Varieties of close translation

It may be useful to distinguish literal from word-for-word and one-to-one trans- lation. Word-for-word translation transfers SL grammar and word order, as well as the primary meanings of all the SL words, into the translation, and it is normally effective only for brief simple neutral sentences: 'He works in the house - now', il travaille dans la maison maintenant. In one-to-one translation, a broader form of translation, each SL word has a corresponding TL word, but their primary (isolated) meanings may differ. Thus in passer un examen - 'take an exam', the two verb couplets can be said to correspond with each other, but, out of context, they are not semantic equivalents. Since one-to-one translation normally respects collo- cational meanings, which are the most powerful contextual influence on trans- lation, it is more common than word-for-word translation. Literal translation goes beyond one-to-one translation in including, say, le courage, derMut and 'courage' as literal equivalents; it is particularly applicable to languages that do not have definite and/or indefinite articles.

Literal translation ranges from one word to one word ('hall', Saal, salle, sola, zal) through group to group (un beau jardin, 'a beautiful garden', ein schoner Garten), collocation to collocation ('make a speech', faire un discours), clause to clause ('when that was done', quandcelafut fait), to sentence to sentence ('The man was in the street.' L'homme etait dans la rue.). The longer the unit , the rarer the one-to-one. Further, single-word metaphors ('ray of hope', rayon d'espoir), extended plural-word metaphors ('force someone's hand', forcer la main a quelqu'un) and proverbs ('all that glitters is not gold', tout ce qui brille n'est pas or), illustrate a second figurative semantic scale. I extend literal translation to corre- spondences such as un bilan sanguin, 'a blood check' and apres sa sortie, 'after going out' (but apres son depart, 'after his departure'), since it can be flexible with grammar whilst it keeps the same 'extra-contextual' lexis. Thus, 'literally', arbre is

'tree' not 'shaft', but words like aufheben, einstellen, Anlage have no literal trans- lation. Here, as in many other cases, my definitions are 'operational' to suit translation discussion (rather than theory), not 'rigorous' or 'exhaustive' (and so on) to suit linguistics.

I believe literal translation to be the basic translation procedure, both in communicative and semantic translation, in that translation starts from there. However, above the word level, literal translation becomes increasingly difficult. When there is any kind of translation problem, literal translation is normally (not always) out of the question. It is what one is trying to get away from, yet one sometimes comes back to it with a sigh; partly because one has got used to the sound of what at first seemed so strange and unnatural; beware of this. Une tentation cuisante: can you get nearer than a 'painful' or an 'intense' temptation? 'Burning temptation' is the nearest, it is still not literal. Literal translation above the word level is the only correct procedure if the SL and TL meaning correspond, or correspond more closely than any alternative; that means that the referent and the pragmatic effect are equivalent, i.e. that the words not only refer to the same 'thing' but have similar associations (Mama, 'mum'; le prof, 'the prof) and appear to be equally frequent in this type of text; further, that the meaning of the SL unit is not affected by its context in such a way that the meaning of the TL unit does not correspond to it. Normally, the more specific or technical a word, the less it is likely to be affected by context. Further, a common object will usually have a one-to-one literal translation if there is cultural overlap, though most languages have strange lexical gaps (e.g. 'fingers', 'waist', 'knuckles', 'shins'). A term for a common object sometimes has other common senses ('bank', 'peace') - so that language, particu- larly in English with its monosyllables, appears inefficient.

FAITHFUL AND FALSE FRIENDS

However, my main point is that we must not be afraid of literal translation, or, in particular, of using a TL word which looks the same or nearly the same as the SL word. At school and university I was told I must never do this, but 'theatre' is theatre is Theater is teatro is teatr; only in Czech is it divadlo (the same applies to

'music', where the Czech is hudba). The translation of objects and movements is usually more literal than that of qualities and ways of moving. Many common adjectives of feeling cut up meaning in their own way, so that we cannot trust a transparent translation of 'sincere', 'loyal', 'trivial', 'important', 'truculent','brutal'; only one or two like 'excellent' and 'marvellous' are usually transparent.

And again, the more general and abstract words ('phenomenon', 'element', 'affair') may or may not be translated transparently; there is often a shift at that abstract level (qualite as 'property') but the translation is still usually one-to-one. In general, there are more faithful friends than faux amis, and we must not hesitate to use them, since any other translation is usually wrong. This presupposes that, in context, the readership of O and T have similar interest and language levels. Otherwise the translation may well be different.

Many theorists believe that translation is more a process of explanation, interpretation and reformulation of ideas than a transformation of words; that the role of language is secondary, it is merely a vector or carrier of thoughts. Consequently, everything is translatable, and linguistic difficulties do not exist. This attitude, which slightly caricatures the Seleskovitch School (ESIT, Paris), is the opposite of the one stating that translation is impossible because all or most words have different meanings in different languages, i.e. all words are culture- specific and, to boot, each language has its peculiar grammar. My position is that everything is translatable up to a point, but that there are often enormous difficulties.

WORDS IN THEIR CONTEXT

All the same, we do translate words, because there is nothing else to translate; there are only the words on the page; there is nothing else there. We do not translate isolated words, we translate words all more or less (and sometimes less rather than more, but never not at all) bound by their syntactic, collocational, situational, cultural and individual idiolectal contexts. That is one way of looking at transla- tion, which suggests it is basically lexical. This is not so. The basic thought- carrying element of language is its grammar. But since the grammar is expressed only in words, we have to get the words right. The words must stretch and give only if the thought is threatened.

I am not making any plea for literal or one-to-one translation, since, if it is translationese (and there is far too much translationese published), it is wrong. But the re-creative part of translation is often exaggerated, and the literal part under- estimated, particularly in literary translation, but also in other types of texts which have nothing linguistically wrong with them, which are competently written.

Take the following extracts from an advertisement by Bendicks Ltd, where we might expect the widest divergences:

(1A) 'B are a unique confection, often copied, never equalled.' (IB ) B sont de confection unique, souvent imites maisjamais egales,

(1C) / cioccolatini B sono un prodotto senza eguale spesso imitato, mai eguagliato. (1D) B ist ein einzigartiger Konfekt, der oft nachgeahmt aber nie nachgemacht warden ist.

(2A) 'Blended together they provide a very distinctive and widely appreciated example of the chocolatier's art.'

(2B) Ce melange est I'exemple tres distingue et largement apprecie de /'art du chocolatier.

(2C) La lorofusione e un perfetto esempio dell'arte distintiva e vastamente apprezzata del cioccolatiere.

(2D) - ein ausgezeichnetes und weithin geschdtztes Beispiel fachlichen Konnens.

One notices first how close these translations are; and they could even be closer, being in some cases elegant (and unnecessary) variations on the original, which is presumably English (e.g., mais in IB; senza eguale in 1C, which is blurred by mai eguagliato). Secondly, syntactical changes in the translation appear to be precipitated by the lack of a suitable word for 'blend'. Again, as German cannot risk chocolatier (a pity), it has recourse to the more generic fachlich ('professional'). German also introduces an effective word-play (nachgeahmt, nachgemacht) which alters and improves the sense of the English. (Nachmachen means both 'to make up'

and 'to copy'; there is word-play with nachahmen.)

ELEGANT VARIATIONS

Elegant variations on literal or one-to-one translation are common, and sometimes satisfy the translator's understandable wish to write in a style or phrase that is entirely natural to him. More often, however, they are irritating to the critic, introduced to exhibit the translator's flair for colloquialisms or synonymy, and, even when insignificant, unnecessary. They are not justified in semantic or even communicative translation. They are a temptation (and an indulgence) for any translator.

Literal translation may appear tedious, but there is satisfaction in weighing it against this or that more elegant version and finding it more accurate and econo- mical. Thus the first sentence of Bendicks" advert:

Bendicks of Mayfair have established a reputation respected throughout the world for the manufacture of chocolate confectionery of the highest quality.

'Bendicks of Mayfair' ant ttabli lew reputation, reconnue dans le monde entier, pour la confection de chocolats de la plus haute qualite.

The translation is lexically and grammatically literal, but for 'confectionery', which has to be changed and is satisfactorily conflated with 'manufacture', and the two unnecessary elegant variations, leur and reconnue.

BACK-TRANSLATION TEST (BTT)

The validity of literal translation can sometimes be established by the back- translation test: e.g. (crudely) a 'black frame' should translate back centrally as un cadre noir, ein schwarzer Rahmen, etc. The back-translation test is not valid in the case of SL or TL lexical gaps: thus 'a murky street', 'a bright vision' (or unepersonne maladive) will not translate back satisfactorily. But 'the literary dictates of his time' will never translate back satisfactorily into les modes litteraires de son temps.

Note also that the figurative element in language militates against literal translation when it is a cultural or a stock metaphor, but favours literal translation when it is universal and/or original.

ACCEPTED TRANSLATION

Some transparent institutional terms are translated literally in at least Western European languages even though the TL cultural equivalents have widely different functions: thus 'President', 'Senate', 'Prefect', 'Chancellor', 'Mayor'. Note also that concept-words such as 'radicalism' or 'realism' are translated lit- erally and often misleadingly, as their 'local' connotations are often different. Any 'core' denotative meaning is swamped by the connotative pragmatic meaning. The terms are normally so important in their relation to the TL culture that a literal translation rather than transference is indicated - a translated word more than a transferred one is incorporated at once into the target language. However, for new institutional terms, a translator must be careful about translating the terms directly into the TL, if they already exist but have quite different functions in the TL culture.

CONSTRAINTS ON LITERAL TRANSLATION

I am not suggesting that any more or less context-free SL word must always be translated one-to-one or literally by its 'usual' TL equivalent. The SL word may: (a) be used more frequently (within the register); (b) have a wider semantic range than the corresponding TL word. Thus hardiesse may translate as 'effrontery' (pejorative) as well as 'daring' (positive, honorific) depending on the context. But la plaine which appears almost to coincide in frequency and semantic range with 'the plain' will always translate as 'plain', unless it is the alternative spelling of la plane ('plane').

If a perfectly natural SL unit produces a clumsy literal translation, e.g. il ne parvenait pas a se degager de sa surprise as 'he wasn't succeeding in freeing himself from his surprise', then the translation is 'wrong', however expressive the rest of the SL text ('he was unable to rid himself of his feeling of surprise', 'he couldn't overcome his surprise').

'Ordinary language', which in English is usually descriptive language, not colloquial but neutral, is equally appropriate in written and spoken language, marked by phrasal verbs, familiar alternatives (such as 'bloke', 'kids', 'cash', 'job',

'make love'), empty verbs and verb-nouns ('make his way to'), and can hardly ever be translated literally.

NATURAL TRANSLATION

However, there are all kinds of insidious resistances to literal translation. You may feel it is not translation, it is mechanical, it is automatic, it is humdrum, it is not clever. You have been told at school not to practise it. It does not enrich your knowledge either of the source or of the target language. It is too easy. We have to resist these arguments. Apart from translationese (i.e. inaccurate translation) the only valid argument against what I might find an acceptable literal translation of an ordinary language unit is that you find it unnatural. Take I'heure est venue or les maisons basses: if you insist you would not normally say 'the hour has come', only

'the time has come'; not 'the low houses', only 'the squat or low-lying houses', I would suspect you were deluding yourself, but I believe that, except for an expres- sive text, you should write in a manner natural to yourself, a manner that expresses your own sense of good style. This is yet another of the tensions within translation. In fact, by repeating several times to yourself a slightly 'unnatural' unit of language, or by saying it in a soft tone of voice, you can sometimes make it sound more natural, and convince yourself it is a good translation. If it still remains unnatural to you, you should avoid it. In this sense, the argument in favour of a translation having the impress of a translator's own way of writing has precedence over the principle of literal translation.

Note that it is sometimes advisable to retreat from literal translation when

faced with SL general words for which there are no 'satisfactory' one-to-one TL equivalents even though one is over-translating. Thus Darstellungen is more common and concrete than 'representations', and, in the context, 'drawings',

'pictures' or 'diagrams' may be quite suitable.

A further point. One can say that, in the human view, all objects are symbols

and all words are either representations or symbols of objects. In this sense, literal translation can go either way. Commonly, atropinique can mean, literally, 'made of

atropin' or 'atropin-like'; brulure, a 'burn' or 'a burn-like sensation'. Less commonly, Die Apfelsinne fdllt mir ein may literally mean 'The orange' (previously mentioned) or 'The idea of the orange occurs to me'.

RE-CREATIVE TRANSLATION

Literal translation is the first step in translation, and a good translator abandons a literal version only when it is plainly inexact or, in the case of a vocative or informative text, badly written. A bad translator will always do his best to avoid translating word for word. Re-creative translation - 'contextual re-creation' as Delisle (1981) calls it - which means, roughly, translating the thoughts behind the words, sometimes between the words, or translating the sub-text, is a procedure which some authorities and translation teachers regard as the heart or the central issue of translation ('get a. far away as possible from the words'). The truth is the opposite: 'interpret the sense, not the words' is, to my mind, the translator's last resource; an essential resource, certainly, and a touchstone of his linguistic sensi- tivity and creativity, not to mention his alertness and perspicacity, when words mislead. Further, contextual re-creation is likely to be more common in interpreta- tion, if delegates are speaking off the cuff, than in written language translation, where words are more carefully measured and perhaps closer to thought. But most translation is not creative in this sense. You have to like struggling with words before you reach the longer passages.

In recent years, London has, in its notices and advertisements, become a trilingual often a quadrilingual (you add Dutch) city. Coming so late to reality, the British Transport Board has profited from others' mistakes and has done well. One of the most striking things about the adverts (of all things) has been how close to each other they are, how near to literal translation:

Que ce soil votre premiere ou votre cinquantieme visile en Grande-Bretagne, parions que, avant la fin de la joumee, vous n'aurez pas manque de remarquer milk curiosites nouvelles, typiques du pays et de ses habitants.

Ob Du Groftbritannien sum ersten oder zum funfzigsten Mai besuchst, wetten wir, daft Du jeden Tag immer noch neue Besonderheiten bei Land und Leuten entdeckst.

THE NOTION OF THE 'NO-EQUIVALENT' WORD

The difficulties of literal translation are often highlighted not so much by linguistic or referential context as by the context of a cultural tradition. Bagehot wrote about 130 years ago that 'Language is the tradition of nations . . . people repeat phrases inculcated by their fathers, true in the time of their fathers but now no longer true.'

If you consider Faust's famous struggle to translate the word logos, a word that is virtually context-free, and therefore has to be translated for itself (Weinrich's notorious slogan 'Words are untranslatable, texts can always be trans- lated' - see his brilliant book Linguistik der Luge - is salutary but sometimes the reverse of the truth), how Faust moves hesitantly and subjectively from Wort ('word'), Sinn ('sense', 'meaning', 'thought'), Kraft ('strength', 'power', 'force') to finally Tat ('deed', 'fact', 'action', 'activity') and making his own comments quite independently of the Greek or the referential truth ('I can't possibly rate the Word as highly as that -1 must translate it differently, if only my mind will make it clear to me, so I'll write "sense", "meaning" and I have to think carefully, I'll have to think that line out again, not be over-hasty, can it be "sense" which makes and produces everything, I'll write "force" ("strength", "power") but as I write that, something is warning me I can't stay with that, so I can safely write "deed", "act", "action"') - all this illustrates a painful struggle with four key words, one of which, Kraft according to Gadamer (1976), is conditioned, not by its context in the play or the New Testament, but by its past - its connection with Newtonian physics and its development (integration) in the German public consciousness by Ottinger and Herder: 'the concept of force was made comprehensible on the basis of the living experience of force. As this integration occurred, the technical concept grew into the German language and was individualized to the point of becoming untranslatable.'

To write off as 'untranslatable' a word whose meaning cannot be rendered literally and precisely by another word is absurd, particularly when it could at least be better delineated by componential analysis into four or five words, though as a footnote, not in the text of the play. Looking at translation in an ideal sense,

Gadamer has pointed out that 'no translation can replace the original . . . the translator's task is never to copy what is said, but to place himself in the direction of what is said (i.e. in its meaning) in order to carry over what is to be said into the direction of his own saying'. Again, this reliance on the vouloir-dire and the significance of what the SL text deliberately left unsaid can be dangerous, and applies only to the most difficult texts, where some kind of interpretation and hermeneutics are essential if the translator is to be active, to 'become again the one saying the text'. Here the moment of period and time, as well as the translator's personality, the judgments he has made in the course of his emotional and intel- lectual development, the pre-judgments (Vorurteile) and preconceptions with which he meets a particular problem (after a year, he will translate the same text in a different way: is this chance or personal change?) - all this is important when one considers translating texts that appear to be on the borders of language and thought, and the struggle is with grammar as well as words, the nuances of mood (modals), and time (tense) and duration (aspect).

But in the vast majority of cases, Gadamer is not going to help the translator at all. His statement 'No translation is as understandable as the original' is mis- leading. Many translations have been and are a good, simple introduction, a lead-in into the original - particularly translations of languages such as German with an artificial word-order inflicted on them by their scribes, their clercs, i.e. the in fact non-SVO (subject-verb-object) languages, which postpone the lexical elements of their verbs to the end of the proposition. However, Gadamer uses his statement

'the most inclusive meaning of what's said comes to language only in the original

saying and slips away in all subsequent saying and speaking' as an argument against literal translation. 'Hence the translator's task must never be to copy what is said', which seems dangerous to me, suggesting that the translator has to anticipate

changes of meaning into language of the future. In fact, Kant, Hegel, Heidegger

and Gadamer are initially more understandable in translation than in the original.

However, where Gadamer is a healthy corrective is in his insistence on the

personality of the translator, the translator's state of consciousness and awareness, the limitations of the usually recognised types of context.

THE ROLE OF CONTEXT

My last point, in fact, is that, in translation, the translator indeed has to be aware of all the varieties of contexts - so many it is idle to list them again - but this does not mean that context is the overriding factor in all translation, and has primacy over any rule, theory or primary meaning. Context is omnipresent, but it is relative. It affects technical terms and neologisms less than general words; it permeates a structured text and touches disjointed texts rather lightly. Where a writer deliber- ately innovates, the translator has to follow him, and blow the context.

A translator with his eye on his readership is likely to under-translate, to use more general words in the interests of clarity, simplicity and sometimes brevity, which makes him 'omit' to translate words altogether. (A translator has to account for every SL word, not to translate it.) Under-translation is justified if an informa- tive text is deficient in clarity. It is not justified if it is unnecessary and is a mere retreat from a literal translation. You must not write down to your reader.

A good literal translation must be effective in its own right. If it shows SL interference, that must be by the translator's conscious decision. Some mild translationese in a tourist brochure has a gentle charm, like the local colour of a transferred word. 'New means of creative work ripened and brought fruit later.' The translator unconscious of SL interference is always at fault. The less context- bound the words (e.g. lists, technical terms, original metaphors, 'unacceptable' collocations), the more likely a literal translation - whilst the more standard are the collocations, colloquialisms, idioms, stock metaphors, the less likely is a literal translation. Inevitably, there is a proper place for literal translation as a procedure in all good translations.

Half the misunderstanding about translation in Britain is due to the fact that so many teachers tell their pupils to avoid translating an SL word by a similar-looking TL word whenever possible. Thus the pupils expand their TL vocabulary and distort their translations.

From Korunets:

Literal translating, which is to be employed when dealing with separate words whose surface form and structure, as well as their lexical meaning in the source language and in the target language, fully coincide. These are predominantly international by original morphemes, lexemes/words, rarer word-groups having in English and Ukrainian (and often in some other languages too) a literally identical or very similar presentation and identical lexical meaning: administrator адміністратор, director директор, region регіон, hotel/motelготель/мотель, hydrometer гідрометр, Tom Том, etc.

It is easy to notice that in some words thus translated not only the morphemic but also the syllabic structure can sometimes be fully conveyed (cf. an-f/ан-ти, es-cort ес-корт, d/recf-огдирект-ор, ex-pres/d-enfeKC-прези-дент).

In many a case, however, the lingual form of the source language words is only partly conveyed in the target language. This happens when the common word is borrowed by each of the two languages in question from different source languages or when its lingual form is predetermined by the orthographic peculiarities of the target language: anti-alcoholic протиалкогольний, music музика, constitution конституція, zoology зоологія, atomic weight атомна вага, chemical process хімічний процес, national opera theatre національний оперний театр.

The literal translation in all the examples above provides an equivalent rendering of the lexical meaning of each language unit despite the augmentation in the number of morphemes/syllables in the target language (cf. atom/сатом-н-а, а/сопо//салкоголь-н-ий).

Close to the literal method of translating or rather reproduction is the so-called way of translator's transcription by means of which partly the orthographic and partly the sounding form of the source language lexemes/words is conveyed: archbishop архієпископ, card картка/карта, European європейський, Muscovite москвич/москвичка, organized організований, stylized стилізований, fnerapeuf/стерапевтичний.

Literal translating, however, can pervert the sense of the source language lexemes or sentences when their lingual form accidentally coincides with some other target language lexemes having quite different meanings: artist митець, художник (rarer артист); decoration відзнака, нагорода (and not декорація); matrass колба (and not матрац); replica точна копія (and not репліка), etc. Hence, literal translating has some restrictions in its employment and does not always help to render the lexical meaning of words or even morphemes having often even an identical lingual form (spelling) in the source language and in the target language.

1 Maksimov, G.Miram, O. Seme nets, S.Shvachko, O.Vasylchenko, etc), Rus­sian (L Barkhudarov, G.Chernov, A. Chuzhakin, A.Fedorov, V.Komissarov, R.Minyar Beloruchev, Ya. Retsker, A. Shveitser, etc) and western (M Baker, M.A.K Halliday, D. Hatim, R. Jones, J Munday, V. Taylor-Bouladon. L.Vysson, etc) scholars.

According to this understanding translation is a process of transforming speech messages in the source language (SL) into the speech messages in the target language (TL) under condition that theirsense and communicative in­tention remain unchanged [Чернов 1987: 6]. It is quite natural that in the

84