The central problem of translating has always been whether to translate literally or freely. The argument has been going on since at least the first century BC. Up to the beginning of the nineteenth century, many writers favoured some kind of `free' translation: the spirit, not the letter; the sense not the words; the message rather than the form; the matter not the manner. This was the often revolutionary slogan of writers who wanted the truth to be read and understood - Tvndale and Dolet were burned at the stake, Wycliffs works were banned. Then at the turn of the nineteenth century, when the study of cultural anthropology suggested that the linguistic barriers were insuperable and that language was entirely the product of culture, the view that translation was impossible gained some currency, and with it that, if attempted at all, it must be as literal as possible. This view culminated in the statements of the extreme `literalists' Walter Benjamin and Vladimir Nabokov.
The argument was theoretical: the purpose of the translation, the nature of the readership, the type of text, was not discussed. Too often, writer, translator and reader were implicitly identified with each other. Now the context has changed, but the basic problem remains.
I put it in the form of a flattened V diagram:
SL emphasis TL emphasis
Word – for – word Translation Adaptation
Literal Translation Free Translation
Faithful Translation Idiomatic Translation
Semantic Translation Communicative Translation
This is often demonstrated as interlinear translation, with the TL immediatelv below the SL words. The SL word-order is preserved and the words translated Singly by their most common meanings, out of context. Cultural words are translatedliterally. The main use of word-for-word translation is either to understand the mechanic of the source language or to construe a difficult text as a pre tanslation process.
The SL grammatical constructions are converted to their nearest TL equivalents but the Iexical words are again translated singly, out of context. As a pre-translation processthis indicates the problems to be solved
faithful translation attempts to reproduce the precise contextual meaning of the original within the constraints of the TL grammatical structures. It `transfers' cultural words and preserves the degree of grammatical and lexical `abnormality' (deviation from SL norms) in the translation. It attempts to be completely faithful to the intention and the text-realisation of the SL writer.
Semantic translation differs from `faithful translation' only in as far as it must take more account of the aesthetic value (that is, the beautiful and natural sound,) of the SL text compromising on `meaning' where appropriate so that no assonance, word play or repetition jars in the finished version. Further, it may translate less important cultural words by culturally neutral third or functional terms but not by cultural equivalents - une nonne repassant un corporal may become-'a nun ironing coprocal cloth' - and it may make other small concessions to the readership. The dr,nnmion between `faithful' and `semantic' translation is that the first is uncompmauising and dogmatic, while the second is more flexible, admits the creative exceptation 100% fidelity and allows for the translator's intuitive empathy with the original.
I his is the `freest' form of translation. It is used mainly for plays (comedies) and poetry the themes, characters, plots are usually preserved, the SL culture conwvricH to the'CL culture and the text rewritten. The deplorable practice of having a pI;m or poem literally translated and then rewritten by an established dramatist or pun has produced many poor adaptations, but other adaptations have `rescued' period plays.
free translation reproduces the matter without the manner, or the content without the form of the original. Usually it is a paraphrase much longer than the original,also-called `intralingual translation', often prolix and pretentious, and not translation at all
Idiomatic translation reproduces the `message' of the original but tends to distort nuances of meaning by preferring colloquialisms and idioms where these do not exist in the original. (Authorities as diverse as Seleskovitch and Stuart Gilbert tend to this form of lively, `natural' translation.)
Communicative translation attempts to render the exact contextual meaning of the original in such a way that both content and language are readily acceptable and comprehensible to the readership.
COMMENTS IN THESE METHODS
Commenting on these methods, I should first say that only semantic and communicative translation fulfil the two main aims of translation, which are first, accuracy,, and second, economy. (A semantic translation is more likely to be economical than a communicative translation, unless, for the latter, the text is poorly written.) In general, a semantic translation is written at the author's linguistic level, a communicative at the readership's. Semantic translation is used for `expressive' texts, communicative for `informative' and `vocative' texts.
Semantic and communicative translation treat the following items similarly: stock and dead metaphors, normal collocations, technical terms, slang, colloquialisms, standard notices, phaticisms, ordinary language. The expressive components of `expressive' texts (unusual syntactic structures, collocations, metaphors, words peculiarly used, neologisms) are rendered closely, if not literally, but where they appear in informative and vocative texts, they are normalised or toned down (except in striking advertisements). Cultural components tend to be transferred intact in expressive texts; transferred and explained with culturally neutral terms in informative texts; replaced by cultural equivalents in vocative texts. Badly and/or inaccurately written passages must remain so in translation if they are `expressive', although the translator should comment on any mistakes of factual or moral truth, if appropriate. Badly and/or inaccurately written passages should be `corrected' in communicative translation. I refer to `expressive' as `sacred' texts; `informative' and `vocative', following Jean Delisle, as `anonymous', since the status of their authors is not important. (There are grey or fuzzy areas in this distinction, as in every aspect of translation.)
So much for the detail, but semantic and communicative translation must also be seen as wholes. Semantic translation is personal and individual, follows the thought processes of the author, tends to over-translate, pursues nuances of meaning, yet aims at concision in order to reproduce pragmatic impact. Communicative translation is social, concentrates on the message and the main force of the text, tends to under-translate, to be simple, clear and brief, and is always written in a natural and resourceful style. A semantic translation is normally inferior to its original, as there is both cognitive and pragmatic loss (Baudelaire's translation of Poe is said to be an exception); a communicative translation is often better than its original. At a pinch, a semantic translation has to interpret, a communicative translation to explain.
Theoretically, communicative translation allows the translator no more freedom than semantic translation. In fact, it does, since the translator is serving a putative large and not well defined readership, whilst in semantic translation, he is following a single well defined authority, i.e. the author of the SL text.
It has sometimes been said that the overriding purpose of any translation should be to achieve `equivalent effect', i.e. to produce the same effect (or one as close as possible) on the readership of the translation as was obtained on the readership of the original. (This is also called the `equivalent response' principle. Nida calls it `dynamic equivalence'.) As I see it, `equivalent effect' is the desirable result, rather than the aim of any translation, bearing in mind that it is an unlikely result in two cases: (a) if the purpose of the SL text is to affect and the TL translation is to inform (or vice versa); (b) if there is a pronounced cultural gap between the SL and the TL text.
However, in the communicative translation of vocative texts, equivalent effect is not only desirable, it is essential; it is the criterion by which the effectiveness, and therefore the value, of the translation of notices, instructions, publicity,
propaganda, persuasive or eristic writing, and perhaps popular fiction, is to be assessed. The reader's response - to keep off the grass, to buy the soap, to join the Party, to assemble the device - could even be quantified as a percentage rate of the success of the translation.
In informative texts, equivalent effect is desirable only in respect of their (in theory) insignificant emotional impact; it is not possible if SL and TL culture are remote from each other, since normally the cultural items have to be explained by culturally neutral or generic terms, the topic content simplified, SL difficulties clarified. Hopefully, the TL reader reads the text with the same degree of interest as the SL reader, although the impact is different. However, the vocative (persuasive) thread in most informative texts has to be rendered with an eye to the readership, i.e., with an equivalent effect purpose.
In semantic translation, the first problem is that for serious imaginative literature, there are individual readers rather than a readership. Secondly, whilst the reader is not entirely neglected, the translator is essentially trying to render the effect the SL text has on himself (to feel with, to empathise with the author), not on any putative readership. Certainly, the more `universal' the text (consider `To be or not to be'), the more a broad equivalent effect is possible, since the ideals of the original go beyond any cultural frontiers. The metalingual sound-effects which the translator is trying to reproduce are in fact unlikely to affect the TL reader, with his different sound-system, similarly, but there may be compensation. In any event. the reaction is individual rather than cultural or universal.
However, the more cultural (the more local, the more remote in time and space) a text, the less is equivalent effect even conceivable unless the reader imaginative, sensitive and steeped in the SL culture. There is no need to discuss again the propriety of `converting' Keats' `Season of mists and mellow fruitfulness or Shakespeare's `Shall I compare thee to a summer's day?' into languages 01 countries where the autumns and summers are unpleasant. Cultural concessiom (e.g., a shift to a generic term) are possible only where the cultural word is marginal, not important for local colour, and has no relevant connotative ol symbolic meaning. Thus, in a Bazin text, it is inadequate to translate: Il est le plu. pelican des peres as `He is the most devoted of fathers' or `He is a symbol of paterna love, a pelican.' A compromise version, retaining the cultural element (pelican) might be `He is as devoted as a pelican to his young.' Authoritative statements being addressed to a readership rather than individual readers, if written in `public language should produce equivalent effect: Pericles, Jefferson, Lincoln, Churchill De Gaulle - the names suggest a universal appeal that asks for a loud and moderr echo in translation.
Communicative translation, being set at the reader's level of language an( knowledge, is more likely to create equivalent effect than is semantic translation a the writer's level; but a text written some hundred years ago gives the reader of th translation an advantage over the SL reader; the inevitably simplified, under translated translation in modern language may well have a greater impact than th original. Hence unser (our) Shakespeare, as educated Germans used to know hi work earlier in the century.
Equivalent effect is an important intuitive principle which could be testei but, as is often the case, the research would not be worth the effort; however, it i usefully applied in reasonable discussion, particularly within the `skill' (as oppose to the `truth', the `art' and the `taste') area of language. In translating `I haven't th foggiest idea', (aucune idee); would: Keine blasse Ahnung or Nicht die geringsv Ahnung or Ich habe keinen blassen Sehimmer davon have the closest equivaler effect? (A translation is pre-eminently a matter for discussion rather than fiat. To often it is still being imposed as a teacher's `fair copy' or model. In fact; the simple: sentence -'The gorgeous girl walked gingerly through the closet' - would, in or i spite of any context, be translated variously by a dozen experts in a dozen differer languages.)
I have dealt at length with the `equivalent effect' principle because it is a important translation concept which has a degree of application to anv type of tex but not the same degree of importance.
METHODS AND TEXT-CATEGORIES
considering the application of the two translation methods (semantic and com~ w mm;u ive) to the three text-categories, I suggest that commonly vocative and wlmvuuive texts are translated too literally, and expressive texts.not literally wuyh. "franslationese is the bane of tourist material and many public notices nal~rirculation est interdite de 22 h d 6 h; jeglicher Verkehr ist verboten von 22 bis 6 I 'lo , ';ill sexual intercourse is forbidden between 10 p.m. and 6 a.m.'). In the UK of foreign language (FL) publicity and notices is now high but there ii. iuon enough of them. In `informative' texts, translationese, bad writing and lack, wlidence in the appropriate linguistic register often go hand in hand; the wnnlvncv with familiar-looking but unfamiliar collocations (station hydrominerale; Imdrmnineral station' - read `spa') is simply to reproduce them. On the other, The inaccuracv of translated literature has much longer roots: the attempt to translation as an exercise in style, to get the `flavour' or the `spirit' of the within; the refusal to translate by any TL word that looks the least bit like the SL text, even by the SL word's core meaning (I am talking mainly of adjectives), so iliai ilic translation becomes a sequence of synonyms (grammatical shifts, and one word to two or three-word translations are usually avoided), which distorts its essence.
In expressive texts, the unit of translation is likely to be small, since words rather than sentences contain the finest nuances of meaning; further, there are lil.rlv to be fewer stock language units (colloquialisms, stock metaphors and collocations, etc.) than in other texts. However, any type and length of cliche must I,a i rnnslated by its TL counterpart, however badly it reflects on the writer. Note that I group informative and vocative texts together as suitable for communicative translation. However, further distinctions can be made.
Unless informative texts are badly/inaccurately written, they are translated more closely than vocative texts. In principle (only!), as they are concerned with rxira-linguistic facts, they consist of third person sentences, non-emotive style, paxt tenses. Narrative, a sequence of events, is likely to be neater and closer to i ranslate than description, which requires the mental perception of adjectives and Images.
The translation of vocative texts immediately involves translation in the problem of the second person, the social factor which varies in its grammatical and lexical reflection from one language to another. Further, vocative texts exemplify the two poles of communicative translation. On the one hand translation by standard terms and phrases is used mainly for notices: `transit lounge', Transithalle, sulle de transit. On the other hand, there is, in principle, the `recreative' translation that might be considered appropriate for publicity and propaganda, since the situation is more important than the language. In fact, provided there is no cultural gap, such skilfully written persuasive language is often seen to translate almost literallv.
Scanning the numerous multilingual advertising leaflets available today, I notice: (a) it is hardly possible to say which is the original; (b) how closely thev translate each other; (c) the more emotive their language, the more thev varv from each other; (d) the variants appear justified. Thus:
Young; fresh and fashionable. ,'Jun,;, frisch and modisch.,']eune, Irais et elegant.
Indeed, this is Vanessa. !n der Tat, so konnen Sic Vanessa beschreiben. Tels sont les qualiftcatifs de Vanessa.
This model links up with the latest trends in furniture design. Dieses Model schliesst be? den letsten Trends im Mobeldesign an. Ce modele est le dernier cri dons le domaine des meubles design. The programme exists out of different items. Das Programm besteht au.c verschiedenen Mobeln. Son programme se compose de differents meubles. . . . which vou can combine as you want . . . die Sie nach eigenem Bedurfnis zusammenstellen konnen . . . d assembler selon vos besoins ... (The three versions reflect the more colloquial style of the English (two phrasal verbs) and the more formal German, as well as English lexical influence (`design', `trend').)
Where communicative translation of advertisements works so admirablv, producing equivalent pragmatic effect, there seems no need to have recourse to `co-writing', where two writers are given a number of basic facts about one product and instructed to write the most persuasive possible advert in their respective languages.
I should mention that I have been describing methods of translation as products rather than processes, i.e., as they appear in the finished translation.
As for the process of translation, it is often dangerous to translate more than a sentence or two before reading the first two or three paragraphs, unless a quick glance through convinces you that the text is going to present few problems. In fact, the more difficult - linguistically, culturally, `referentially' (i.e., in subject matter) - the text is, the more preliminary work I advise you to do before vou start translating a sentence, simply on the ground that one misjudged hunch about a key-word in a text - say, humoral in le bilan humoral (a fluid balance check-up) or Laetitia in l'actrice, une nouvelle Laetitia (a Roman actress or an asteroid) - mav force you to try to put a wrong construction on a whole paragraph, wasting a lot of time before (if ever) you pull up and realise you are being foolish. This is another way of looking at the word versus sentence conflict that is always coming up. Translate by sentences wherever you can (and always as literally or as closely as you can) whenever you can see the wood for the trees or get the general sense, and then make sure you have accounted for (which is not the same as translated) each word in the SL text. There are plenty of words, like modal particles, jargon-words or grammatically-bound words,which for good reasons you may decide not to translate. But translate virtually by words first if they are `technical', whether they are linguistic (marigot), or cultural (sesterce), or referential (sessile) and appear relatively context free. Later, you have to contextualise them, and be prepared to back track if you have opted for the wrong technical meaning.
Research is now proceeding on how people translate, but there may be many factors (mood, deadline, need for a change of method) which will not betaken into account . Throughout the pre-translation process, you keep a clear image of what is actually happening, if only as a premiss that has to be continuously amended. This appliest poetry as to technical translation. Thus: Le soleil, sur le sable, o lutteuse , mLntnic En l'or de tes cheveux cha,uffe un bain langoureux (Mallarme, Tristesse d'ete) may suggest the sun bathing the golden hair of a sleeping girl lying on the sand my;gling (against what?) in languorous heat, and this image has to be kept constantly in parallel with the oblique and elliptical version of it rendered by the language.
As a postscript to this chapter, I add further definitions of translation methods.
Service translation, i.e. translation from one's language of habitual use into mother language. The term is not widely used, but as the practice is necessary in most countries, a term is required.
Plain prose translation. The prose translation of poems and poetic drama Initiated by E. V. Rieu for Penguin Books. Usually stanzas become paragraphs, prose punctuation is introduced, original metaphors and SL culture rctained, whilst no sound-effects are reproduced. The reader can appreciate i lie sense of the work without experiencing equivalent effect. Plain prose translations are often published in parallel with their originals, to which, after a `careful word-for-word comparison', they provide ready and full access.
information translation. This conveys all the information in a non-literary text, sometimes rearranged in a more logical form, sometimes partially summarised, and not in the form of a paraphrase.
Cognitive translation. This reproduces the information in a SL text converting the SL grammar to its normal TL transpositions, normally reducing any figurative to literal language. I do not know to what extent this is mainly a theoretical or a useful concept, but as a pre-translation procedure it is appropriate in a difficult, complicated stretch of text. A pragmatic component is added to produce a semantic or a communicative translation.
Academic translation. This type of translation, practised in some British universities, reduces an original SL text to an `elegant' idiomatic educated TL version which follows a (non-existent) literary register. It irons out the expressiveness of a writer with modish colloquialisms. The archetype of this tradition, which is still alive at Oxbridge (`the important thing is to get the flavour of the original'), was R. L. Graeme Ritchie, evidently a brilliant teacher and trans lator, who was outstandingly more accurat scraps of Ritchie's weaknesses: La Notreworked her way in'; La pluie brouilla les objets Cette vie se surpassera par le martyre, et le marty transcend itself through martyrdom and no coming'.
These last two concepts are mine, and onl will be useful as terms of reference in translation
Jumat, 06 Maret 2009