On the challenge of translation
Last summer, I read a piece by Brian Merchant on how modern AI technology has decimated translation jobs. It featured interviews with a number of translators and makes for sobering reading. Quite apart from the very real human loss and de-skilling, I thought Merchant did a good job of describing what we’re set to lose by handing over translation to machines:
The quality of translations across the board, from video games to corporate communiques stands to decline, with AI output, according to interviewees, often being homogeneous, blind to local details, or flat-out wrong. Nuances about places and cultures, recognizable to a knowledgeable human interpreter risk disappearing, sanded down by blunt-force automation. It’s not overly dramatic to say that we risk losing the capacity for cultures to understand one another better if we’re all simply feeding output into each other’s automated translation systems.
Needless to say, I agreed with every word of this. There’s a pop tendency to think of language as being synonymous with words, which suits generative AI companies who can reduce language to a numbers game. Communication is about much more than words, though, and linguistics has a whole sub-discipline devoted to how we make meaning with language beyond the bare bones of words and grammar. This is known as pragmatics: the contexts, intentions and social relations that give life to what we say. Pragmatics is what we lose when we filter language through LLMs.
When you’re reading, studying and translating ancient literature, a lot of energy goes into reconstructing precisely these pragmatic contexts. I’d argue that philology in its older sense—the sense of a discipline that doesn’t distinguish language from its production, so to speak—is particularly interested in patching together context, and especially of trying to understand intention. Translating even a single word can require detailed interpretative excavation.
In this post, I’m going to take you through the sorts of things a humanities scholar might consider when translating a text from a thousand years ago. I’ll be using the Old English poem The Wanderer as my example (you can read a quick overview of the poem and listen to my translation of it in my previous post, here).
Part 1: The comfort of the crib
To recap, The Wanderer is a lyric poem that is mostly from the perspective of a warrior who has lost his lord and companions, and who finds himself in a reflective mood amid icy landscapes. The poem starts, though, with five lines that frame what is to come, and it is these lines that we’re going to focus on here:
Oft him ānhaga āre gebīdeð,
Metudes miltse, þēah þe hē mōdcearig
geond lagulāde longe sceolde
hrēran mid hondum hrīmcealde sæ,
wadan wræclāstas. Wyrd bið ful āræd.
Even in such a short passage, there is a dense amount of critical and formal interest, and numerous potential problems of translation present themselves. My own crib that I refer to when I used to teach the poem is what I would call a relatively ‘close’ or ‘word-for-word’ translation, and goes like this:
Often a solitary one experiences grace, the Lord’s mercy, though he, sorrowful at heart, had to, for a long time, stir the ice-cold sea with his hands over the sea-way, walk the paths of exile. Fate is fully determined.
I don’t attempt to emulate the poetic metre of the Old English, though I do mostly follow the word order insofar as it doesn’t completely upend the meaning, which makes it sound a little awkward. This is what we might call an unexciting, “transparent” translation, one that aims to take the original and convert it entirely into a modern idiom that is as faithful as possible, using Peter Baker’s gloss to the poem as my guide, since that was the one my own students would use.
However, “transparent” suggests that you can transmit the meaning of the text in a way that doesn’t distort or misrepresent what the original Old English says, while the term “faithful” implies that I have some sort of special access to the meaning intended by the author, not to mention the meanings that would have been understood by (potentially quite diverse) contemporaneous early medieval audiences. Is my translation really “transparent” and “faithful”, and can it ever hope to be transparent and faithful?
Part 2: In what sense is the wanderer alone? Or, how words don’t behave like we want them to
Let’s take that word ānhaga as our starting point for what is at stake here. A recent editor of the text, Anne L. Klinck, says in her note to this line that [quote] ‘the meaning of the word ānhaga… is clear enough, essentially “one who is alone”’. Sure enough, the Dictionary of Old English follows this meaning, stating, among other definitions ‘solitary being, lonely being’, while the older Bosworth-Toller Anglo-Saxon Dictionary offers ‘one dwelling alone, a recluse’. I probably don’t have to point out that even here there is a disparity in meaning, with a term like recluse implying a conscious separation from society rather than chance or imposed separation.
As Anne Klinck notes, however, the etymology of this word brings with it some complications. It is a compound, meaning it is made of two different words; the first part is the word ān, meaning simply ‘one’, but by metaphorical extension, also ‘alone’ or ‘only’. The second part is haga, which means ‘[an] enclosure’, yielding the meaning (quoting Klinck again) ‘one who is enclosed, or encloses himself, alone’. But were Old English speakers alive to any metaphoric potential of the word ānhaga outside of its “transparent” meaning of a ‘solitary person’? Elsewhere in The Wanderer, we find metaphors such as ferðloca (‘life enclosure’) or brēostcofa (‘breast chamber’). These are kennings, the sort of taut metaphorical compounds that characterised Old English and Old Norse poetry, and both are part of a common literary motif of talking about the body as a sort of enclosure for the soul. Did ānhaga have more to it, in that case, perhaps meaning something like ‘one who has turned in on themselves’, a feeling familiar to anyone who has suffered anxiety or depression.
This is already enough of a headache for anyone translating the text, but things are more complicated still; later on in the poem, we find the following lines:
…..sorg and slǣp somod ætgædre
earmne ānhogan oft gebindeð.
(sorrow and sleep together often bind the wretched solitary thinker).
The second line echoes aspects of the first line of the poem, with the universalising oft (‘often’) and close aural consonance between gebideð (‘experience’) and gebindeð (‘bind’). The one that stands out, though, is ānhoga: the word contains the letter <o> rather than <a>, which has led scholars to believe that it might instead derive from a combination of ān and the verb hogian, ‘to think’ or ‘to meditate upon’, and Klinck offers the etymological definition ‘one who meditates alone’. Suddenly the sense ‘recluse’ might make more sense. At some point we maybe had two distinct words, ānhaga and ānhoga, that, due to their similarity, saw their meanings eventually fall together.
But are we to assume that a smart, literate Old English author (or indeed a smart illiterate, oral performer) wouldn’t have been alive to the possible word-play that the two forms of the word entails? The author of The Wanderer shows an interest in both contemplative reflection and the imaginative productivity of confinement metaphors. The problems for the would-be translator of The Wanderer are many, but, as a philologist and critic, I have the luxury to be able to pick over etymology, consult dictionaries and criticism, compare use of ānhaga and ānhoga in other poetry, look at near-synonyms for both words, then present you with a brief history of the word. And when it comes to making a crib for reference or guiding students through the poem, I can just tell them to “play safe” and go for ‘solitary one’: no fuss, no ambiguity, no subtlety.
The creative translator may want different effects, though, depending on their sense of the poem. The choices are manifold, and each comes with its own anxieties or distortions. ‘Outsider’, for example, gels nicely with the theme of exile, but possibly occludes any contemplative force that ānhoga connotes, and indeed the possible freedom implied in being an outsider is difficult to square with a word that might mean ‘one who is enclosed’. ‘Loner’ has the attraction of being unexpected, but comes fraught with negative connotations that don’t necessarily seem apparent in the Old English original. The Oxford English Dictionary’s definition, ‘a person who avoids company and prefers to be alone’, also doesn’t really reflect the fact that the wanderer did not, in fact, choose their own separation from society. There are numerous other ways in which we might choose to translate ānhaga, including words like ‘individual’, ‘introvert’, even ‘misanthrope’, each of which are plausible while presenting their own problems, and ways of rendering ānhoga are similarly various, if not more jarring. ‘Thinker’ or ‘solitary thinker’ work, which is what I choose for in my own crib of the text, but other near-synonyms for thinker all sound slightly odd: scholar, philosopher, sage, intellectual. Each of these is tinged with an uneasy sense of anachronism.
Part 3: Lost in translation (no, it doesn’t get easier)
And if you think this is the end of the translator’s problems, then we are only getting started. One further challenge is that an Old English author might well have carefully selected one or the other form of ānhaga/ānhoga for stylistic effect, being alert to the connotative and etymological differences between the two. If we opt for ‘solitary one’ in line 1 of the poem, and ‘solitary thinker’ in line 40, the senses might be somewhat faithfully represented in each instance, but we in fact lose the aural and visual consonance between the two words, and by extension the various resonances in meaning that might have been apparent to a contemporary medieval audience. We can’t just select an equivalent pair of modern English words with overlapping senses, and that happen to have a single difference in vowel quality, and hope to replicate the same sort of effect. Jacques Derrida wrote on this precise problem, suggesting that:
‘Whenever the unity of the word is threatened or put into question, it is not only the operation of translation which finds itself compromised; it is also the concept, the definition, the very axiomatics, the idea of translation that must be reconsidered’.
The ‘unity’ of a word can be thought about in any number of ways, from the etymological and semantic, to the aural and orthographic, but every translation risks such damage. There is no easy way around this, and indeed, after first claiming not to be interested in words, Derrida spends a good eight thousand words of that particular essay thinking through how best to translate Portia’s line ‘When mercy seasons justice’ from The Merchant of Venice into French.
So far, I’ve talked about the problems inherent in translating a single word and we haven’t even touched upon the rest of the passage that I quoted. There are lots of other choices we have to make: does gebīdan more properly mean ‘await’ or ‘experience’, for example? The solitary one can be argued to experience grace in spite of (or even because of) the physical and mental hardships they endure, but at the same time one might feel deprived of God’s favour in such difficult circumstances, and thus biding patiently for evidence of it. Elsewhere we have the metudes miltse, ‘the lord’s mercy’. Do we integrate it into modern syntactic tastes, or leave it detached and disruptive, forcing us to pause over it? And how do we deal with the noun metod itself? This is an example of a poetic synonym that Old English authors drew on when they needed to meet the demands of metre and alliteration; there are lots of others that mean roughly ‘lord’, including frēa and dryhten. But were these words completely neutral synonyms, or was there more going on? Some translators might look to etymology and the fact that metod is related to the Old English verb metan, ‘to measure, evaluate’, (archaic Modern English to mete [out]), and thus go with the poetic ‘measurer’. Such a translation emphasises the Christian god’s role as judge, and the unusual stylistic use of ‘measurer’ in modern English might make us dwell upon it a little more.
Elsewhere, we could straightforwardly translate the word hrīmcealde as ‘ice cold’. But equally, we might want to go for ‘rime cold’, with the word ‘rime’ being an archaic dialect word meaning ‘hoar-frost’ or ‘frozen mist’ and ultimately derived from the first part of the compound, hrīm. Such straightforward replacement has the potential to signal to the reader that this is poetry out of time, but also risks grating if the rest of the translation retains a modern idiom. For the word mōdcearig, we could choose ‘mood sorrowful’, incorporating the etymological descendent of mōd, much as Ezra Pound rendered the compound mōdwlonc (‘proud in spirit’) as ‘mood-lofty’ in his translation of another Old English poem, The Seafarer. But what works well for Pound, works less well here: if we take the mōd in mōdcearig to be equivalent to ‘spirit’ or ‘heart’ or ‘mind’, then ‘mood-sorrowful’ perhaps doesn’t catch the sense of desperation inherent in being sad to one’s core.
Part 4: Translating the most famous line in Old English (no pressure)
Which leads us to the perhaps the most famous line in all Old English literature: wyrd bið ful āræd. In my own crib, I render this as ‘fate is fully determined’. Richard Hamer’s translation of the poem from a mass-market paperback has ‘fate is relentless’. Either way, the result is unavoidably portentous: you have a destiny, and there’s nothing you can do about it (fans of The Last Kingdom may catch echoes of Uhtred son of Uhtred’s ‘destiny is all!’). Scholars have written extensively about this line, so what I have to say here is very much the truncated version.
First, we have to consider whether the idea of a teleological end point, of something preordained, is coherent with the internal logic of the poem and a wider medieval Christian view of the world. There is an apocalyptic fate of sorts, a gæstlic (spiritual) time when eall þis world wela weste stondeð (the wealth of the world stands desolate), but that is a different question entirely from the personal fate of an individual anhaga, one who, according to Christian thinking, should have possession of free will. Indeed, as Kathleen Davis has argued, much of the poem deals with gaining understanding via experience: for þon ne mæg weorþan wīs wer, ær he āge/ wintra dæl in woruldrice (‘therefore no-one can become wise before they have their share of winters in this world-kingdom’). Choice, the opportunity for spiritual growth, is central to the philosophy of the poem: we’re not bound to be a hopelessly lost eardstapa, weary in mind, as long as we turn our soul to religious understanding.
A translation of wyrd as ‘fate’, then, could be unhelpful. Probing the etymology of the word helps: wyrd, it seems, is derived from the proto-Germanic verb *werþan-, which arrives in Old English as weorþan, meaning ‘to come about, happen, become’. Klinck notes that various editors have, on this basis, chosen to translate the word instead as man’s lot, or in other words, that which comes about or happens, or simply events. Elsewhere, we find the Old English word wyrd-writere to refer to a chronicler or historian, or one who records events. In this interpretation, the word retains a sense of fatalism without suggesting that we are beholden to something inevitable.
The final word of the passage, that is ārǣd, is more often than not parsed as ‘resolute’, or’ inexorable’, or even ‘ordained’, reinforcing the traditional interpretation of wyrd as fate. In a short but persuasive piece, Kathleen Davis has argued that this interpretation should also be rethought: the word ārædan, of which āræd is the past participle, can mean something along the lines of ‘determined’ or ‘settled’, but often also means ‘to interpret, construe, decipher, read’ (indeed, it is from the related form rædan that we get the modern word ‘read’). In poetry, the word always seems to be used with this sense of interpretation, of making sense of things. She goes on to say:
‘I suggest [that] this clause references the relation between the ānhaga’s exceptional status and his grasp of wyrd; by virtue of his experience and wisdom, he has come to terms with , and can decipher what to others may seem mysterious, like a riddle… For him, wyrd is ‘determined’ only in the sense that he has come to a determination regarding the events of life, and the depth of his understanding enables him to become snottor on mode – a man able to convert experience into wisdom’.
Davis’s conclusion is, I think, convincing, but it is interesting that she doesn’t offer a full translation of the half-line, instead simply leaving both wyrd and āræd untranslated in her quoting from the poem. This is partly because she is giving a complex, philological, interpretation of the phrase’s possible semantic range, not providing a ‘readable’ translation for a general readership. If I were to come up with something reasonably ‘accurate’ – something that you might expect to offer in a classroom setting – then we could go with something like events are fully understandable, perhaps with an underlying modality, i.e. events can be fully interpreted. While fate is fully resolute sounds ominous, perhaps even a little bombastic, ‘events are fully understandable’, gives a sense of hope – we might even imagine an unsaid ‘so don’t worry about it’ tagged onto the end. Indeed, if we were feeling particularly adventurous, we might even go for ‘shit happens, deal with it’. In my creative translation of the poem, I decided on ‘things go as they must’, which I wanted to echo the sort of proverbial folk wisdom that you find throughout Old English and other medieval literatures.
I don’t want to suggest for a moment that any of these translations are definitive, or even that we shouldn’t translate the phrase as ‘fate is fully resolute’ – god knows grammars and introductory textbooks still stick with some version of this phrase – but to emphasise that translation is not a straightforward or completely neutral act. Many people will disagree with Davis’s argument and my proposed translations. But the point is that translation opens up meaning in way that is as difficult and frustrating as it is thrilling, and when you read any text translated by a human, you should always be attuned to the fact that what you are engaging with has, at every point, gone through a process of careful selection and mediation. An LLM will always choose the statistical path of least resistance; a half-decent translator can’t help but seek out those boggier, overgrown paths and see where they lead.
In the final post of this series we’ll take a look at a couple of famous translations of Beowulf to see how different perspectives on a text can tease out new ways of understanding it.
This post is based on a short lecture series I gave to undergraduate students at the University of Oxford in 2020 and 2021.
Dr Nik Gunn 
Leave a comment