INTRODUCTION

Theorists sometimes maintain that cognitive translation (the transfer of cold information) is perfectly possible and may be possibly perfect - it is the hard core, the invariant factor; the only snag comes when: (a) there is an emphasis on the form as well as the content of the message or; (b) there is a cultural gap between SL and TL readers (different ways of thinking or feeling, material objects) or there is a tricky pragmatic relation, i.e. between on the one hand the writer and on the other the translator and/or reader. There is a certain truth in these generalisations, though they miss one point, that the adequacy of a translation basically depends on the degree of difficulty, complexity, obscurity of the whole passage, rather than the one or the other aspect. Further, any passage that stresses SL form can be perfectly explained and therefore over-translated into the TL, though it will not have the naked impact of the original. However, if one must make generalisations, I can say that normally the translation of serious literature and authoritative statements is the most testing type of translation, because the first, basic articulation of meaning (the word) is as important as the second (the sentence or, in poetry, the line) and the effort to make word, sentence and text cohere requires continuous compromise and readjustment.

Biihler's expressive function of language, where content and form are on the whole equally and indissolubly important, informs two broad text-categories: serious imaginative literature and authoritative statements of any kind, whether political, scientific, philosophical or legal.

The two categories have obvious differences: (a) authoritative statements are more openly addressed to a readership than is literature; (b) literature is allegorical in some degree; authoritative statements are often literal and denotative and figurative only in exceptional passages, as in broad popular appeals ('islands' amongst the literal language), such as 'The wind of change is blowing' - Un grand courant d'atr souffle (both stock metaphors); 'I have nothing to offer but blood, toil, tears and sweat' (Churchill, 13 May 1940) -Je n'ai a vous offrir que du sang, de la sueur, du travail, des larmes (figurative language, but these are symbols, to be understood literally as well as figuratively); 'the underbelly of the Axis' (Churchill, January 1943) - le bas-ventre de l'Axe - not le point vulnerable (an original metaphor). Further, the element of self-expression in authoritative statements is only incidental but the translator has to pay the same respect to bizarreries of idiolect as in fantastic literature: La France y volt un renfort decisifde noire latinité a L'avantage de tous les hommes - 'France sees it as a decisive strengthening of our Latinity benefiting all men' (De Gaulle).

A further generalisation for the translator: literature broadly runs along a four-point scale from lyrical poetry through the short story and the novel to drama.

POETRY

Poetry is the most personal and concentrated of the four forms, no redundancy, no phatic language, where, as a unit, the word has greater importance than in any other type of text. And again, if the word is the first unit of meaning, the second is not the sentence or the proposition, but usually the line, thereby again demonstrating a unique double concentration of units. Thus in:

... But Man, proud man Brest in a little brief authority,

Most ignorant ofwhat he's most assured His glassy essence, like an angry ape,

Plays such fantastic tricks before high heaven As make the angels weep...

(Shakespeare, Measure for Measure, II.II. 117)

the integrity of both the lexical units and the lines has to be preserved within a context of: (a) corresponding punctuation, which essentially reproduces the tone of the original; and (b) accurate translation of metaphor. Consider Tieck's version:

... dock der Mensch, der stolze Mensch,

In kleine kurze Majestat gekleidet,

Vergessend, was am mind'sten zu bezweifeln,

Sein gldsem Element - wie zom 'ge Affen,

Spieltsolchen Wahnsinngaukelndvordem Himmel,

Daft Engel weinen...

(trans. Tieck and Schlegel, MaftfiirMaft)

Here the word - and line - units have been preserved with the punctuation; the image 'plays such fantastic tricks' becomes 'plays such madness, conjuring' but the other images are preserved, whilst 'most ignorant of becomes 'forgetting' and the positive 'most assured' becomes the double negative 'least to be doubted', which is a common modulation. The greatest and unnecessary loss here is the 'fantastic tricks' metaphor. Original metaphor is the controlling element in all creative language, evoking through a visual image - even abstract images such as justice or mercy become people or objects - not only sight but the four other senses (e.g., fur as touch, food as taste, flowers as smell, bells or birds as sound) as well as the concomitant human qualities, good or evil, pleasure or pain, that these images (sensory, sensuous, sensual, sensitive, perhaps even sensational, to liven up language) can produce. Poetry presents the thing in order to convey the feeling, in particular, and however concrete the language, each represents something else - a feeling, a behaviour, a view of life as well as itself. Original metaphors the translator has to reproduce scrupulously, even if they are likely to cause cultural shock. Shakespeare's 'Shall I compare thee to a summer's day' (Sonnet 18), as Neubert has commented, will leave Arabic or Eskimo readers cold, but the Arabic or Eskimo reader must make the effort to fint out the truth of the simile, which is at least half-revealed in the next line: 'Thou art more lovely and more temperate'. A cultural metaphor (e.g., the technical term '(Summer's) lease') is not so important.

The translator can boldly transfer the image of any metaphor where it is known in the TL culture. But for lines such as Walter de la Mare's:

And even the thought of her when she is far

Narcissus is, and they the waters are

(Reflections)

or Kingsley Amis':

Should poets bicycle-pump the heart

Or squash it flat?

(Something Nasty in the Bookshop)

faced with literal translations in cultures where Narcissus and the bicycle-pump are not known, the reader is not so much culturally shocked as baffled. In such poems there is a case for creating a culturally equivalent TL metaphor, or converting the SL metaphor to sense or, where there is space, adding sense to the metaphor; but if the translator regards the metaphor as important, it is his duty to carry it across to launch it on the target language and its culture.

Whilst I think that all images have universal, cultural and personal sources, the translator of poetry cannot make any concession to the reader such as transferring the foreign culture to a native equivalent. If autumn in China is the season not of Keats's 'mists and mellow fruitfulness' but of high clear skies and transparent waters, and the sound of clothes laundered for the cold weather pounded on the washing blocks, then the reader must simply accept this background and, if he wants to feel it, repeated reading is more likely to make it his possession than are detailed background, explanation of allusions and so on. Nevertheless, the European must be aware that, forthe Chinese culture, jade is not jade-coloured but white ('jade snow', 'jade beads', 'jade moon'), that comparisons with eyebrows assume the custom of painting women's eyebrows green, that the phoenix has no myth of resurrection, that dragons are close and kindly, that cypresses suggest grave-yards, as in the West (see Graham, Poems of the Late Tang).

The transition from Chinese to English culture is made easier because all the images mentioned are not unfamiliar to an English reader. The difficulty comes when and if local flowers and grasses are used as metaphors.

I am sceptical about the idea that a translator of poetry is primarily com-municating - that he is, to his readers in the conventional definition of communicative translation, trying to create the same effect on the target language readers as was created by the poet on his own readers; his main endeavour is to 'translate' the effect the poem made on himself. A translator can hardly achieve even a parallel effect in poetry - the two languages, since all their resources are being used here as in no other literary or non-literary medium, are, at their widest, poles apart. Syntax, lexis, sound, culture, but not image, clash with each other. Valery wrote: 'My aim is not literary. It is not to produce an effect on others so much as on myself- the Self in so far as it may be treated as a work... of the mind. I am not interested in writing poetry without a view to its function.'

Compare John Caimcross, who was not trying to disprove that French, or poetry, or French poetry, or Racine, was untranslatable, or to present Racine to his English readers, or to present his English readers with Racine, but set about translating simply because the English words started forming themselves in his ear, and so he quotes Racine again: Ce quej 'aifait, Abner, fai cru le devoirfaire - 'What I have done, Abner, I had to do' (Athalie, 1.467), which is itself an echo of yeypaAa, ytypafya - Pontius Pilate's 'What I have written I have written'. Take it or leave it.

Now I think that in most examples of poetry translation, the translator first decides to choose a TL poetic form (viz. sonnet, ballad, quatrain, blank verse etc.) as close as possible to that of the SL. Although the rhyming scheme is part of the form, its precise order may have to be dropped. Secondly, he will reproduce the figurative meaning, the concrete images of the poem. Lastly the setting, the thought-words, often the various techniques of sound-effect which produce the individual impact I have mentioned have to be worked in at later stages during the rewriting (as Beaugrande has stated in his fine translation of Rilke). Emotionally, different sounds create different meanings, based not on the sounds of nature, nor on the seductive noises in the streams and the forests, but on the common sounds of the human throat: Sein odernichtsein - das isthierdie Frage appears to have a ring of confidence and challenge in it which is foreign to Hamlet's character - is it the redoubled ei sound? - that opens up the whole question of the universal symbolism of sounds. All this plangency, this openness is missing in 'To be or not to be- that is the question' which is almost a word-for-word translation, though the German hier 'that is here the question' - appears to underline the challenge which is not in Shakespeare. The fact is that, however good as a translation, its meaning will differ in many ways from the original - it will, in Sorrow's phrase, be a mere echo of the original, not through Gogol's glass pane - and it will have its own independent strength. A successfully translated poem is always another poem.

Whether a translator gives priority to content or manner, and, within manner, what aspect - metre, rhyme, sound, structure - is to have priority, must depend not only on the values of the particular poem, but also on the translator's theory of poetry. Therefore no general theory of poetic translation is possible and all a translation theorist can do is to draw attention to the variety of possibilities and point to successful practice, unless he rashly wants to incorporate his theory of translation into his own theory of poetry. Deliberately or intuitively, the translator has to decide whether the expressive or the aesthetic function of language in a poem or in one place in a poem is more important. Crudely this renews Keats's argument concerning Truth and Beauty: 'Beauty is Truth, Truth, Beauty - that is all you know, and all you need to know', when he maintains that they define and are equivalent to each other, as well as the later argument between art as a criticism of life (Matthew Arnold) and art for art's sake (Theophile Gautier) which characterised two French poetic movements as well as much turn-of-the-century literature - 'All art is useless', wrote Oscar Wilde, whose own art belies the statement. Clearly Keats, who was not thinking of translation, oversimplified the argument. If Truth stands for the literal translation and Beauty for the elegant version in the translator's idiom, Truth is ugly and Beauty is always a lie. 'That's life', many would say. But a translation theorist would point out that both these versions, the literal and the elegant poem, would normally be equally unsatisfactory as translations of a poem or of anything else. Some fusion, some approximation, between the expressive and the aesthetic function of language is required, where in any event the personal language of the poet which deviates from the norms of the source language is likely to deviate even more from those of the target language. Thus Karl Kraus complained that Stefan George, by 'doing violence' to the English sense of Shakespeare's sonnets and to German verbal and grammatical usage, had produced 'a unique abortion'! But, in my belief, George is the closest and most successful of all translators.

Thus:

Lebwohl! zu teuer ist dein besitzfur mich Und du weiftt wohl wie schwer du hist zu kaufen Der freibriefdeines wens entbindet dich Mein recht aufdich ist vollig abgelaufen.

which is:

Farewell! too dear is your possession for me And you well know how hard you are to buy The charter of your worth releases you My claim to you has fully run its course.

which becomes:

Farewell! thou art too dear for my possessing,

And like enough thou know'st thy estimate:

The Charter of thy worth gives thee releasing;

My bonds in thee are all determinate.

(Sonnet 87)

George's translation is notable for its tautness and flexibility, and particularly for its emphasis on the correspondingtheme-words ('dear', 'charter', 'releasing', 'bonds', 'determinate'). Where he is unable to reach Shakespeare is in the polysemy of 'estimate', 'releasing', 'bonds', and 'determinate', and thus he restricts the meaning of the quatrain - and above all in the splendid logical statement of Shakespeare's opening with its communicative dynamism on 'possessing', where George is forced into an inversion.

Angus Graham, in his discussion on the translation of Chinese poetry, says that the element in poetry which travels best is concrete imagery. A crib or trot of Chinese such as:

Kuang Heng write-frankly memorial.

Success slight Liu Hsiang transmit classic.

Plan miss.

could be rewritten as:

A disdained K'uang Heng, as a critic of policy,

As promoter of learning, a Liu Hsiang who failed.

Here the poet is miserably contrasting his failures with the success of two statesmen, but contrast this with:

Tartar horn tug North wind,

Thistle Gate whiter than water

Sky, hold-in-mouth Koknor Road

Wall top moon thousand mile

I note that, even in a Times Literary Supplement review, Erich Segal comments on most translators 'metarophobia', their unease in the presence of metaphor. Pindar speaks of man being skias onar 'the dream of a shadow' but Richmond Lattimore turns it round to the conventional 'shadow of a dream'. According to Aeschylus, Prometheus stole the anthoslpyros, the 'blossom of fire', but according to half the translators he merely 'plucked the blossom'.

Dichten equals condensare, as Ezra Pound wrote in ABC of Reading, mistakenly thinking that dichten is related to dicht ('dense' or 'thick'), but stating a truth. Original poetry itself has no redundancy, no phatic language, but the translator usually needs a little extra space, he relies on redundancy in overtranslating, say, veule as 'flabby' or 'weak and soft' and here he is often hemmed in by the metre. Racine's wonderful line Le journ'est pas plus pur que le fond de mon coeur may become: 'My heart is candid as the light of day' (Dillon) or: 'The daylight is not purer than my heart' (Caimcross) and whilst the second translation is closer and more successful, it cannot match the fullness and softness of the original; the alliteration, the monosyllables, the repeated r's, the emotive fond are missing.

I have said that original metaphors have to be translated accurately, even if in the target language culture the image is strange and the sense it conveys may only be guessed. Undjener, der'du'zuihmsagtetrdumtmitihm: Wir(Ce\an, InMemoriam Paul Éluard): 'And he who addressed him as "thou" will dream with him: We.' The translator Michael Hamburger has to use 'thou', although the connotations of friendship and love - what I would call le plaisir de te tutoyer - will be lost on the reader of the translation or perhaps soon on the reader of the original, now that the intimate <1ii, tu has been taken over by the Left and all the under-thirties. Le plaisir de te tutoyer has almost gone, unless you are old, but so, thankfully, has das erste Du stammelte aufihren keuschen Lippen.

Sound-effects are bound to come last for the translator, except for lovely minor poetry such as Swinburne's. Inevitably, he must try to do something about them and, if not, compensate, either by putting them elsewhere or substituting another sound. German, the Brudersprache to English, oftenfinds its adjectives and nouns - fremde Frau, 'alien woman'; laue Luft, 'tepid air' - unreproduced, but longer alliterations.

Und schwdlle mil und schauerte und triefte (G. Benn, Untergrundbahn)

can usually find a modest, suggestive equivalence:

To swell in unison and stream and shudder (trans. M. Hamburger)

John Weightman has stated that French poetry is untranslatable into English. I cannot accept this. Firstly, because a lot of French poetry (Villon, Rimbaud, Valery) has been more or less successfully translated into English; secondly because although there are obvious minuses - the syntactical differences; the huge English vocabulary compared with the small French vocabulary, so that many French words appear to be generic words covering many English specific words that themselves lack a generic word (e.g., humide, mouille: 'humid, damp, dank, moist, wet, clammy, undried'; noir. 'black, dark, dim, dull, dusky, deep, gloomy, murky'), making French 'abstract' and intellectual whilst English is concrete and real-yet, in the actual particular of a text, English has infinite creative resources, English has the disyllables as well as the monosyllables, English in the eighteenth century got close to all the so-called French properties and, given empathy, given sympathy, there is no reason why, one day, even Racine should not find his inadequate but challenging English translator.

John Caimcross sets out three considerations for the translation of Racine:

(1) the translator must adopt ten-syllable blank verse; (2) Racine must be translated accurately; (3) Racine's verse is particularly difficult owing to his capacity of evoking music from the most unpromising material -1 could think of more.

Hippolyte's confession of love -1 would not call it that, it is too restless and feverish - to Aricie (Phedre, 11.524-60) is often considered to be precieux, i.e.,

affected, conventional, too polished, sophisticated, class-bound, with too many stock metaphors, but for me they have always been Racine's most beautiful lines, crystallising the neurotic exposed mental and nervous obsession which is the essence of the Racine characater. Taking the critical lines 539-48, it appears to me that in any modem version, the language must be kept modem and formal, the polar oppositions (fuir, trouver, suivre, eviter, lumiere, ombre) retained, the stresses and repetitions preserved, the image of the hunted, haunted animal (Hippolyte) kept clear, and some attempt made to keep the simple language, the soft sounds with occasional alliteration.

Consider first the version of John Caircross. In general it is accurate, though a new image is unnecessarily created ('Cut from my moorings by a surging swell') and some oppositions blurred:

Present I flee you; absent you are near

Presenteje vousfuis - absente,je vous trouve

and the stresses often changed. The translation, written in 1945, has a few old- fashioned phrases: 'in thrall', 'a single blow has quelled'. With all this, lines such as

Before you stands a pitiable prince...

Who, pitying the shipwrecks ofthe weak...

Deep in the woods, your image follows me.

Dans lefond desforets votre image me suit

The light of day, the shadows of the night.

La lumiere dujour, les ombres de la nuit.

(the latter a one-to-one translation)

Everything conjures up the channs I flee

I seek but cannot find myself again -

Maintenantje me cherche, et ne me trouve plus

(note the unusual number of French monosyllables) are close to the original and successful.

George Dillon, like Caimcross, uses blank verse, and prefers formal accuracy to musicality. His translation is closerthan Caimcross's, so that lexical inaccuracies such as 'surprise' for trouble and 'hurt' for dichireram disconcerting, as is the weak line: 'Your image moves with me in the deep forests' (the alliteration is compensatory). Some stresses and contrasts are more clearly rendered than Caimcross's:

With you and with myself, vainly I strive

(1.541)

All summon to my eyes what I avoid

(1.545)

I seek myself and find myself no more

(1.548)

(the latter the most successful line) - such lines show how simply and precisely Racine can be translated. Both Dillon and Caimcross hit on the same translation for line 544 and there are occasions where one or two lines of Dillon's could improve Caimcross's rather better overall version; Dillon's

Only my deep sighs echo in the wood;

My idle couriers have forgotten my voice.

is better than Caimcross's

My idle steeds no longer know my voice And only to my cries the woods resound.

(I do not know why Caircross has reversed the lines.)

Robert Lowell's 'imitation' of Phedre is another matter. These rhymed pentameters attempt to explicate the image of the speech:

Six months now, bounding like a wounded stag I've tried to shake this poisoned dart, and drag Myself to safety from your eyes that blind When present, and when absent leave behind Volleys ofburning arrows in my mind.

I do not know how such lines would strike a reader or spectator new to Phedre. For myself, with Racine's images burned into my mind, I find them unsatisfying, because, like Hippolyte, I am continuously looking for and failing to find even the simplest images which Lowell would have had no difficulty in retaining or recapturing. In fact I find the greatest loss in Racine translations is the resonance of the only 1800 words that are used in the twelve plays.

THE SHORT STORY/NOVEL

From a translator's point ofview, the short story is, of literary forms, the second most difficult, but here he is released from the obvious constraints of poetry - metre and rhyme - whilst the varieties of sound-effect are likely to play a minor role. Further, since the line is no longer a unit of meaning, he can spread himself a little - his version is likely to be somewhat longer than the original though, always, the shorter the better. He can supply cultural glosses within the text - not, as in poetry or drama, delete or banish them to some note or glossary: L 'ascenseur ne fonctionnait pas, en raison des Economies de courant - 'With the war-time electricity cuts, the lift wasn't working.'

Since formal and thematic concentration and unity may distinguish the short story from the novel, the translator has to be careful to preserve certain cohesive effects.

The second type of key-word is the word or phrase that typifies the writer rather than the particular text: sich verirrenjagen, beirrt, ndmlich, bccngcn and all the Beamte words may be said to typify Kafka, as powerful verbs like entrainer, tpier, agir, frtmir, exiger, grelotter, tressaillir, obse'dermay typify Mauriac. Some of these words go into a ready one-to-one translation into English, and get their connotational significance from repetition and context (situational and linguistic) which can more or less be reproduced by the translator. Words like jagen and entrainer are difficult: jagen suggests 'hectic chase' and entrainer (Quelle force m entraine?*), 'impelirresistibly'.

For key-words, translators have to assess their texts critically; they have to decide which lexical units are central, and have the more important function, and which are peripheral, so that the relative gains and losses in a translation may correspond to their assessment. (I realise that many translators will claim they do all this intuitively, by instinct, or by common sense, and they do not need translation theory to make them aware of relative importance.)

There is no advantage in making generalisations about the translation of serious novels. The obvious problems: the relative importance of the SL culture and the author's moral purpose to the reader - it may be exemplified in the translation of proper names; of the SL conventions and the author's idiolect; the translation of dialect; the distinction between personal style, literary convention of period and/or movement; and the norms of the SL - these problems have to be settled for each text.

The signal importance of the translation of some novels has been the introduction of a new vision injecting a different literary style into another language culture, and when one looks at Weltliteratur translations in this sense -1 think of Proust, Camus, Kafka, Mann, Pavese - it is clear that the translators have often not been bold, which means not literal, enough: these are the million cases where a literal translation is aesthetically not inferior to a free translation, fashionably justified as 'sub-text', formerly the 'spirit' or the 'genius' of the language or the author.

DRAMA

The main purpose of translating a play is normally to have it performed successfully. Therefore a translator of drama inevitably has to bear the potential spectator in mind though, here again, the better written and more significant the text, the fewer compromises he can make in favour of the reader. Further, he works under certain constraints: unlike the translator of fiction, he cannot gloss, explain puns or ambiguities or cultural references, nor transcribe words for the sake of local colour: his text is dramatic, with emphasis on verbs, rather than descriptive and explanatory. Michael Meyer, in a little noticed article in Twentieth Century Studies, quoting T. Rattigan, states that the spoken word is five times as potent as the written word - what a novelist would say in 30 lines, the playwright must say in five. The arithmetic is faulty and so, I believe, is the sentiment, but it shows that a translation of a play must be concise - it must not be an over-translation.

Meyer makes a distinction between dramatic text and sub-text, the literal meaning and the 'real point': i.e. what is implied but not said, the meaning between the lines. He believes that if a person is questioned on a subject about which he has complex feelings, he will reply evasively (and in a circumlocutory manner). Ibsen's characters say one thing and mean another. The translator must word the sentences in such a way that this, the sub-text, is equally clear in English. Unfortunately, Meyer gives no examples. Normally one would expect a semantic translation of a line, which may be close to a literal translation, to reveal its implications more clearly than a communicative translation, that simply makes the dialogue easy to speak. Lines such as 'Aren't you feeling the cold?' and 'I think your husband is faithful to you' have potential implications of escape and suspicion respectively in any language, provided there is cultural overlap between them. (They would not have the same implication if the climate or the sexual morality respectively differed considerably in the SL and the TL culture.)

Finally a translator of drama in particular must translate into the modern target language if he wants his characters to 'live', bearing in mind that the modem language covers a span of, say, 70 years, and that if one character speaks in a bookish or old-fashioned way in the original, written 500 years ago, he must speak in an equally bookish and old-fashioned way in the translation, but as he would today, therefore with a corresponding time-gap - differences of register, social class, education, temperament in particular must be preserved between one character and another. Thus the dialogue remains dramatic, and though the translator cannot forget the potential spectators, he does not make concessions to them. Given the emphasis on linguistic form, and the subtlety of the SL, his version is inevitably inferior but also simpler and a kind of one-sided introduction to the original. Kant is easier to read in French than in German, perhaps even for a German.

Whilst a great play may be translated forthe reading public's enjoyment and for scholarly study as well as for performance on stage, the translator should always assume the latter as his main purpose - there should be no difference between an acting and a reading version - and he should look after readers and scholars only in his notes. Nevertheless, he should where possible amplify cultural metaphors, allusions, proper names, in the text itself, rather than replace the allusion with the sense. ('Hyperion to a satyr' becomes 'a sun god to a monster' in Chinese.)

When a play is transferred from the SL to the TL culture it is usually no longer a translation, but an adaptation.

CONCLUSION

Finally in discussing the translation of serious literature, I must make it clear that I am trying to look at the future. There is no question that translators such as Stuart Gilbert, who translated Malraux and Camus into English and Joyce into French, had a quickening effect on translation: possibly reacting against the stiff and literary translation style which so fouled up the translation of Russian literature at the turn of the century. Profoundly influenced by Hemingway who was mainly responsible for bringing fiction closer to normal speech, Gilbert produced a lively enough equivalence: Aujourd'hui, maman estmorte oupeut-etre hier, je ne saispas becomes 'Mother died today, or maybe yesterday, I can't be sure'; Je prendrai I'autobus a deuxheures etj'arriverai dans Fapres-midi. - 'With the two o'clock bus I should get there well before nightfall' (examples from Camus, L Ftranger). You can see that half the time Gilbert is trying to be more colloquial than the original, yet every time he might have said that the further colloquialism was in the sub-text, i.e., implied or implicated in the original. Nevertheless it is hard to see how one can justify translating Ilfaisait tres chaudas 'It was a blazing hot afternoon', and there are a thousand other examples of such 'deviations' which show that these translators may have been aiming at 'intuitive truth', an instinctive naturalness (there is no question usually of ignorance, of carelessness, such as is so common in translations from the German) rather than accuracy at any level. I am suggesting that some kind of accuracy must be the only criterion of a good translation in the future

- what kind of accuracy depending first on the type and then the particular text that has been translated - and that the word 'sub-text' with its Gricean implications and implicatures can be made to cover a multitude of inaccuracies.