translate from a TT unless with the consent of the author.
4. May only alter a text with the author's consent.
5. Have the right to accept or refuse a translation.
6. May demand the documents necessary for the translation.
7. Must respect professional secrets.
8. Must translate personally and ensure that their name appears on TT.
9. In the case of co-translation, the names of all the translators must appear on TT.
10. Must demand the same conditions if co-translating.
11. Must refuse work detrimental to a fellow translator.
12. Must not accept work conditions inferior to those established by the profession.

The priority of the author over the translator is clear in points 3 and 4. The translator's only elective right is mentioned in point 5, the capacity to refuse to translate a text. Points 1, 2, 3, 4 and 7 might perhaps better be described as clients' rights. The blunt individualism of point 8 ("Le traducteur s'engage à traduire personnellement") is curiously and immediately modified by the co-translation rules of points 9 and 10. All in all, this negative ethics delimits certain areas where translators should not enter, but accords them little positive identity within their own territory.

What is the explicit guarantee of equivalence in this code of ethics? Since authority here only partly ensues from the author, there is relatively little inspiration or grace at stake. The code instead refers to some basic traits of professional competence, expressed as adequate knowledge of the ST and TT languages and the pertinent subject matter. Translators are authorised as translators not because of any individual identity, but because of the linguistic and communicative skills they have acquired. The level of meritocratic performance is brought down to that of a passive collective capacity. The prime guarantee of equivalence - the first point named in the above code, and the first that comes to mind in popular definitions of translation - becomes the translator's linguistic competence, a space supposedly untouched by active passages of transfer or dilemmas of allegiance. If translation is thus seen as little more than a linguistic phenomenon, it becomes impossible to see why there should be any ethical problems to solve.

Some translation theories turn this technocratic evasion into an incipient morality, finding in "natural" language a kind of universal common ground to be attained and retained. Such is the paradise targeted by Nida's "closest natural equivalent" (1959, 33) and, more belligerently, by Newmark's campaign against jargon, pretension and superficially asymmetric discrimination.

What is this neutral space full of "natural" language? To whom does it belong? Apparently to everyone, since universal equality is one of the principles of its very construction. But if the tongue includes a privileged set of mechanisms for creating specific and relatively untransferable belonging, how can it possibly map out a neutral ground able to abolish all distinctions between "we" and "they"? How can knowledge of "natural" tongues suddenly make the translator equally "natural" and thus neutral?

There have been numerous outcries against the naturalness of language. One of the loudest came from Roland Barthes, who bluntly declared the apparently democratic tongue to be not only undemocratic but quite simply fascist, since "fascism does not prohibit the saying of things; it obliges things to be said":

"In French (which provides the most extreme examples), I am obliged to present myself first as a subject before I can manifest an action, which will then forever be no more than my appendage: what I do is merely the consequence of what I am. In the same way, I am always obliged to choose between the masculine and feminine genders; the neutral and complex genders are prohibited. And again, I have to indicate my relation to the second person by using either tu or vous ; I am allowed no recourse to social or affective suspense or mystery." (1977, 14)

The complaint, which concerns no more than the primary linguistic articulations of the I-here-now, is most readily understandable in terms of Barthes' homosexuality; the assumed naturalness and equality of the tongue turns real desire for equality into something quite unnatural and unequal. Missionary-translators may similarly suffer from a lack of neutrality, since in many languages God must be either male or female, and natural common ground is hard to find. And yet Barthes managed to express amorous sentiments in fragments of beautiful French; and Bible translators manage to translate.

An evasive reply to Barthes is that he should have either spoken another language - English perhaps offers a higher degree of superficial neutrality - , or set about reforming the French language itself, from a position of authority (the above complaint is from his inaugural lecture at the Collège de France). In the case of masculine dominance in English gender, the latter course has been undertaken reasonably successfully and change has proved relatively painless. In the case of Barthes' French, change was in fact sought through the literary and critical reworking of the tongue, creating a multiplicity of contradictory rules in such a way that complexity itself ultimately left space for a certain undefined liberty, although strategies of selective blindness would also have been possible (Anne Garreta's 1986 novel Sphinx responds to Barthes by avoiding gender-marked adjectives from beginning to end). There are many ways of attaining neutral linguistic ground, but this neutrality cannot be attributed to nature.

The tongue is neither naturally neutral nor a fascist conspiracy. The neutral ground sought by democratic ethics can be expressed in language, but its expression requires work, transformation and creation, be it by the translator or by the discontent author. Since there is no neutrality prior to this linguistic work, the ethics of professional practice are to be sought in the active role of the professional. Neutrality is not natural; it has to be created.
 

To translate is to attempt improvement

The rejection of natural neutrality makes it possible to address several thorny questions commonly avoided by the ethics of anonymity. The most important of these problems is the translator's right or duty to improve originals (the question is significantly absent from the code of ethics summarised above, which bluntly adjudicates this right to the author). Since translators cannot help but take position - since even neutral positions have to be created - , their ethics should break with passive non-identity, forcing them actively to evaluate the texts they work on, making them take on a major degree of responsibility for the texts they produce.

The question of improving a text concerns various domains. With respect to "facts", few would disagree with Newmark's sound if pedestrian advice:

"When extralinguistic reality is wrong in the source text, the translator must say so. Misstatements must be either corrected or glossed. This responsibility is more important than monitoring the quality of the writing in the source-language text." (1981, 128-129)

The real problem begins on the level of what Newmark describes as "monitoring the quality of the writing", which should be extended to include the monitoring of pertinence, relations between implicit and explicit material, and strategies of addition and deletion. Were Peter the Great and Perrot d'Ablancourt actually improving their texts? Was Major Spears effectively improving de Gaulle? Such questions are less easily answered than appeals to "extralinguistic reality". But there can be little doubt that Peter the Great, d'Ablancourt and Spears all thought they were carrying out improvements, in accordance with either personal criteria or group interests.

After all, if the transferred texts had been perfect and required no change, why should they have been translated? Is not translation itself a major change? The questions demands some thought.

What might be wrong with a merely transferred text? Obviously, if a transferred text's language or codes unacceptably restrict its performative capacity in a new cultural context, the first improvement must be the elaboration or transformation of its language or codes, the opening of a new potential "we" for new potential receivers. By definition, from the perspective of the translator receiving the text, the first improvement must be fact of translation itself.

Translational improvement thus initially means enabling a text to reach certain receivers who would otherwise find that text unavailable or incomprehensible. In certain cases this requires that improvement pass through the reproduction of defects, so that the original text or author may be recognised as defective and thereafter be avoided or corrected in future texts. To translate is not always to correct; but it is always to attempt improvement, sometimes according to a long-term vision.

On this level, improvement is obviously a very relative notion; it is always in terms of the specific purposes of the person or group interested in creating a new "we", in extending reception in a certain direction (and not in others).

The pertinent question is then not whether the translator should improve a transferred text, but according to whose criteria improvements should be made. For whom is the text to be improved? And with authority from which side of the summit?
 

Translators' first loyalty should not fall one side or the other

Since the question of improvement can be formulated as a classical problem of multiply divided loyalties, it may be conceptualised in terms of the figures to which the translator might turn when in search of authorisation. There are at least five such figures:

- The author or sender, whose consent might be necessary for alterations (as stated in the ATLF code of ethics).

- The receiver, who might have a right to know about the defects of the original (the ATLF code is prefaced by something called the "right to exact information of all kinds", described as "one of the fundamental rights of Man [l'Homme]" ).

- The client or initiator ("Besteller", "donneur d'ouvrage"), who might demand that the translation be written in accordance with a specific aim or purpose (as broadly defended in the wake of Holz-Mänttäri 1984).

- The ST culture, since even the most disastrous poverty or iniquitous domination might have the right to treatment in terms of a regime of cultural equality (a 1982 UNESCO text advises member countries that "international cultural cooperation depends on respect for cultural identity, for the dignity and value of each culture, for independence, for national sovereignty and for non-intervention").

- The TT culture, since translation has the potential to alter the defensive capabilities of the receiving community and the long-term expressive capacities of the receiving language (certain Canadian translators engage in "elegant rewriting" in order to protect and enrich Québécois, cf. Delisle 1984, 118).

Of these five possible sources of authority for improvement, which is to have priority? And if there is no absolute priority, how, in any given situation, might translators know in which of these directions they should look first?

Several short-sighted solutions should be dispensed with before a general answer can be given.

First, although the ideal translation might be thought to be one in which all these parties would find visible improvement, the ideal of equivalence suggests exactly the opposite, namely that what these parties are primarily interested in is invisible improvement of the kind that can be mistaken for zero-degree value change. From the translator's point of view, to translate is to improve. But from the perspective of authors, clients, receivers and cultures, what we are calling translation is the production of equivalence. No one really wants to know about the translator's value added; few are automatically prepared to let unknown and potentially untrustworthy individuals decide what is or is not an improvement, especially when those individuals start to reveal their ignorance by asking clients too many questions. The more the translator manifests his or her individuality, the less chance equivalence has of finding believers. The improvements most likely to suite all parties are thus those made by silent hands, leaving fragile translational fictions untouched.

Second, it might be assumed that the fairest kind of improvement is that which is most explicit. The best translation would then be one in which nothing is hidden from the receiver, all problems are elaborated, all original defects are noted and expansion is worked up to the outer threshold of relative equivalence. But in whose favour would such an ethics of explicitness function? Certainly not the client, who will only have to pay more. And not especially the author, who will usually doubt and resent distortions of the original form. Receivers are surely those who stand to gain the most from maximum expansion. But how many receivers have the time to go through vast tomes of elucidation before finding the part of a text they consider of most value? How many would rather have a shorter text, elaborated according to purpose-specific criteria of pertinence? How many would exchange their apparently universal "right to information" for a socially determined right to "freedom from redundant or impertinent information"? Since excessive textual quantity is one of the best ways of concealing important information - books are best hidden in libraries - , an ethics of explicitness cannot be expected to provide any general panacea for problems of immoral manipulation.

Third, an ethics of commercial service, most effectively based on the translator's responsibility to the client's purpose, would simply place mercenary behaviour beyond the reach of ethical critique, suggesting that the most improved translation is the one which gets paid the most and that none of the other possible sources of authority count for anything at all. In the absence of any translational guiding hand (such as Luther's insistence on faith), an ethics of commercial service would appear to be more like a non-ethics, although the rather more interesting but associated criterion of "professional detachment" will be commented upon below.

Fourth, an ethics based on symmetrical respect for hypothetical cultural equality would seem inadequate to the fundamentally asymmetric principles of translation itself: authors and translators are by definition not equal; text flows between senders and receivers are rarely balanced or reciprocal; the right to information is not automatically a universal blessing. An ethics of improvement must recognise that translation is a profoundly asymmetrical phenomenon.

Fifth, an ethics based on asymmetrical cultural specificity would in fact fare no better than its symmetrical counterpart. If one says that each culture has and should have its own way of translating - and thus its own way of improving texts according to its own criteria - , there is no way of recognising and assessing translation as an actively communicational phenomenon necessarily relating at least two cultures. An ethics of cultural specificity would be like suggesting that improvement in a marriage can only result from having husband and wife pursue their separate criteria, which in fact amounts to saying that the best marriage is divorce. Is the best intercultural relation then culture-specific? Is the best translation really non-translation in disguise?

There can be no doubt that certain phenomena are culture-specific. But translation, precisely because it is an intercultural phenomenon, should not be one of them. To pretend otherwise would be to push relativism to the paranoic extremes translation should ideally overcome; it would be to draw up rules based on unchanging cultural distances and then apply them to a phenomenon which has as its function precisely the transformation of those distances; it would be to forget that the kings, princes and priests taking their sovereignty from cultural specificity can never totally trust translators anyway. Because, despite whatever cultural specificity they might be attributed, translators have always been intercultural.

Questions of divided loyalty cannot be decided in terms of looking in one direction rather than another. Criteria of equivalence, explicitness, purpose-adequacy, hypothetical cultural equality and cultural specificity fail to provide any convincing orientation as to the general nature of translational improvement. This is because they are not in themselves translational criteria; they are not derived from any careful contemplation of what translation is and does. In order properly to decide how and when to improve through translation, one must first position oneself in an appropriately intercultural space, and only then consider the fortunes of individual senders, receivers, clients or cultures.
 

Professional detachment is attachment to a profession

A properly translational ethics must precede questions of individually divided loyalty. It must be developed beforehand, in the space of the collective professionalisation which produced the ideal of the individual translator in the first place. Translators' prime loyalty must be to their profession as an intercultural space, an intersubjective place in which criteria of translational quality can and should be determined.

On what basis should these decisions be made? It is important to realise that certain factors lie beyond professional control.

First, there is rarely any question of improving the source-text, which by definition lies beyond the space of translation. The ST is generally to be regarded as a fait accompli, outside the control of the translator and only entering translation through irreversible transfer. No one can change de Gaulle's speech of 1940. What can be improved is the transferred text, the original as it arrives in a new context: a translator could and did attempt to improve de Gaulle's speech as it existed in Britain in 1966.

Second, the question of improvement does not directly concern the content of the translated text. When Geoffrey Kingscott (1990) discusses the moral responsibility of the translator in terms of issues like "the manufacture or sale of armaments, the use of animals in laboratory experiments, and pornography", the questions involved might concern the translator's opinions and beliefs as an individual, but do not concern the translator's profession as such. The individual translator can refuse to work in these areas, but does so as an individual, like any other worker reluctant to be involved. There are no strictly professional grounds for saying that such texts should not be translated.

Source-texts and non-translational ideologies must thus lie beyond the space in which a professional ethics can be developed.

But this does not mean that the profession should then refuse any ideologically based alterations whatsoever. Content is one thing; axiological presentation is quite another. In this respect, Kingscott is more provocative than exact when he argues that:

"...the translator or interpreter, when he or she is translating and interpreting, is in the same position as an advocate. An advocate, during the course of his career, may occasionally appear on behalf of an unfortunate victim, but it is more than likely that his client will be a double-dyed villain who would make him shudder with disgust if he had not learnt to take an attitude of professional detachment.

"Our clients rely on us to put their case, in the foreign language, as they would like to see it put, not as we would like to see it put." (1990, 48)

The analogy might be more instructive than it appears. If a client knows how a case should be put, why should he or she need an advocate? Obviously, so that a spontaneous presentation of the case can be improved; so that certain embarrassing details can be left out or hidden, other advantageous elements extended or added, a more formal or logical order instituted. The advocate, like the translator, is employed to improve a given text. But should the translator therefore mimic the advocate's professional detachment with respect to the client's purpose as such?

When a barrister argues a client's case, it is in a symmetrical situation where a further professional will argue the opposing case, producing a partitioned dialogue leading to a formal conclusion or judgement. That is, professional advocates are employed to facilitate exchange within a highly formalised regime. If they are detached with respect to the moral value of their client's actions and opinions, it is because they are firmly attached to the ethical values of the discursive regime within which they work: they will not abuse the judge; they will respect court rules; they will speak the formalised language of the profession; they will tailor the client's case to suite the rules and conventions of the applicable legal code.

Now, if the translator is really a kind of advocate, there is no reason why the analogy should be limited to professional detachment. One should also ask what professional criteria - what regime of formalised exchange - might justify this detachment as ethical commitment. If there is detachment, it is only because there is attachment to something else. Barristers are attached to the rules and procedures of the court, which has as its purpose the dispensing of justice. Translators should presumably be attached to the rules and procedures of their profession, justifying their actions and decisions in terms of translation's own ultimate aim. But where advocates have the symmetry of accusation and defence, translators have only the asymmetry of imposed directionality; where advocates leave the final decision to a judge, translators themselves are surely the only people fully qualified to assess the shortcomings of their intercultural work. For these reasons, the fact that the advocate's aim is to serve a client does not necessarily mean that the same purpose is valid for translators.

The apparent conflict between translational improvement and professional detachment thus in fact concerns the translator's attachment to a profession whose ultimate aims have yet to be formulated.
 

Translation has purposes of its own

Some descriptions of the purpose or aim of translation appear to focus on what translators actually do or produce, without reference to the authority of external actors. In view of the negative arguments presented above, these descriptions would seem to be the only ones worthy of our continued interest. Unfortunately, even the most basic of such attempts tend to relapse into the ethics of service.

Ladmiral defines the purpose [finalité] of translation (1979, 15; 1981, 25):

"The purpose of a translation is to enable us to go without reading the original text."

This would seem so basic as to be axiomatic, and yet it gives rise to several problematic questions: Who is this "us"? (contradictory equivalence suggests that translations are not just for their receivers); Are transliterated terms not properly part of a translation? (I have argued that they are); Why should a translation not lead to a reading of the transferred text? (such has long been the pedagogical use of literalist cribs); Can the particularity of "a translation" be confused with the general function of translation as a social phenomenon? (I have argued that, in ethical questions, the general case is very different from particular translators and individual translations). And then the most obvious question: If the purpose of a translation is to save "us" from the original, what is the purpose of "going without reading the original"? Are foreign languages and codes no longer to be learned? And so on. One senses that some further, more noble aim is required.

Laugier (1973) defines the social purpose [finalité sociale] of translation:

In a fine polemical paper, Laugier distinguishes radically between the social purpose of translating technical texts - which should enable technology to operate - and that of translating literary texts - which, according to formalist aesthetics, should create literariness in the space of reception - , such that the technical purpose would be well served by translating signified for signified, whereas the literary purpose would then require the almost mechanistic literalism of signifier for signifier.

The obvious ethical problem with this general approach is that it limits the social purpose of translation to that of serving the pre-established (and always available?) purpose of ST texts, thus once again displacing the question of purpose away from translation itself. If the purpose of translation is to reproduce ST technical or literary functions, what then is the purpose of technical and literary functions? And so on, again. Like Ladmiral, Laugier falls back into an ethics of service, with all its potentially mercenary consequences.

We are thus returned to the fundamental questions: Can translation be seen as having a purpose of its own? How might such a purpose be formulated? For whom should it be formulated?

I believe that a basic answer to these questions has already been intimated in the above discussion of professionalism. Let me now bring together the main steps in the argument.

If one can accept that, from the perspective of translators, to translate is to attempt improvement, if one can accept that translators' first loyalty should be to their profession, and if it can also be accepted that translators are by definition intercultural subjects with intercultural professional relations, it becomes possible and interesting to propose that, from the perspective of translators, the ultimate aim of translation is to improve the intercultural relations with which they are concerned.

Such would be the formal principle of ideal translational improvement most adequate to translators as an intercultural collectivity.

As much as this principle might appear to be saying nothing, as much as it is intentionally devoid of historical content, I suspect it is saying something simply to the extent that it has very rarely been formulated in such terms. But I freely concede that one should work towards giving it a more positive historical content.

Goethe was close to this principle, and no doubt more elegant, when he described the aim of translation as being to "increase tolerance between nations" (in Lefevere 1977, 34). But elegance is a relative virtue, and tolerance presupposes a degree of sovereignty and symmetry that nations are fast losing. A modern translational regime cannot assume that one act of tolerance will be rewarded with another, that there will be a judge to listen impartially to accusation and defence, that all parties will remain equally purposeful and command respect as such. Positive virtues like tolerance, understanding and enlightenment should not be dissociated from translation, but a world of asymmetry and limited sovereignty requires a slightly more cunning mode of thought.

Instead of arrogantly assuming that the immediate enemies of translation are intolerance, misunderstanding and ignorance, we must look closely at the historical ways in which translation opposes and interacts with belonging, which in itself surely has positive values. Far from proclaiming the goodness of all translation, an ethical notion of improvement should seek to identify the form of historically progressive dialectics between translation and belonging. That is, it should look for elements of globally beneficial translational regimes; it should be addressing wholly contemporary issues like the intercultural status of English, the stimulation of European plurilingualism, and the representation of revived Romantic nationalisms. Obviously, these questions involve far more factors than can be grasped through translation theory alone.

At this point I can only put forward several examples that might be of interest for future discussion, especially now, in the early 1990s, at a moment of major regime change and thus extreme uncertainty.

An exception to prove the rule:

The excellent film Patton at one point portrays an American-Soviet celebration immediately after the fall of Nazi Berlin. There is no communication between the American and Soviet generals until, when the Soviet proposes a toast, the American general orders his interpreter: 'Tell him he's a son of a bitch.' The interpreter understandably doubts that this is likely to improve intercultural relations: 'I can't tell him that, sir!' The general insists, the interpreter interprets, and the Soviet's relayed reply comes back as: 'You're a son of a bitch too.'" And then there is a toast.

The anecdote calls for two observations. First, the interpreter is undoubtedly correct to take an active role and offer advice to the Prince of half-Berlin. Second, his decision ultimately to convey the text is justified only by the extreme symmetry of the situation at all levels, both in the immediacy of the transfer situation and the authority of the communication participants.

This global symmetry was to become an international political and military regime ensuring some forty years of relative peace. But beyond the cold-war regime, such texts should probably not be translated.

An unexpectedly asymmetric regime:

On 9 January 1991, the American Secretary of State James Baker met the Iraqi Minister of Foreign Affairs Tariq Aziz in Geneva. A letter from President George Bush, addressed to Saddam Hussein, was transferred from Baker to Aziz, but the Iraqi refused to convey the letter to Saddam. One might say that the fluently bilingual Aziz refused to transfer and translate for his president. The reason given for this non-translation was that "the language of the letter is inappropriate to communication between heads of state". In other words, in an asymmetric situation with respect to both the transfer situation and the relative power of the communication participants, the letter was not going to do anything to improve intercultural relations.

A suspicious mind might suppose that George Bush's letter was written in the spirit of cold-war confrontations, in terms of the relative symmetry of the Cuban missile crisis, if not entirely in the language of General Patton. If so, the intermediary's refusal, ostensibly based as it was on cultural values like respect and pride, is indicative of a very different international regime. But few failed to appreciate it at the time.

Should translators really take their decisions on the level of world-saving questions like intercultural regimes? In practice, it would be a lot to ask and could risk some quite absurd results. But I do not believe translators should passively accept the role of mere technicians, working on means and never considering anything but the most immediate or commercial ends, burying themselves in a practice falsely cut off from history and theory. Translators should also be intellectuals; they should have ideas about who they are and what they hope to achieve as a collectivity. If Sartre (1965) could describe the main contradiction of the intellectual of his day as "universalism of profession" opposed to "particularism of class", the problems of the translating intellect might well be described in terms of "interculturality of profession" opposed to something like an immediately commercial "particularism of task".

In ethical questions, priority must be given to the interculturality of the translator's profession.

Should one refuse the "superiority of the white man"?

A Spanish encyclopedia for children contains the following sentence: "El hombre blanco ha marcado el avance del progreso humano durante más de dos mil años." This was translated as: "The white race has led human progress for more than two thousand years." In this particular case, where the text is "authorless" and the intended receiver is aged between ten and fourteen, even the most professionally detached translator would have doubts about serving a client's interests by translating in this way.

Thankfully, the sentence did not reach its intended public. The translation was sold to an American publisher working from Taiwan, where it was radically edited, with appropriate complaints being sent to the Spanish publisher, who then conveyed them to the man who had written the sentence. The editor in Taiwan pointed out that the Tang dynasty in China was far more advanced than the European culture of the time, and that the sentence was thus factually incorrect (cf. Newmark). The Spanish "author" then replied that he didn't care about China and, anyway, there was "nothing to compare with Romanic art and Gothic cathedrals". I refused to translate this latter argument, not because I think Tang art is better than Gothic cathedrals - who am I to say? - , but because the argument itself had deviated from the original sentence (which concerned "leading progress" and thus cultural contact) and was moreover highly unlikely to improve intercultural relations.

In some cases, dark-age silence or separate development is to be preferred to a regime of mud-slinging.

Should one refuse "the Catalan nation"?

One of the recurrent problems of translating from Spanish or Catalan into English is the political organisation of Spain into seventeen "autonomous regions", some of which call themselves "nations". Particularly in Catalan, the terms for "nation" and "national" tend to refer to Catalonia and not to Spain, which is then usually referred to as "l'Estat" (the State). This means that "el nivel estatal" (State-level) concerns the whole of Spain, and "el nivel nacional" (national-level) sometimes concerns just Catalonia, upsetting the English-language concept of a nation divided into "states" (as in United States or the states of Australia, Nigeria or India). Part of the problem is that the Romance-language term "nation" refers to cultural tradition more than to the nation-state as such, as it tends to do in English (much the same terminological problem is raised by French-Canadian nationalism). But the difficulties here are not just linguistic. Since the internal political regime of "autonomous regions" is specifically post-Franco, use of the term "nación" to refer to a region is by no means neutral in Spanish. It implicitly demands further independence. The translator is thus asked to convey or refuse this implicit demand.

In such situations, the translator must decide whether or not to call Catalonia a nation. It is not an easy decision.

Since I believe that, in the context of European integration, improved relations require stimulation of minor cultures, I also believe that translation should be used to convey awareness of such cultures. Ideologically, I am in favour of bending the dominant English concept of "nation" so as to be able to talk about "the Catalan nation"; I have no qualms at all about unambiguously referring to Catalonia as a "stateless nation" and have done so in one way or another whenever indicated by the Spanish or Catalan text. However, when considering this problem, I searched my translations for the term "nation" and was surprised to find that, as a technician translator concerned with means and not ends, I have repeatedly referred to the Spanish State as a "nation", even when not indicated by the Spanish or Catalan text. In a text on EEC agricultural policies, a phrase like "No hay estructura administrativa al nivel estatal o autonómico para el desarrollo..." is efficiently rendered as "There is no specific administrative framework on the national or regional level for the development and application of the EEC regulation." ("Farm Structures and Pluriactivity in Spain", 1989). That is, I find that I have been prepared to convey regionalist ideology when it appears, but not to impose it when it does not appear or is not at issue. An ethics of translation should be able to address moral dilemmas when they arise, but it should not exhume them unnecessarily.
 

Regimes and the training of translators

The translator's profession can be approached from the outside, through the critique and distancing of alternative sources of authority. But it can also be considered from within, through reflection on the way principles, norms, rules and procedures (the components of regimes) are or should be transmitted from translator to translator, master to apprentice, teacher to student.

Do teachers of translation teach their students principles, norms, rules and procedures? Undoubtedly, yes. But are these regime elements properly translational? The vast majority are inevitably linguistic or have to do with area studies, bearing on adequate comprehension of content and on target-language production rather than on translation as such. Guidelines are widely formulated and taught as lexical, syntactic, discursive, pragmatic, probabilistic, psychological or sociological laws. Prestige is easily won by anything resembling standardised research into the nature of these apparent laws, as if they had inner secrets to reveal to passive microscopes. As argued above, the guise of the natural thus supports a falsely benign neutrality requiring no strong ethical elaboration. Such approaches effectively let sleeping students lie. They have little to do with translation.

How much of what is taught really concerns translation as a positive activity? Beyond apparent rigour and a lot of exercise, honest reflection on actual teaching practices might come up with a few homely imperatives like "explain abbreviations", "avoid footnotes", "do not translate names of real people" or "be more literal when translating citations". But none of these really rise above the level of personal preferences or endemic conventions; none of them can claim absolute applicability in all situations; none of them strictly concern ways of infallibly improving through translation. And then, who has the authority to impose such preferences anyway? Whose theorisation can be accorded such absolute validity? Can the teaching of translation really convey strict guidelines at all?

I believe that the effective teaching of translation has much more to do with using theory to open up a series of possibilities, a series of alternative ways of translating, and then inductively questioning those alternatives in terms of the specific and general aims of translation. In most cases, this process requires little more than occasional cries like "Think of your reader!" or "What will your client say?", initially overshadowed by the basic imperative "Decide!", since students tend to take some time to accept that translators have the right and duty to make decisions. But any reasonable course should include enough morally dubious or badly written texts for serious questions to be raised about expansion and abbreviation, addition and deletion. The ensuing discussions can usefully be directed in terms of questions like "Why translate this text?", "For whom might this text be translated?" and ultimately, "Why translate?", in the noble hope of transforming the copiers of rules into a profession of decision-makers, forcing theorisation to show itself as a part of translational practice.

For each future generation or world situation, the answering of such questions, the practical process of collective theorisation, might ideally constitute the consensual basis of a peculiarly translational regime.

All teaching influences students, for or against the opinions of the teacher. Although each generation of translators is undoubtedly free to work as it sees fit, the answers it comes up with will initially be to the questions posed by the previous generation. It is no crime to exert influence, to direct attention one way or another. But it does seem less than ethical to dress up current translational preferences in the costume of immutable rules, and thus attempt to deny future generations the right to decide about their own professional ethics.

Only when translation rules are recognised as ethical decisions, when notions of split loyalty and potential treason are accepted as something more than idle metaphor, and when the technologies of means are accompanied by hard thought about why intercultural relations are important and how they can be improved, only then might we develop properly translational regimes.
 
 
 

Next chapter