Introduction

Many people, not all of them outside Italy, think that the Divine Comedy is a rather misshapen story. And indeed, if it were just a story, it would be back to front: the narrator has an exciting time in Hell, but Purgatory, when it is not about art, is about theology, and Heaven is about nothing else. What kind of story has all the action in the first third, and then settles back to stage a discussion of obscure spiritual matters? But the Divine Comedy isn’t just a story, it’s a poem: one of the biggest, most varied and most accomplished poems in all the world. Appreciated on the level of its verse, the thing never stops getting steadily more beautiful as it goes on. T. S. Eliot said that the last cantos of Heaven were as great as poetry can ever get. The translator’s task is to compose something to suggest that such a judgement might be right.

This translation of the Divine Comedy is here today because my wife, when we were together in Florence in the mid-1960s, a few years before we were married, taught me that the great secret of Dante’s masterpiece lay in the handling of the verse, which always moved forward even in the most intensely compressed of episodes. She proved this by answering my appeal to have the famous Paolo and Francesca episode in Inferno 5 explained to me from the original text. From various translators including Byron we can see what that passage says. But how did Dante say it? My wife said that the terza rima was only the outward sign of how the thing carried itself along, and that if you dug down into Dante’s expressiveness at the level of phonetic construction you would find an infinitely variable rhythmic pulse adaptable to anything he wanted to convey.

One of the first moments she picked out of the text to show me what the master versifier could do was when Francesca tells Dante what drove her and Paolo over the brink and into the pit of sin. In English it would go something like:

We read that day for delight
About Lancelot, how love bound him.

She read it in Italian:

Noi leggevam quel giorno per diletto
Di Lancelotto, come l’amor lo strinse.

After the sound “-letto” ends the first line, the placing of “-lotto” at the start of the second line gives it the power of a rhyme, only more so. How does that happen? You have to look within. The Italian eleven-syllable line feels a bit like our standard English iambic pentameter and therefore tends to mislead you into thinking that the terzina, the recurring unit of three lines, has a rocking regularity. But Dante isn’t thinking of regularity in the first instance any more than he is thinking of rhyme, which is too easy in Italian to be thought a technical challenge: in fact for an Italian poet it’s not rhyming that’s hard.

Dante’s overt rhyme scheme is only the initial framework by which the verse structure moves forward. Within the terzina, there is all this other intense interaction going on. (Dante is the greatest exemplar in literary history of the principle advanced by Vernon Watkins, and much approved of by Philip Larkin, that good poetry doesn’t just rhyme at the end of the lines, it rhymes all along the line.) Especially in modern times, translators into English have tended to think that if this interior intensity can be duplicated, the grand structure of the terzina, or some equivalent rhymed framework, can be left out. And so it can, often with impressive results, each passage transmuted into very compressed English prose. But that approach can never transmit the full intensity of the Divine Comedy, which is notable for its overall onward drive as much as for its local density of language.

Dante is not only tunnelling in the depths of meaning, he is working much closer to the surface texture: working within it. Even in the most solemn passage there might occur a touch of delight in sound that comes close to being wordplay. Still with Paolo and Francesca: in the way the word “diletto,” after the line turning, modulates into di Lancelotto, the shift from -letto to -lotto is a modulation across the vowel spectrum, and Dante has a thousand tricks like that to keep things moving. The rhymes that clinch the terzina are a very supplementary music compared to the music going on within the terzina’s span.

The lines, I found, were alive within themselves. Francesca described how, while they were carried away with what they read, Paolo kissed her mouth. Questi (this one right here), she says, la bocca mi basciò, tutto tremante (kissed my mouth, all trembling). At that stage I had about a hundred words of Italian and needed to be told that the accent on the final “o” of basciò was a stress accent and needed to be hit hard, slowing the line so that it could start again and complete itself in the alliterative explosion of tutto tremante. An hour of this tutorial and I could already see that Dante was paying attention to his rhythms right down to the structure of the phrase and even of the word.

The linked rhymes of the terza rima were a gesture towards form, marking the pulse of the onward surge of the great story, which was driven by its poetry; and would be infinitely less great without its poetry, just as Wagner’s Ring cycle would be infinitely less great without its music. But Dante’s formal requirements for himself went down to the very basics of the handling of language. It was all very precise, and yet it all added up. Though it was assembled from minutely wrought effects, the episode really did have rhythmic sweep. My wife, clearly touched by my sudden impersonation of a proper student, said all the rest of the poem was like that too, including the supposedly colourless theological bits. Every moment danced, and the dance was always moving forward.

Over the next year or so, while I was reading Dante in the original to satisfy the requirements of the English tripos at Cambridge, I looked at several rhymed translations and found them strained. On the other hand the translations done in prose had whole chunks that were too dull to read, especially in the second and third books. The total effect of looking at so many translations was to be convinced that the job was thankless. One thing I could see clearly, however, was that any even halfway successful translation would have to rhyme, although the question was how, especially in those long stretches, later on in the poem, where not a lot seemed to be happening.

You could see easily why Byron went no further than his Paolo and Francesca; he couldn’t keep up the excitement if there was nothing for the excitement to bite on, as it were. Starting from the level of incident, he, or any other translator, couldn’t get down to the level of language. They could raise themselves to the level of thought: some of the translators were, mutatis mutandis, as educated as Dante was, who was one of the most educated men of his time even in the conventional sense, quite apart from the proto-scientific sense in which he was original without parallel. But they couldn’t, or wouldn’t, get down to the level where syllables met each other and generated force. That had to be the aim, impossible as it seemed; to generate the force, both semantic and phonetic: the force of both meaning and sound. Indeed, in the original, some of the meaning was in the sound. Unless a translator did something to duplicate how the poem sounded, he, or she, wouldn’t get near what it meant.

So the task, if task it was, went on the back burner and stayed there for about forty years. I was barely aware that I was even thinking about it. Helping me to be diffident on the topic were my memories of what my wife had said about the Dorothy Sayers translation. Back at Sydney University, when we first met in the late 1950s, she was studying Italian: the beginning of what would turn out to be a distinguished lifelong career of teaching and editing Italian literature, with particular attention to Dante. As part of the course, she had been required to produce a substantial paper; and she had made her subject the Dorothy Sayers translation, which was famous at the time and indeed is still well-known now: the total amount of money it has made for Penguin must be colossal. But my wife, always the closest of readers, was able even then to detect that Sayers had simultaneously loaded her text with cliché and pumped it full of wind.

Finally, not long after I retired, I thought I could see how a translation might work. I started by drawing the necessary conclusions from my knowledge of what to avoid. The first thing I had learned was that a strict terza rima was out of the question. Even Louis MacNeice, one of the great verse technicians of modern times, had resorted to half-rhyme in order to sustain a long poem in terza rima. And the sad truth about Autumn Sequel is that it is simply terrible. Its predecessor, Autumn Journal, had been a triumph—it should be ranked, in my view, as the great long poem of its time—but for that work he had allowed himself access to his vast stock of classical metres. In Autumn Sequel he stuck himself with the terza rima, and with a not very attractive version of it. I tried writing a longish poem in strict terza rima and could see that it was creaking with strain the longer it went on, thereby accomplishing the opposite of the desirable effect of a narrative form, which is to get you into the swing of the thing. With the dubious exception of Shelley’s The Triumph of Life, nobody has ever written a terza rima poem in English that makes you forget the form in which it is composed, and a terza rima translation of Dante like Laurence Binyon’s makes a feature of Binyon’s virtuosity rather than Dante’s mastery.

So if it couldn’t be in strict terza rima, and if a loosely rhymed terza rima wasn’t worth writing, what regular measure would carry the freight? Over the years I had written thousands of couplets and although I enjoyed using them for comic effect I knew that they wouldn’t fit this task. The Divine Comedy isn’t comic. A few couplets might come in handy to clinch each canto but on the whole the couplet suffers the drawback that Johnson spotted in the work of Pope: in a rhyme-starved language like English, the same rhyme sound keeps cropping up too early. Even if the words that rhyme are kept deliberately different each time—night/bright, light/sight, etc.—the sound is the same, and calls the wrong kind of attention to itself. And yet I wanted the rhyming words to be close enough together to be noticed.

Finally I realised that I had been practising for this job every time I wrote a quatrain. In my poems, ever since I started writing them in the late 1950s, the quatrain was the most common measure unless I was writing in free verse. Decades of practice had made the quatrain so natural to me that it had become a mode of thought. For this project, if the quatrain could be augmented with extra lines whenever the occasion demanded or opportunity offered, it would yield the ideal combination of strictness and ease. To reinforce the strictness, I would avoid feminine rhymes. When writing couplets, one veers inexorably towards feminine rhymes, and the effect, though often usefully flippant, is always in danger of recalling Gilbert and Sullivan. To match Dante’s gravitas, a strong, solid dignity would be required. Well, I had had plenty of practice in writing quatrains with masculine rhymes. So really I had spent all this time—the greater part of a lifetime—preparing my instruments.

Quietly I got the project under way, starting with the first line of the first canto. It would be a mistake to do the famous passages and then join them up: the welds might show. If my modified quatrain measure was to justify itself, it would have to work from the beginning, and be still working at the end—because the Divine Comedy ends as it begins, with the same feeling of exuberance that has been maintained throughout. Dante registers this exuberance even in his most desperate moments. He doesn’t stop singing just because something dreadful is happening. What he says is: something dreadful is happening even as I sing. It’s an interplay of form and content: the most ambitious that any literary artist ever attempted. Dante’s torrential cascade of poetic effects adds up to a claim. No poem could be equal to my subject: no poem except this poem.

I wanted my translation to provide parallels for such effects while maintaining all of Dante’s sense of economy. No poet, not even Shakespeare, could say quite so much so quickly, so the translator must know how to be brief. But there are things he can add without slowing the pace. (Or, indeed, taking more space: since English needs less room than Italian, there is latitude for adding things while keeping to the same length as the original.) Dante had barely finished the poem before the first commentaries began appearing. Commentary was thought necessary because Dante had composed every canto of his poem as if it were a weekend article based on news that had only just happened, and whose details did not need to be outlined. There must have been readers who, not having heard the news about who robbed whom and which pope double-crossed which prince, were puzzled even at the time.

Since then, the chances of being puzzled have multiplied, and there is also the increasingly pressing matter of making even references to the Gospels clear to readers who might not be familiar with them. The first commentaries inaugurated seven hundred years of scholarship and criticism which have gone on to this day, and the summary of all that knowledge and informed speculation is there at the foot of the page in any scholarly edition. But this translation would be for non-scholars in the first instance. Footnotes would be a burden to it. Ideally, the thing itself should carry all the information it would need. How to do this?

The reservoir of material at the foot of the page of a scholarly text (sometimes there are five lines of text and fifty of apparatus) provides the translator of the poem with an ideal opportunity to upload salient facts into the verse narrative and make things clear. There are stretches in all three books when it really helps to be told who belongs to which family; whether one family is at war with another; and what precisely happened to which city when it was betrayed. On the vexed question of theology—this crops up especially in the third book, Heaven—it can help to be told that a certain scholar represents a certain position. Almost always the relevant information is there on the page, down at the bottom. In numerous instances I felt justified in lifting it out of the basement and putting it on display in the text. It might seem bold to assume that Dante, if he thought the reader might not know, would have explained which features of a certain scholarly dispute he was referring to, but we can also assume that he didn’t want the reader to be presented with an insoluble puzzle. Dante wanted to be read. Every stratagem he employed tells us that. He was talking about the whole of creation at once but he wanted to glue the reader to the bench.

To help him do it for the present age, I opened up a way to make simple, sparing but sufficiently wide-ranging use of what we might call the basic scholarly heritage. My aim, when importing an explanatory detail, was to make the text more readable instead of less. Also I have cut back on his lists of names here and there, when a list is all it is. Perhaps boldly, I would say that all the reader needs to know is in the poem as I have presented it. As Dante, in Heaven, travels all over the sky, some of the references to the Zodiac might seem obscure, but they probably seemed so at the time: and really that’s all you have to know, that the references sounded learned. Dante’s first readers could take the more obscure points for granted while they followed the main points. My job was to make some of the main points more self-clarifying by putting in some of the explanations that had been accumulating at the bottom of the page for the best part of a millennium, but to do so without slowing the tempo. The result one hopes for is a readily appreciable outline of Dante’s Christian vision. Is it possible, though, in an age without belief?

Well, how much does Dante believe? The truth is that he didn’t want to believe anything if he couldn’t test it. As we trace the story through one hundred cantos, three regions and every shade of emotion from despair to bliss, we find that he believed in his journey to salvation. But if his belief had been without its doubts, there would never have been a journey; and at the very end we find ourselves concluding that the great poet, setting out the reasons for his faith, has reached conclusions that will eventually make blind faith impossible. Dante could ask questions about theology because he was in love with questioning itself; he was in love with the design of the divinely inspired universe because he was in love with design, which he could see in a fold of cloth and a fallen leaf.

Dante had a scientific mind: one of the first scientific minds we know of in the modern age, for which, indeed, he can be said to have built the foundations. His poem stands at the dawn of modern science, and therefore of the times we live in now: its essential moment is in its final vision, when Dante the traveller, at the apex of heaven, looks into the source of creation and sees the imprint of a human face. The Divine Comedy is the precursor of the whole of modern history, and I hope this translation conveys enough of its model to show that he forecast the whole story in a single song: a song of lights. The joy of discovery is what drives the poem, and if my translation gets some of that exultation into English verse then it will have done its work.

For all his majestic weight, there is also a lightness to Dante, and I hope to have got that in. Finally but essentially, tempo is one of the two main poetic elements in play. The other is texture. What we have, in this miraculous work of art, is a mutually reinforcing balance of tempo and texture, of a kind that had never been heard before over such a distance: fourteen thousand lines of it. Much as he worshipped Virgil, Dante was better at it than his master. If he could have read Homer, he would have found that he was better at it than Homer too. In the original Italian, you can hear it. But most of us will never read Dante in the original Italian. It’s a pious wish that translators are always making: they hope that the reader, intrigued by the translation, will be driven to learn Italian, etc. Common sense tells us that it will seldom happen.

To know a foreign language thoroughly is a big task, and to know its literature is a full time job. For more than fifty years, my wife’s scholarship, her tenacity and seriousness of purpose, have been there to remind me of what it means to be dedicated to Dante and to help pass on the body of knowledge associated with his name. Her work culminated first with her gold-medal-winning edition of Dante’s Monarchia for the Società Dantesca Italiana, the only national edition of any work of Dante edited by a non-Italian, and a labour of love that took her thirty years. It culminated all over again with the completion of her digital edition of the Divine Comedy’s manuscript tradition, a tool for all Dante scholars, and a thing of extraordinary beauty and utility. I hope she will forgive me for straying onto her territory, but really there is no contest. Beside her lucid and scrupulous scholarship, a translation counts for very little. But I have done my best with it, always encouraged by the memory of how, in Florence, she first gave me an idea of what it meant to be in the service of her great poet.

—LONDON, 2012