Translation of Literature: Challenge, Sacrifice and Potential

This extended essay by Fay Davies was awarded the prestigious Ithaka Prize on Tuesday, 19th March, 2013. It was originally published in Portsmouth Point magazine.



Winner of the 2013 Ithaka Prize
(photo: Chris Reed)

One might think of language as a system of signs understood by humans to refer to certain concepts in the mind. Its purpose is communication – rendering thoughts into spoken or written words so that they may be received by other people. This is only one definition of many, but however it may be viewed, one aspect of language is clear: it is far from precise. No word has an absolute, indisputable meaning, and every individual's interpretation is unique.

There is more than one language in the world: more than one system of arbitrary signs which refer to our perception of reality. These can be learnt and understood, and of course translated. But the process of translation is not the same as the act of mere understanding. There is never a corresponding word in one language for every word in another, and languages have different structures: grammars. The process of translation inevitably becomes an act of rewriting, as translator must read and interpret a phrase, then render the thought in another language. Adding another stage into the imprecise process of communication, the translator creates a new piece, and it will always be a distortion of the original. If the subject of this translation is literature, it cannot hope to replace the author's work, but only perhaps provide us an insight into the original.

Of the variety of literary forms, poetry is perhaps the most difficult to translate. To the existing problems of translation, poetry adds the constraints of rhythm, rhyme, tone and form. In his introduction to 'Ezra Pound, Translations', Hugh Kenner writes 'If he doesn't translate the words, the translator remains faithful to the original poet's sequences of images, to his rhythms, or the effect produced by his rhythms and to his tone'[1]. The very mention of some kind of choice suggests a difficulty. The translator, it seems, can choose not to translate 'the words'. He must 'remain faithful' to either the images, or the rhymes, or the tone. He must sacrifice certain elements in favour of what he thinks is most important and most representative of the original work, and this is subject to individual interpretation.

To help illustrate the subtle, elusive and sometimes ineffable difficulties that translation of poetry presents, I will be comparing two translations of Dante Alighieri's 'Inferno', Book One of The Divine Comedy: that of Dorothy Sayers, 1949, and Steve Ellis, 1994. The Divine Comedy has been translated into English by at least thirteen notable translators since 1805, and each translation is wildly different – demonstrating the unavoidable imprecision of the art.

I will also be exploring evidence of their different intentions, facilitated by the introductions in which both translators state their intended style. Ellis, for example, starts by asserting that his translation is 'a colloquial version'[1], and that 'it tries to recapture some of the vigour and directness of Dante's original'. Importantly, 'vigour' and 'directness' are abstract qualities, and there is no absolute means of reproducing them. The kind of language that incited vigour in Dante's era and in Dante's culture is certainly not the same as today. Sayers, the second translator, claims that 'the vocabulary and the sentence-rhythms of verse are not, and never can be, exactly the same as those of contemporary prose'[1]. Thus, although she writes that she has 'considered the whole range of intelligible English speech', she is of the opinion that her diction must be altered to match the demands of verse. A noticeable difference between the two versions is Sayers's use of the ancient 'thee and thou', marking her work with the echoes of ancient speech.

Finally, I will address the theoretical concerns at the heart of translation. The sacrificial notion, evident in Kenner's words, is a common one: we see translations in terms of what has been lost, what cannot be brought over to the new language. However, perhaps it may be possible to change such a perception: to see the art in a more positive light.

Starting the analysis is an example that demonstrates the idea of sacrifice, as Kenner notes. In Canto XVIII, line 51, Dante uses a pun. 'ma che ti mena a si pungenti salse?' He is inquiring as to the reasons that his companion found himself in hell, and a fairly literal translation would be 'But what brings you into such a biting pickle?' Salse, the word for pickle or sauce, is also the name of a ravine near Bologna where they threw the bodies of criminals, and herein lies the pun. This line is problematic for two reasons: historical and linguistic. The first is that modern readers are unlikely to know of such a ravine: regardless of the need for translation, the detail may not be understood. This is something that may be a problem with any text, as they all tend to presuppose a certain amount of knowledge on the reader's part. Yet, the older the text and the further it is from the culture and society of the reader, the more likely it is to contain allusions that are not understood.

The second reason that this line poses difficulties is more crucial to the nature of translation: the English translation of 'salse', whatever it may be, is not 'Salse'. The pun no longer exists. Sayers and Ellis overcome this problem in two different ways. Sayers writes, as a translation of this line, 'What wormwood pickled such a rod?' Her translation retains the pun element. She could never use the same pun as Dante, but she creates one that her readers are more likely to understand. In a way she ignores Dante's words; she does not give a literal translation but seeks to create a similar device of her own. While 'pickled' may have been inspired by 'salse', it occupies a completely different role in this sentence as has become a verb where it was originally a noun. No matter, her translation does have roughly the same meaning: the narrator asks this occupant of Hell how he came to be there. And, as she preserves the use of a pun, she in turn preserves the humour and playfulness of Dante's line.

Ellis, on the other hand, does not preserve the pun, writing 'Why is your sauce here so spicey?' Ellis has provided a reasonably literal translation of Dante's line, allowing the pun to be sacrificed and not providing another. The relative simplicity of this line is typical of Ellis's colloquial style. Of course, it is not without its own merits: the playful alliteration of 'sauce' and 'spicey' has, like Dante's original line and Sayers's pun, a comical effect. Significantly, while one translator brings over the pun to their version and one doesn't, both bring over the humour of the line. Perhaps this is most important: for the translators to attempt to reproduce the same feeling in the reader that they think a reader of the original would have felt. Incidentally, what they believe a reader may have felt is most likely what they themselves felt upon reading. Thus, subjective interpretation plays a large part in the process of translation.

Humour is something that both translators are careful to preserve, but it is a complex topic. Sayers notes this, writing in her introduction that '[humour], of all qualities, has been the most hopelessly obscured by his translators and critics'. Humour is problematic because it is produced by a network of words, cultural expectations and the subversion of these. Often, the causes behind it are elusive and a result of multiple strands. Sayers sums this up, describing Dante's particular humour as 'something more like a faintly ironic inflection in the voice than anything humorous in the words themselves'.  An example of this 'ironic inflection' and the way it has been preserved by the translators is Canto XI, lines 76-78, where Dante's guide Virgil reprimands him for the stupidity of his questions. Sayers translates it as:

            “What error has seduced thy reason, pray?”
            Said he, “thou art not wont to be so dull;
            Or are thy wits woolgathering miles away?”

The humour lies in the fact that the narrator represents Dante himself, therefore passages such as this are ironic and self-deprecating. In her translation, Sayers increases the sense of insult with her use of metaphor and words that are particularly offensive. What translates fairly literally as 'it looks elsewhere' becomes 'woolgathering miles away'. 'Why does the mind wander so' is replaced with the accusation 'so dull'.  Sayers's Virgil is, in this instance, more emphatically insulting and disrespectful than Dante's. As a result, she amplifies the sense of self-deprecation and, arguably, the humour of the section.

Ellis, however, does not supplement Dante's words with metaphor:

            He says, 'Why do your wits wander
            out of their normal way so much?
            or what's your mind got hold of?'

Arguably, to say 'what's your mind got hold of' excuses the matter as absent-mindedness or anxiety; there is nothing of the ridicule of 'dull' or 'woolgathering'. The humour is not completely lost, as it is inherent in the very notion of Dante putting himself down; yet it would seem less explicitly comical than Sayers's version. Despite her observations that humour tends not to be a result of the 'words themselves', it appears that Sayers intends to accentuate the humour by these means; particularly by the use of 'woolgathering'.  Interestingly, Ellis's unembellished version is closer to the kind of subtle irony that she describes.

Occasionally, word order alone can mean that different shades of humour are produced. In Canto XVII, lines 91-93 there is another instance of self-deprecation, this time more subtle: as he boards the shoulders of a demon, the Dante acknowledges his fear and futility. As Sayers translates:

            So I climbed to those dread shoulders obediently;
            “Only do” (I meant to say, but my voice somehow
            Wouldn't come out right) “please catch hold of me.”

In these lines, there is a revelation: that Dante was unable to say what he intended. Sayers conveys this revelation in parentheses: the syntax is broken, stuttering, anxious. In this translation, the character Dante makes no attempt to hide disorder: it is immediate, honest, and not at all dignified. The confused uncertainty of 'somehow' adds to the pitiful humour: he has no control and no answers. The phrase in parenthesis uses more colloquial diction than is normal in this translation: the abbreviation 'wouldn't' and the simplistic 'come out right' seem curiously modern. Sayers in fact writes in her introduction that she favours a modern phrase 'if the passage was humorous or conversational', revealing her intention to produce changes of tone when she deems it appropriate.

Ellis phrases these lines differently:

            I perched on those vast shoulders:
            I wanted to say, 'Hold me tight,'
            but found my voice wouldn't come.

The revelation comes last, and the irony comes from writing what he intended to say and later revealing his inability to say it. Crucially, Ellis's syntax is less cluttered, and unbroken by parenthesis. This gives his version a greater sense of control than that of Sayers. Yet, he still portrays disarray with the verb 'found', as that he 'found' his voice wouldn't come is similar to Sayers's use of 'somehow'. It implies a confusion; a lack of control over his faculties. 

The way Dante writes this section, the syntax is also unbroken but the speech comes last:

            I' m'assettai in su quelle spallacce:
            si volli dir, ma la voce non venne
            com' io credetti: 'Fa che tu m'abbracce.'

The revelation of his failure to speak is the first thing Dante notes. Arguably it comes as less of a surprise than in Ellis's or Sayers's versions, as no intended speech has been recorded before we learn of his failure to say it. The phrase 'ma la voce non venne / com' io credetti' ('but did not come as I thought') is parenthesised: writing it as the less dominant clause of the sentence Dante unveils a  shame over his fright.

The mere position of the revelation, whether it is first, last or central, subtly affects the humour that is created. It is impossible to judge exactly why each version is comic, and which is more comical than the others, because it depends on individual interpretation. It is also difficult to know exactly why Ellis and Sayers chose the order that they did: perhaps it was to convey the exact feeling they interpreted from Dante; perhaps it was because it best accommodated their chosen rhythm. In any case, the translators have not produced a faithful representation of Dante's lines. They have changed his structure for reasons external to the direct demands of translation, something that may seem unexpected when the purpose of translation is to provide as best an insight as possible into the author's original work.

Each of the examples so far demonstrate Dante's tendency to mix styles, particularly that of epic and comic. This mixture of styles is no more apparent than in Dante's occasional scatological humour: his crudity. Of the two translators it is Ellis who most exaggerates this, and indeed his colloquial style can more aptly accommodate it. Sayers tends to soften this aspect. Her relative hesitancy to use vulgar terms could be due to a desire to echo a sense of the medieval, meaning that she chose to write in tune with contemporary perceptions of the era or of epic poetry. This would include use of high diction, archaisms and a sense of prudery as opposed to crudity. One instance of this crudity can be found in the final line of Canto XXI, as the leader of a group of devils replies in an obscene fashion to the salutes of his minions. 'Ed elli avea del cul fatto trombetta', Dante writes, translating fairly literally as 'and he made his rear into a trumpet'. Ellis takes no pains to be polite, writing:

            but first, each blew a raspberry
            at their leader, for a salute,
            and he trumpets back with his arse.

While his translation may be intentionally colloquial, this last line is incongruously vulgar. Next to the harmless and childlike act of blowing a raspberry, the word 'arse' comes as somewhat of a shock. The final line is noticeably more playfully rhythmic than what came before it, as if proudly announcing its inelegance. It is the incongruity, then, which creates the humour of this section, rather than the act or the language in isolation. Sayers, in contrast, masks the vulgarity of this last line: 'He promptly made a bugle of his breech'. She does not make the line stand out, linking it to the rest of the section with rhyme, using the dated terms 'bugle' and 'breech', and adding grace with the alliteration. Interestingly, Dante's version places 'trombetta' at the end of the line. This word, which is the key to the metaphor and the humour, fall last, ending Canto XXI. As shown by the reasonably literal translation above, English would also allow this to be the case, yet Ellis and Sayers choose to change the line. It may seem that the change has occurred for no plausible reason, bringing us back to the previous example in which both Ellis and Sayers inexplicably change the word order. This change, too, is unexplained; but it is vital to keep in mind that the process of translation involves not only the constraints of language, but also the constraints of the new poem which is in creation. It is a separate entity from the original text, and needs to cohere and work as a whole. Thus, lines may be written in ways that differ from the original simply because they fit in better with what has come before them. In this case, both translators at least allow the line to retain some kind of comic impact, whether through pure vulgarity or playful alliterative frivolity.

Another instance where Sayers and Ellis diverge in their level of vulgarity is lines 58-60 of Canto XI. Listing the occupants of a particular circle of Hell, Ellis writes:

            of hypocrisy, flattery, enchantment,
            counterfeiting, thieving, simonists,
            pimps, swindlers, crap like this.

The last small phrase is, again, incongruous. The modern, colloquial term is common to the English language of today but at odds with more archaic religious words such as 'simonists'. Perhaps this is Ellis's attempt at mixing the epic and the comic. Unsurprisingly, Sayers puts it more gently:

            Hypocrites, flatterers, dealers in sorcery,
            Panders and cheats, and all such filthy stuff,
            With theft, and simony and barratry.

'All such filthy stuff', if only because of its higher syllabic count, is decidedly less dismissive than 'crap like this'. This impression is reinforced by the way that the list subsequently continues, seeming longer, and the more complex, mixed and broken syntax contributes to a more dignified tone. The tone and the words are more similar to Dante's original, although like Ellis he writes 'e simile lordura', translated possibly as 'and the like filth', at the end. Turning 'and the like filth' into 'crap like this', Ellis takes the dismissive remark to an extreme, undermining all that went before it.

Alliteration forms a particular problem when translating poetry: not only does the translator have to preserve the sense and tone of the line, but they are also restricted by a necessity to repeat first letters. In the 'Last Judgement' sections, Sayers shows her command of such techniques as, to use her own phrase, she 'reproduces Dante's complete alliterative scheme'. One such section is in Canto VI line 95 onwards:

            “Till the last loud angelic trumpet's sounding
            For when the Enemy Power shall come arrayed
            Each soul shall seek its own grave's mournful mounding,
            Put on once more its earthly flesh and feature
            And hear the Doom eternally redounding.”

Here, Virgil is telling Dante what the condition of the spirits will be after the Last Judgement, and Dante widens out from his personal tale to take on a biblical, authoritative tone. Sayers manages to retain the alliteration, and while she arguably sacrifices the meaning of certain words, their proximity to the originals is sufficient. For example, 'mounding' was originally 'tomba', more  accurately translated as 'tomb' but still an adequate representation. 'Doom eternally redounding' is perhaps more accurately translated as 'that which echoes in eternity'.Yet, any sacrifice she has made is worthwhile because the epic subject is suited to the alliteration and its rhetorical poetic effect. It keeps a pulse through this section, creating a mounting movement spurred on too by the rhyme. However, Ellis does not attempt to reproduce Dante's alliteration except where literal translation facilitates it:

            “till the angels sound their horns
            and the enemy chief comes here:
            each will repossess his flesh and figure,
            hear what booms through eternity.”

It is possible that his decision not to reproduce the alliterative scheme stemmed from the difficulty of the task, but the choice may indeed have been deliberate. The solemnly lyrical effect of the alliteration in Sayers's version, the sense of gravity and tension that it produces, is arguably unsuited to the poem that Ellis writes. In his translation of this section, it appears that he has preferred to consume the exalted subject matter with a more matter-of-fact tone. The casualness with which he writes 'comes here' contrasts with Sayers's 'shall come arrayed': he trivialises the act and makes it seem like an accident; a mere happening. The omission of alliteration prevents too much embellishment, and he achieves more understated description of this momentous biblical occasion. This seems to better reflect Ellis's style and the self-consciously modern nature of his translation.

Another instance where his style seems emphatically modern is in Canto. 'just as I've heard of that lance that Achilles and his father had, /harmful at first and then benign'. To say 'I've heard of that lance' is noticeably hesitant; the narrator distances himself from the legend with the word 'that', as if he professes not to be intimately acquainted with classic mythology. Ellis notes that he uses primarily the 'language of the 90s', and it would appear that he reproduces not simply the language of the current era, but also the mindset.

Throughout the comparison, the problems of translation have manifested themselves in the issue of the pun; in the challenge of reproducing humour and a mix of the epic and comic; in the necessity to write a piece that is good in its own right, staying faithful to one's own style as well as to the original. The last point is perhaps the most interesting one, and although cited as a 'problem', perhaps it is something that should be celebrated. The style established by each translator may be developed unconsciously, but as noted, they seem to start with some kind of agenda. Ellis claims that he wishes to 'recapture some of the vigour'; Sayers writes of her decision to use archaic forms of English. So, while the task of translation is of course imprecise, it also gives freedom. It implies not just sacrifice, but also potential: potential for reinvention, which Sayers and Ellis both take  advantage of. Translations are overshadowed by the presence of the original; a kind of holy grail that we wish to access but can only do so imperfectly. This is perhaps even more emphatic in the case of Dante, who has such legendary status as a writer. Sayers writes of her translation 'it is not, of course, Dante; no translation could ever be Dante'. Yet, to say 'it is not Dante' is not the same as saying 'it is not good'. The purpose of translation is to make a work accessible to those who do not understand the language of the original: to, as the Literary Review writes in praise for Ellis's translation, provide a 'door to Dante'. Sayers and Ellis demonstrate the true freedom of translation by making subtle changes that stem not out of necessity, but out of choice. The vast differences in their versions are a result of the way that interpretation is subjective and unique. So, a translation should not be judged solely on its ability to mimic the original. It is also vital that a translation can stand in its own right, enjoyed by readers for what it really is: a new creation. A testimony to the plurality of meaning.

BIBLIOGRAPHY

Dante – The Divine Comedy – 1: Inferno; Italian text with translation and comment by John D Sinclair (Oxford University Press 1961)

Dante – Inferno; Translated, introduced and annotated by Steve Ellis (Vintage Classics Random House 2007)

Dante – The Divine Comedy 1: Hell; Translated by Dorothy L. Sayers (Penguin Books 1949)

Hugh Kenner's introduction to Ezra Pound, Translations (New Directions, 1963)