Recently, I was doing some peer review on a translation. I thought it was a good translation, and recommended it be published with only a few tweaks, but it was one of several things I’ve seen recently that raised a question which might be interesting to discuss. We do a lot of translation on this website, and if you’ve read them you’ll know that we tend to render every word in English. The translation I was reviewing, and it’s not the only one I’ve seen, kept several words in Latin. I’ll say up front that there’s a case for either approach and I don’t think either is objectively worse, depending on your perspective and goals; but it may be worth writing down why I go for the former option.
First, let me give you a concrete example of what I mean. Here’s a passage from the Bachrach and Fanning translation of Flodoard of Rheims’ Annals (chosen because it’s the easiest example of this in print for me to get – as I write this, which is several weeks before it will go up, I’m in isolation with COVID so my resources are a bit more limited than they otherwise would be):
…King Louis gave the castrum of Amiens to Erluinus. Heribert’s sons took the munitio of Clastres, in the pagus of Vermandois, due to the treachery of Raoul, one of King Louis’s fideles. This Raoul secretly slipped out of the stronghold when Heribert’s sons entered it and plundered the treasures before abandoning the deserted municipium.B.S. Bachrach & S. Fanning (trans), The Annals of Flodoard of Rheims (Toronto: 2011), p. 39.
As you can see, in these two sentences, five words are untranslated: castrum, munitio, pagus, fideles and municipium. Were I translating the same passage, I would have rendered them as ‘citadel’, ‘fortress’, ‘district’, ‘followers’ and ‘fortress’ again. So what arguments could justify either approach?
The main argument for leaving some words in Latin is that the word is ‘untranslatable’. In some cases, English can’t convey any of the nuance. Take Vergil’s Aenid, which repeatedly refers to its main character as pius Aeneas. That seems like it should be ‘pious Aeneas’, but a Classically-trained friend of mine once spent a good fifteen minutes explaining to me that this is a very flat translation: pius doesn’t mean that Aeneas is religious, necessarily (although it does include that meaning with its ambit) but that he is dutifully loyal in appropriate ways, especially towards his family. In other cases, English can’t convey wordplay or puns. Gregory the Great’s famous pun that beautiful English slave-boys looked like non Angli sed angeli (‘not Angles but angels’) does work in English, but his follow-up puns don’t – they were from the kingdom of Deira (modern Yorkshire), and Gregory responded bene Deiri, de ira eruti (‘Deira is a good name for it – they will be snatched from God’s wrath’). Or – my favourite – the description of Dominicans as domini canes, ‘the dogs of the Lord’, which also doesn’t really work in English (‘Dominicanines’?). Finally, there are technical terms which some translators deem more appropriate to keep in Latin, in much the same way that people who work on eighteenth-century France tend to keep the French word gabelle rather than putting it into English as ‘salt-tax’. I see these most often in the case of titles (dux is probably the most significant, but marchio also gets this treatment) and fortifications (some well-known medievalists have argued that the endless different Latin words for ‘fortified place’ – arx, castrum, oppidium – and so on all have different technical meanings and so leave them untranslated). I personally find this last point the least sympathetic, not least because at least a few of the arguments that X or Y is a technical term are tendentious and not translating them smells too much of stacking the deck. That is, by leaving these terms untranslated, an artificial sense of a coherent technical terminology is created which might be as if not more misleading than just putting the word in English. Nonetheless, there are certainly ambiguities. Dux is a case in point: I don’t think anyone would object too strenuously to translating dux Normannorum applied to the Viking leader Ragnar in 845 (as indeed it is) as something like ‘Viking chief’ or ‘Scandinavian warlord’; or to translating the same phrase applied to William the Conqueror as ‘duke of Normandy’ in 1066 – but what about when applied to Richard the Fearless c. 970?
All these points are valid, but to leave words untranslated because of them seems to me like an abdication of responsibility. This is a personal point of view – another person may very well see it as due caution – but let me try and explain. To start with, I don’t think translation is the process of transparently rendering a Latin text into an English one; it is (in an academic context, at least) a work of mediation and explanation, a tool to help with understanding the original version. This is so far from being a controversial opinion that it might be almost a commonplace – to go back to the example from Flodoard at the top of this post, the word ‘stronghold’ in the translation isn’t in the Latin but has been added by the translators, making the structure of the sentence smoother in English but also providing a gloss on munitio and municipium – but I think an approach which says that some words are translatable and others aren’t runs the risk of implying it. To take the Bachrach/Fanning translation as an example again, they choose to leave dux in the Latin but render comes as ‘count’, which implies that a comes is straightforwardly comprehensible to English speakers in a way that a dux isn’t – and I am on record on this very blog as not thinking that’s true.
What that means is that not translating Latin words hamstrings a translation’s value as a tool to aid understanding. An important part of translating a text is deciding what the words mean, and refusing to do that in (say) 3% of cases means that the translation is only 97% useful as a tool. It certainly means you as a translator have to make interpretative judgements; but the whole translation is an interpretative judgement, making refusal in particular cases somewhat arbitrary. Sometimes, choosing one rendition of a word or phrase is tricky – but this is exactly the sort of thing that scholarly apparatus exists to discuss: nuance, wordplay and technicalities find their home in footnotes. This does mean that translations are not very useful for doing detailed linguistic analysis of sources. On the other hand, that’s not their job: a translation isn’t a source, it’s a translation of a source, and anything which rests on particularities of the language needs to be related back to the original text. For this reason, I think my comments apply even to translations aimed at an audience of academics, who might have more grounding in the source languages than interested non-academics – anyone who is going to be doing serious research based on an historical text will need to have a copy of the original to hand anyway if the language matters, because by their nature translations are useful but not dependable for this: trust, but verify, as the saying goes.
As I’ve said several times, I think leaving words untranslated is a legitimate choice, and I don’t find it reprehensible or unjustifiable. However, I do think that translating all the words, even the difficult ones, is more helpful, and that’s why I do it.
(Oh, and it almost goes without saying that calling this a post about ‘translation philosophy’ is a little pretentious – there is a large literature which is actually about translation philosophy, to which my only exposure is a couple of comments at the front of Penguin Classics. Consequently, these are only the little musings of a rough-and-ready practitioner and may seem rather naive to anyone who’s actually well-versed in these ideas!)