With everything that’s written and said about translation, it’s rather strange that there is so little discussion of the actual process of translation itself. Maybe translators fear that if they reveal their secrets anyone will be able to copy them. Perhaps they don’t want to look too hard at the magic they create in case they suddenly find themselves unable to repeat the trick. Maybe some just think it’s just too boring. But whatever the reason I’ve rarely seen anyone explain exactly how they do the actual task of translating.
I’d like to put an end to that. In fact, I’ve touched on the subject before, in this blog post some years ago. And had the MET conference taken place last year in San Sebastián as planned, Tim Gutteridge, Victoria Patience and I were planning to hold a panel discussion on the translation process, but the pandemic intervened and now it seems as if that round-table is unlikely to take place.
What I’d like this post to achieve would be to start a discussion in which other people also reveal their own processes, either in the comments here or on the various social media platforms where it gets shared – not to say that “my process is better than yours” but to give us a chance to learn from each other and all improve.
So this is what I do. To start with, I simply take my source text, import it into MemoQ and start translating. It’s possible that I will have read the whole text before in advance, but it’s quite possible that I won’t. My first translation draft is very rough. I’m not looking to create a finished text here – the aim is to extract the meaning so I can work on it afterwards. I know colleagues who create a much more polished first draft, but I find that extremely difficult to pull off. It’s as if the processes of deciphering meaning and writing polished English are just too far apart in my brain for me to do them simultaneously. Perhaps that’s why I’m not an interpreter. Too often the right words simply won’t come until I’ve got the basic meaning there to work with, although I do get an occasional pleasant surprise when I look at the rough draft later.
I increasingly feel that the real work comes at the second stage. This is when I look at the first draft and start to make it read like proper English, correcting clumsy or over-literal phrasing, changing awkward words and expressions, making checks and ensuring consistency. Whereas no-one will ever see one of my first drafts, I’d be quite happy to share the result of the second draft process. It wouldn’t be perfect, but it would be pretty decent.
I usually make my third and final draft within MemoQ but looking at the text in the preview function, which shows it more or less as it will appear in the final format. The changes here are usually minor, although sometimes seeing the whole translation like this shows up paragraphing problems and other issues going beyond single translation segments. Because of this, if, for some reason, MemoQ can’t produce a preview or if there are problems with codes or segmentation I will export the file before this stage and do this draft in the final format. Whichever way I do it, though, I’m also on the look-out for any remaining awkwardness: at this stage I try to be ruthless in refusing to allow anything that doesn’t quite read like English to get through.
When I’m happy with the final draft, and assuming I’m still in MemoQ, I run the spellcheck and then run the Quality Assurance check, which looks for possible anomalies involving technical aspects like codes, as well as double spaces and inconsistent translations. When I’ve resolved all the problems this throws up, I export the file into the original format. If that format is Word, I will run its spelling and grammar check and then run PerfectIt, which is a tool integrated into Word that I basically use for catching inconsistencies affecting both grammar and style. I then consider the translation to be ready for delivery to the client.
This process has served me well for several years, but in recent months I’ve become increasingly dissatisfied with it. Having my translations revised by Tim and Victoria in RevClub has made me feel that I’m too closely tied to the source text more often than I’d like. But how can I get further from it, when I start with what is necessarily a very literal rendering and then move away so gradually? I could add another revision stage, but the law of diminishing returns would begin to apply. And in any case, I would find it difficult to justify reducing my hourly earnings even more by increasing the time spent on each translation even further.
Another option would be to put more into my first draft. Once again, though, time is an issue. When I try to force myself to write a better first draft, the right words and expressions seem to elude me and take too long to find. And, while in theory if I had a better first draft the second wouldn’t take so long, I have a suspicion that the overall time needed wouldn’t be greater. Of course, I could perhaps drop the third draft if I worked like this, but I believe my translations would then lose quality. Quite often I pick up small, silly mistakes in the final read-through which I really don’t want getting through to my clients.
I certainly need to do something about the first draft, though. It’s the part of the process that takes the most time yet adds least value, because all I’m doing is producing a very rough translation as a starting point for the real work. What I really need to do is find a way of spending the bulk of the time I devote to a job at the stage where the work is really being done the – the second draft – making that a translation rather than a revision stage, which is what I always used to consider it to be.
How could this be done? In my next blog post I’ll look at a possible solution from what regular readers of this blog will no doubt consider an unlikely source.
Interesting. That is exactly my process (though I don’t use a CAT tool). I find that doing a very rough first draft starts to focus my mind on the text, let’s me be spontaneous (which often produces the best solutions) and gets something down on paper as a basis for the real work. Plus, I sometimes find something which I researched in detail at the time is explained further on in the text or becomes clearer as the text goes on. It might not be the most efficient but it’s the process that works for me. I’ll be interested to read your next post on the subject as I’m always open to suggestions.
Thanks, Alison. The fact that you don’t use a CAT tool is interesting. Sometimes I wonder if it’s a handicap for certain texts, but I’ve been using one for so long I’d find it difficult to do without it, I think.
I can use one but choose not to. I know I produce a better translation without one. One client sent me a long, intricate museum text and picked a couple of inconsistencies. I suggested a CAT tool might have been useful in that instance and he said no. He said he could rectify the inconsistencies but didn’t want me to compromise my style.
I think the advantages of CAT tools probably outweigh the disadvantages, especially as I have a tendency to skip sentences when I’m translating and MemoQ makes that almost impossible. But I can see that in some circumstances having the text broken up into segments isn’t always the most helpful way of looking at it.
I have a very similar process, but I also get Word to read it aloud to me in my final checking stage. I find I spot things that I would miss by just reading. I’m going to have a look at Perfectit, thanks for the tip.
I find PerfectIt very useful, particularly for consistency issues. I have heard that people get the text read by Word. That’s something I may have to look at.
One solution is to dictate the first draft so that part of the process takes less time.
That might work. But I do have one or two problems with dictation. One of them is the habit of typing, which is something I’ve been doing all my working life. I find my thoughts often go to my fingers faster than they go to my tongue. The other is a more practical one: I share a home office with my wife, who’s a journalist. I really don’t want to distract her by spending half the day muttering into a microphone. So if I was going to dictate I’d have to change my whole working set-up and perhaps rent an office.
I was going to say the same thing: dictating often produces a more natural first draft as you hear if it sounds wrong as you say the words. It can also be much quicker, giving you time to polish your output in latter stages. I’ve been dictating since the 80s though; it doesn’t suit everyone and is almost a halfway house between translation and interpreting in terms of thinking off the cuff. It definitely doesn’t work for everyone, but perhaps worth a go?
Many years ago, as a business correspondent, I went to interview some people at Philips who were working on speech-to-text technology before it became a widespread thing and they told me: "This is the future. Typing makes no sense at all." And I thought for a bit, did a few mental exercises to try out the idea, and said: "Not for me, it isn’t. My brain doesn’t work like that." It seemed to me then (more than 20 years ago) that I’d been typing so many years my brain was wired into my fingers (I touchtype), so my thoughts went straight there before they even reached my mouth. Now, I’ve spent two more decades reinforcing that tendency. I have tried interpreting but I’m not very good at it, which seems to bear out the theory. Having said that, I might be tempted to at least try dictation if it wouldn’t also completely disrupt my shared office arrangements, so I think it will probably remain the big "what if" for me.
I’m not a natural typist (was stopped by my grammar school headmistress from going to night school typing classes at 16 or so as clever girls don’t need to type – ha!), so when I started my first translation job in 1984 and was told we had to dictate our translations for the typing pool, it made perfect sense. In fact, I was more horrified when they told us a few years down the line that we had to type our own on the newfangled IBM Displaywrite word processors. I still type with two fingers, albeit quickly, but I can achieve much more when I dictate, so it’s a no-brainer for me, especially when you throw RSI into the mix. That said, when I’m writing my own blog articles, I often prefer to type: maybe that completely free creativity (without a source text as guide) flows better through the fingers?
I don’t think anyone really understands how the brain works in these circumstances and I can quite understand how dictation works for many people, because there aren’t so many touchtypists. But I always feel much more certain that I’m saying what I want to say when I write than when I speak, and I always have done, even before I knew how to type. I’ve always hated telephones (something I had to overcome as a journalist, although I never learned to love them) so e-mail has been a godsend to me. Even when I’ve spoken at conferences I’ve always had a full script. I don’t read it, but it makes me feel better to work out what I want to say in writing first. Despite all that, I would actually be interested in seeing whether I could adapt to dictation, but I don’t want to spend lots of money or disrupt my office arrangements to be able to do it.
Thanks a lot for this post, Simon! My process is very similar to yours, especially considering your earlier replies to comments – I work with a CAT tool as well and don’t dictate either.
I get my rough first draft in pretty quickly, as it helps me get a general sense of the text. Most of the texts I translate are not very well-written, and I find it especially hard to produce a good first draft in those cases. After that, I go over the translation many times, in MemoQ and then in the final format – which tends to allow for more flexibility around sentence length, conciseness, etc.
I have a very detailed self-revising and proofreading checklist for each client, as they all have their idiosyncrasies, and I then use the Antidote corrector (especially good for French) and the Read aloud function in Word, provided the text isn’t too long. After delivery, I reimport the final version in MemoQ with Monolingual Review.
For one of my clients, I use Studio as it really makes the process simpler on both sides, and there are a few more steps, but the general idea is the same.
Also, I try to leave as much time as the deadline allows, so that I can go over the text again if necessary. It’s especially true for that last client, as they want me to produce translations that read really well in French, even if I need to veer from the English a bit (they also offer generous deadlines, so that helps). I’m not so sure about the "law of diminishing returns", though, as I’ve found some of my best ideas in the last or last-but-one revisions of a translation!
Thanks for sharing your process, Nelia. I agree that it’s a good idea to leave as much time as possible before doing the last revision, although I have to say that the best ideas usually come at the second stage. The last one is more for having second thoughts about things that I’ve tried but haven’t quite worked. And I think the law of diminishing returns would be that every time I did an extra revision stage I’d find less and less.
In the scenario of technical translation from En to Es as I assume to be your case, I would consider using an automatic translation engine for the first draft.
I’m translating non-technical ES-EN mostly, but it’s funny you should mention MT as that’s exactly what I’ve decided to experiment with, as you’ll see if you read my more recent post.