A few years ago, I was staying, along with a lot of other British academics, in a small village in Sachsen-Anhalt called Wust. We were there for a cultural festival with a strong emphasis on English-language literature, and most of us were accommodated in the village. One day an American professor staying in a self-catering cottage found himself needing to buy some clothes-pegs in the village shop.
Now you might think this sounded easy, but of that gathering of literati, all of whom could have translated chunks of German poetry at the drop of a hat, how many do you suppose knew the German for clothes-pegs? Goethe and Schiller just don't mention them very often, as it happens.
So a bunch of us had to sit down and work it out logically. This was quite easy – German is a logical language – and besides we cheated and asked the shopkeeper. "Wäscheklammer", it turned out to be. What is interesting is the "Wäsche" half. You could not come at this by translating "clothes" literally into German. This would give you "Kleider". But as soon as you say it, you know that isn't what a German would call clothes hung on a line. Kleider are clothes in a shop, or hanging in a wardrobe, or on the person. When in a laundry basket or washing machine, or hung on a line to dry, they become "Wäsche" – washing.
This is a trivial example of how translation isn't just a matter of different words for the same thing, but of different perceptions of the same thing. In English, washing is always washing but clothes can be washing too, depending on where they are at the time. German makes a clear distinction between clothes in a state to be worn and washing, which isn't. When, reading mediaeval German poetry, I first came across the phrase "as shy as a peacock" I was somewhat surprised. The logic, again, is impeccable: the peacock, rather than displaying its tail all the time, mostly keeps it hidden, dragging on the ground, and its gaze too is usually fixed on the ground. Languages bear a strong likeness to the six blind men in the Hindu proverb, feeling different bits of the elephant and coming up with different ideas of it.
Everyone who has learned another language has gone through the stage of filing one word against another, translating word by word, later phrase by phrase. Then, though never quite so as you can notice how, you arrive at a point where you don't translate at all, in the sense of formulating an idea in language A and translating it into language B. Instead you simply switch your mind over and think in the other language.
It's about this time that you generally decide certain phrases are untranslatable – and in themselves, they probably are, at least except by paraphrase. The Icelandic word gaefa, which crops up throughout the sagas, doesn't quite mean luck, though it is usually so translated, and it doesn't quite mean fate, because a man's gaefa is to some extent of his own making. A paraphrase might be something like "the state of affairs in a man's life when he is in harmony with his world". It is a concept, which belongs in the time of the sagas and makes immediate sense there. An ogaefumadr is a man whose relationship with his world has gone all wrong. His best intentions go amiss; people put the worst construction on his actions and what might loosely be called his bad luck will tend to affect those around him, too. In fact, if he didn't live in a pagan world where gods were not regarded as commanding unquestioning obedience, you might call him a Jonah – a concept which would, of course, have meant nothing to him.
The point is that though a word or phrase may be untranslatable, the idea behind it rarely is. One of the most famous examples of non-communication in history is preserved in a drawing done from life by the Inuit interpreter John Sacheuse in 1818. A British ship has just landed in Greenland and the British officers, Parry and Ross, are greeting the inhabitants, the first Europeans ever to do so. Parry has come prepared. He is holding out the age-old European symbol of peace, an olive branch, to the puzzled Inuit. The other officer, Ross, is just looking at a white, empty landscape where plainly no tree of any description has ever been seen, and you sense that any minute now, he is going to nudge Parry and suggest that they are, in more than one sense, barking up the wrong tree.
But it was not impossible for them to convey that they came in peace; they just needed to find the symbol for peace that belonged in that setting, rather than importing their own. Indeed, most of the trouble European explorers later encountered in the Arctic derived from their habit of importing and translating literally rather than adapting and thinking in a different way. To avoid scurvy they loaded themselves with tinned fruit and lemon juice in impractical glass bottles, because they knew that Europeans who ate enough fruit didn't get scurvy. It didn't seem to occur to most of them that the Inuit, who ate hardly any fruit except a few berries in autumn, didn't get scurvy either. It did occur to Ross, who when trapped for four years in the Arctic forced his men to eat seal blubber. He got them home safe and fairly well, but they complained bitterly, and for a long time nobody learned the obvious lesson, that to survive in the Arctic you needed to think Inuit, not wrap yourself in as much of Europe as you could carry with you.
Does this all sound a long way removed from the act of translation? I don't really think so, though it depends whether you define "translation" purely in terms of words. For me, it involves an ability to free oneself from one mindset, one culture, and open up to others. There is a sense in which the Inuit translation of "lemon juice " is "seal blubber". When missionaries translated the Bible into certain South Seas languages, the Lamb of God had to become the Piglet of God in order to convey anything to a sheepless audience. In fact missionaries and their like have always been adept at this chameleon act, and oddly enough it may have been their own moral certainty that made them so adaptable. They were conveying ideas, which they were sure were right for everyone, and to that end they were perfectly prepared to change the words, the symbols, the syntax and way of seeing things to suit local tastes. In this they were ideal translators, and I don't think it is any accident that one of the most instinctive and innovative translators of all time was Martin Luther, who was so determined to translate the Bible into the kind of German spoken by "the man in the market-place and the child in the alley".
His classic example of his methods was the Latin phrase "ex abundantia cordis os loquitur", which a previous translator had rendered "aus dem Ûberfluss des Herzens sprechet der Mund" - out of the overflow of the heart speaks the mouth. Luther, who was not at the best of times a tolerant man, could not pour enough scorn on this phrase. What, he asked, was the overflow of the heart? Was it anything like the overflow of the drains, and – always his acid test - what man in the market-place or child in the alley would ever say it? His own version, famously, was "wes das Herz voll ist, des gehet der Mund über" – when a man's heart is full, his words pour out.
This is classic translation of ideas rather than words or phrases, and as a translator Luther has always been a hero of mine because that is the way I have always wanted to translate poetry. Of course there is an extra complication with poetry, in that its meaning is conveyed on the level of sound as well as that of sense. It may sometimes be as important to convey into the new language the rhythms, rhymes, assonances and alliterations of poetry as the mere sense of the words. In fact, arguably the first thing any translator must decide is, given that something will inevitably be lost in translation, which element of the work in question is the most essential, the most vital not to lose.
If you look at most English translations of Aristophanes before the 1960s, you will see, over and over, some play on words or topical reference rendered with laborious precision into English, with a footnote to explain why, when said on an Ancient Greek stage, it was hilariously funny. Then, in the sixties, Mentor Classics brought out a series of Aristophanic translations by American academics – William Arrowsmith, Douglass Parker, Richmond Lattimore, Dudley Fitts - which took as their starting point the assumption that the most important thing about translating Aristophanes was to make him funny. When translating words which had had their original audience in stitches, it was not enough to explain why the thing had once been funny. One must make it so, and if that meant departing radically from the text in order to recreate the joke in another language and culture, so be it. This was Aristophanes you could laugh out loud at while reading.
And it wasn't that hard to do, because the well-springs of humour do not alter much, and the likes of Aristophanes and Plautus have recognisable modern equivalents. Aristophanes, for instance, is forever making jokes about piglets, because the Greek word for piglet, choiros, also has the slang meaning "female genitalia". One can translate it literally as "piglet", and explain in the footnote why Greeks found it funny. Or one can translate it as "pussy", and recreate the exact effect of the original, for Mrs Slocombe's pussy from Are You Being Served is indeed the lineal descendant of Aristophanes' piglet. The moral is plain: if you want to translate the man in the way he himself would most have wished, i.e. amusingly, you need to think yourself into some such equivalent mode as the seventies British sitcom.
Having said that, I do think it is possible to over-translate, to suppose that no idea will travel unless re-clothed in the dress of the new culture. One sees it frequently in productions of Shakespeare or Greek tragedy, where the first thing the director seems to have asked himself is "which period shall we transpose this into?" Of course this practice can shed interesting new light on a play, but it should not be automatic. Personally I find it vaguely insulting for the director to assume that I will not be able to react to a play about tyranny unless it is set in the Third Reich. It can also be insulting to the original author, for if he was good enough at his job, his words and ideas will reach out a long way beyond his own time and place. When, at the start of Women of Troy, Euripides has Andromache and her child brought onstage on a cart piled high with loot, two more items of booty from the sacked city, no audience with any imagination needs "Troy" translated into "Baghdad", or anywhere else, to get the point. Indeed Euripides' original Athenian audience did not. The year that play was first produced, during the Peloponnesian War, Athens had, with very little justification, invaded the island of Melos, killed all the men and enslaved the women and children – Greeks, like themselves. Many Athenians had felt very dubious about the ethics of this; when Andromache wondered aloud who would own her, the audience knew fine well they were neither in Troy nor the heroic past, without Euripides needing to spell it out.
Language, though, needs translating even when images, times and places do not, and fortunately for the employment prospects of translators, most translations have a shelf life because spoken idiom changes, and not just spoken idiom either. I have always been fascinated by two English translations of the same Irish Gaelic poem, one by the 19th-century poet Padraig Colum and a second by the 20th-century writer Frank O'Connor. Colum's first verse went:
O woman shapely as the swan,
On your account I shall not die!
The men you've slain – a trivial clan –
Were less than I.
I shall not die because of you,
Oh woman, though you shame the swan.
They were foolish men you killed.
Do not think me a foolish man.
There are differences of poetic language here – the inversion of Colum's second line, the exclamation mark and the high style of "slain" have no place in O'Connor's plainer mid-twentieth-century poetic speech. But there also seems to be a difference in attitude. I have no idea which is closer to the spirit of the original, but Colum's speaker and O'Connor's are clearly two different men. Colum's, with his 19th-century certainty and bravado, sounds almost bombastic beside O'Connor's quieter, more wry voice. In this case the 19th-century dress could actually risk obscuring the original poem for a new audience and poems are constantly needing new suits in this way
Translators of poetry, like Parry and Ross in the Arctic, need to look less at the words than at the concepts they signify and the associations they carry with them. They also, like the American translators of Aristophanes, have to decide what in their chosen text is most of the essence and what can be sacrificed if it must be. If I might use a personal example for a moment, I was once translating a poem by the 17th-century German poet Paul Fleming, a love-song called Für eine Jungfrau, written in the person of a lady, for her to send to her lover. It was an endearing, vaguely folksy little piece, with a music produced partly by extensive use of feminine rhymes. This was my translation of its last verse:
May God lay on my lover
as many marks of favour
as stars in heaven may be,
as twigs are in the hedgerows,
as grasses in the meadows,
as fish swim in the sea.
Most of this is straight, literal translation; the exception is the word "hedgerows". In the German the line reads "as twigs are in the forests" (Wälder, which rhymes with Felder, the "meadows" of the next line). This was a folk poem; without rhyme of some kind it was nothing, and forests and meadows cannot easily be got to rhyme or half-rhyme; none of their near-synonyms really match. "Woods" and "fields" is getting close, but it is a masculine near-rhyme and utterly ruins the music of feminine rhymes which is a big part of the poem's raison d'être. I wasn't losing that for anyone. It seemed to me, looking at concepts rather than words, that apart from its sound qualities, the poem's other essential was the folk motif of semi-humorous exaggeration, the strength of the woman's feelings showing in her series of accumulations. "Wälder", for the purposes of this poem, meant "some place in which there might credibly be a great many twigs" and frankly hedgerows would translate that every bit as well as forests.
Of course this judgement as to what's most important is to some extent subjective, which is why two translators concentrating on different bits of the elephant can get such widely diverging results. In a recent Guardian article the journalist David Aaronovitch, writing about a woman who struck him as anti-semitic, three times referred to her as a blonde woman. A couple of women readers wrote back indignantly about the stereotyping of blondes; one asserted that had this woman been a man, Aaronovitch wouldn't have felt it necessary to mention his hair colour. Evidently these readers were reading "blonde woman" as journalistic code for "airhead", which of course it often is. But it surprised me, because I had read it differently. I think Aaronovitch would in fact have said "blond man", had the occasion arisen, because I read his use of "blonde" as code for "Aryan, and therefore probably anti-semitic" – still stereotyping, but of a different kind. I can well see how someone translating this article into another language could, consciously or unconsciously, slant it either way. And it would work in reverse: eine Blondine could be literally translated as "a fair-haired woman" or "a blonde". Both are accurate translations, but to an English-speaking ear they definitely do not give the same result. There is no such thing as a Page Three fair-haired woman.
(While we are on the subject of German blondes, by the way, it is as well to remember that while eine Blondine is a blonde, eine Blonde is a glass of beer. Ordering the wrong one may lead to trouble.)
One thing we can say: the journalist's use of "blonde" in that case cannot have been completely neutral, a mere physical description. Since the woman's hair colour had no relevance to what she was saying, he must have been reacting to some association it had for him. And this is the sort of thing a translator, especially perhaps a literary translator, is liable to pick up on and want to reproduce in the new context.
But this, of course, is where translation can easily slide into interpretation. In your zeal to convey as much as possible of your original author's intentions, you can find yourself tempted to explain and spell out what he deliberately did not. Eventually any literary translator comes up against the problem of how to translate what is not actually said in the original but is nevertheless conveyed to its readers. One instance of this is shared information specific to the original audience. It might be a reference to some historical or fictional character, or a buried quotation. It is not spelt out because it doesn't have to be; its impact is almost subliminal and in fact if it were spelt out the impact would be ruined in the same way that Aristophanes' jokes are if you have to explain them.
The ideal solution for the translator in this situation is to find some local equivalent for the new audience, which will similarly make its impact without being spelt out. Otherwise he'll have to resort to footnotes, and though the information can be translated that way, the impact cannot. On the other hand, few local equivalents are exact, and some are positively misleading. The Russian equivalent of Jack Frost is Moroz Krasni-Noz, Frost the Red-Nosed, and despite the cheerful-sounding name he is a fearsome character who catches travellers out in the forest after dark and freezes them to death. He would certainly put a new slant on any Russian translation of English children's rhymes, but it would hardly be what the author intended. George Szirtes has remarked on the difficulty, when translating a Hungarian poem into English, of conveying what a mention of the sea really means to someone from a land-locked country. The word can be translated easily enough, but a British reader, to whom the coast is always in relatively easy reach, simply will not read it the same way its first audience did.
I'm not sure it is possible for a translator alone to overcome a problem like that; the reader too needs to reach out, to think himself into another skin. What the translator may be able to do is help him realise this need exists. It wouldn't, for instance, necessarily occur to a British reader of a Hungarian work in translation that Hungarian attitudes to the sea might differ from his own. Earlier I suggested that, in the interest of translating concepts rather than words, a translator would sometimes be looking for equivalent images and experiences, stressing people's basic similarities rather than their differences. As Confucius put it, men's natures are alike; it is their habits that drive them far apart. The difference between Aristophanes' piglet and Mrs Slocombe's pussy is a habit; it is the underlying laugh that is nature. But sometimes difference is a part of nature too. Shelley translated a Platonic epigram as
Kissing Helena, together
With my kiss, my soul beside it
Came to my lips, and there I kept it,
For the poor thing had wandered thither,
To follow where the kiss should guide it,
Oh, cruel I, to intercept it!
Not his best effort, to be honest, but what is more interesting is the deliberate mistranslation in the first line, where Plato did not write "Helena" but "Agathon", which of course is a male name. Shelley's reasons had to do with theories of translation rather than censorship. As is plain from the foreword he wrote to his translation of Plato's Symposium, it bothered him mightily that his heroes Plato and Socrates should have had what to him seemed aberrant sexual tastes. Since he venerated them in all other ways, he could only assume that something about ancient Athens – the status and education of women, possibly – was vastly different from our own world and that had Plato lived in our day, he would have thought and behaved differently. And I'm sure he translated male into female because it seemed to him that if he did not, his readers would never see past this one great difference to the Plato he himself could see. He was, I think, trying to translate ideas rather than words, reasoning that the constant, the essence, was love rather than the particular form it took, and that what was a normal expression of love in Plato's day needed therefore to be translated into something that was normal for early nineteenth-century England.
Although Shelley could well have been right about his audience's reaction had he done otherwise, I still think he made the wrong decision, myself. If nothing else, I can't imagine Plato approving of it, which is a consideration that should weigh heavily with any translator. The argument that love is the essence – the nature - and the outward form it takes merely the habit is appealing in some ways, and Edwin Morgan, for one, used to keep personal pronouns out of his love poems, even after there was no legal prohibition to worry about, for much this reason, that he did not want to alienate straight readers. But in the end, sexuality is a bit more than a habit; it is an intrinsic part of a person's nature, and to falsify that in translation does seem essentially different from changing piglets to pussies. This, I think, is one occasion when the translator must ask the reader to do some reaching out and think himself into another culture. The translator must let the text say, in essence, this is Plato: you can see for yourself the breadth of his thought, his quest for knowledge and truth, his poetic vision. You can also see that this was his sexual nature, and if you can't reconcile these two sides of him, then you will have to change the way you are looking at them.
I said earlier that once you are thinking in another language, some things seem impossible to translate except by paraphrase. I think the bane of every poetry translator's life is the really well-written epigram. Brevity is the soul of the epigram; conciseness of expression is itself the essence of the poem, the element that must be translated above anything else, and it is, more often than not, exactly what you can't hang on to. All the phrases that give me the most trouble to translate are short. There is a saying of Goethe's, which has become a proverbial expression of contempt for mediocrity in Germany: so was lebt, und Schiller mußte sterben. Literally translated, this means "the likes of him is alive, and Schiller had to die". This is hard enough to translate anyway, because it depends for its impact on the kind of shared information I mentioned earlier, which doesn't cross cultures. Schiller, with his immense literary genius, died before his time. Turning him into Keats doesn't really work, because though most Germans still know Schiller died young, (partly because of this saying) most modern English readers don't have a clue what age Keats died at. Perhaps you could turn him into Diana. But you'd still have the problem of so was, literally "something like that", which, when applied to a human being, exudes a degree of contempt quite remarkable for two short words, and for which "the likes of him" is a wholly inadequate translation. It really conveys something more like "that object you wouldn't waste your time kicking into a ditch", but there you are, you see: the brevity's gone. The concept I found it hardest to translate in a German poem, ever, was the difference between Amerika spelt with a K, which is how German normally spells it, and America with a C. Spelling it with a C in German amounts to cultural cringe, but how to convey that with anything like the economy of the original stumped me then and still does.
Which doesn't, of course, mean that it can't be done. I am an optimist about translation, because basically I think Confucius was right about men's natures being alike. No two readers, even from the same culture, read a piece exactly the same way, and in these post- Barthes days authors accept that, though apparently the BBC does not. Lately in Hay-on-Wye, a BBC radio interviewer asked me "How do poets make sure all their readers read the poems the right way, the way the poet intended?" and he seemed most taken aback when I suggested that a reader's alternative reading might be valid too. A writer cannot be sure of conveying exactly what was in his mind; the moment his words enter other ears they will be filtered through a network of memories and associations he can't even guess at. The translator, in the middle of a game of Chinese whispers, has the problems both of hearer and conveyor of language.
But the fact that a thing can't be done perfectly does not mean it is pointless to do it at all. The reason I began translating poems in the first place was that, over and over again, I found myself moved by similarity, closeness, common humanity. The Ancient Greek poet Ibykos, falling in love again in his eighties, compares himself to an old racehorse forced to drag itself round the track one more time. A fifteenth-century English schoolboy, needing an excuse for being late to school, panics and comes up with "I had to milk the ducks for my mother". Seneca, in Nero's Rome, asks rhetorically and irritably how a man is supposed to write philosophy with a procession of sausage-sellers crying their wares below his window. In nineteenth-century New York, Walt Whitman wakes in terror from a dream of being naked in the street. A ninth-century Irish scholar in an Austrian monastery describes his one relationship, his companionship with his cat Pangur. And they could all be next door; all are recognisably of the same stuff, and often up the same creek, as us. In the absence of the Babel fish that would allow us to speak all languages fluently and experience all literature in the original, there is certainly something of all these experiences that can be translated for new readers, that in fact, given a little help with the words, translates itself.