Recently I came across an interesting thread on Twitter by user クリシー that talked about how difficult the translation industry was, specifically in that there often aren’t many resources and training opportunities, resulting in “J-E newbies use[ing] stock translations like “It can’t be helped.”(仕方ない) Or “I won’t forgive you!”(許さない)”.
When I read this I was inspired to write another post about translation, something that I spend a lot of time on but haven’t posted about much lately.
It’s been around 7 years since I began translating (primarily fiction/literary works, but also some real-world stuff through Gengo), and I still remember one of the first projects I worked on where it took me an unbelievable amount of time to translate a handful of sentences, I think at least 10-30 minutes for fairly simple prose without tricky grammar, expressions, or cultural references.
I remember really enjoying the experience of racking my brain to try and come up with half-natural-sounding English, and also I remember being surprised by how difficult it was and how much time it took. Even a few years later, while my efficiency has improved, at the core I feel translation is simply an extremely challenging endeavor that takes a lot of brainpower. (I’ve been focusing mostly on fiction translation from Japanese to English, two languages that are quite linguistically different, and perhaps the difficulty is fundamentally less for other language pairs and types of content.)
When I realized that translation is actually a creative endeavor––similar to creating a story, or an improvisational melody on an instrument––I finally started to understand and accept why good translation will always take time and serious thought. It isn’t my place here to prove it is a creative endeavor (some people may disagree), but I feel that just the fact you can have several good translations of something, each which themselves look very different, goes a long way to provide evidence for that conclusion. Nevertheless, I am pretty sure anyone who has any experience translating will agree with me immediately.
But I want to get to addressing the main part of that Twitter thread, which is about the lack of resources and opportunities for budding translators.
As the author of Self Taught Japanese you probably won’t be too surprised to hear that my translations skills were mostly self-taught; I haven’t taken a class on translation, or even read any books that focus on the details of it (except maybe a few pages here or there). Now I’m not saying that direct instruction for translation (or language acquisition) is bad, and perhaps some people will be able to ramp up quicker that way. But I feel that ultimately these things can, and maybe should, be learned more effectively when you take control of your own learning. Also, I want to emphasize that you shouldn’t be literally sitting in a room by yourself with no materials and just magically learning to translate. It’s fine to talk to other people, and especially get feedback from them, but ultimately nobody should be teaching you “this is how you translate”.
Even though translation is a creative process, like painting a picture or singing a song, you can still steadily improve your quality (and your efficiency, to a lesser extent) over time.
But first––and I apologize since this is surely in every article ever written on translation skill improvement––you have to make sure you have good fundamental abilities in both languages. This means the ability to fluently read and understand the source language (Japanese), and the ability to not only understand, but be able to communicate effectively in the target language (English).
And because languages are so complex and diverse, you really need experience in the genre you are dealing with to have a chance at an accurate, high-quality translation. In my case, I focus mostly on fiction/literature not just because I enjoy it, but also because I have read many books in both Japanese and English that can be more or less categorized as literature. While I haven’t spent nearly as much time on writing in Engilsh, I have penned a handful of short stories and portions of several novels (here’s one recent effort). Actually, I feel the editing skills that I picked up when I was doing creative writing are really paying dividends in my translations. As a side point, I highly recommend that translators of fiction put aside time to practice creative writing, because if you can’t write natural prose in English (or whatever your target language is), you aren’t likely to write natural English when you are translating.
If you are not confident with your skills in either the source or target language, make it a point to keep reading and keep writing (listening and speaking is great too) so you can continue to enrich your education in those key areas. Even if you are confident in your skills, keep exposing yourself to more things to further improve your lexicon and your understanding of word nuances.
Now we come to the hard part: what else is needed besides the above to become an ass-kicking translator. Let’s say you are totally bilingual in Japanese and English, with strong skills in reading and writing in the genre in question. If you are starting out in translation, as I mentioned earlier you may still find it can be very hard to capture the essence of the source text into a natural-sounding target text (this was alluded to in that Twitter thread). The whole tradeoff between literal and natural is where much of the creativity is required, since the translator must constantly make little decisions that can have subtle, and sometimes major influences on the reader. And once you factor in how the source text will often have nuances not literally present in the words themselves, plus the details about the historical/cultural context, you have a pretty big task on your hands.
Reflecting on my thought process when I translate, the main flow involves brainstorming words and phrases that seem to provide similar meaning as the original text, mixing those together to produce a natural, grammatically correct English sentence, and then iterating on that. While the act of translation can be very difficult, the actual process is somewhat simple. (Here is an article I wrote on my translation process, though it’s a bit outdated now.)
The key is the iterations. I always go through a series of iterations, where I gradually distance myself from the source text. In the first iterations I am looking for a natural result, but also double-checking that I got the basic meaning right. In the later ones, I am basically ignoring the source text and making sure the result is natural and consistent, not just on the sentence level but as a whole. (See this article on fractals and translation for more about iteration.)
Coincidentally, around the time I was finishing up this article I heard part of a radio interview with famous composer John Williams. He was asked a question about composing, and said something like (this is based on my memory, not a direct quote): “My music may seem simple, but there is a lot of complexity going on behind the scenes. Often I will spend time trying to adjust a single tone, a single note, to get things perfect.” When I heard this, it reminded me of how iteration factors into translation. I hesitate to compare the work of a true master composer to that performed by translators, but ultimately I think at the core the same creative process is at work, even though translation does not involve creating something new “from scratch”.
There is a key skill that I don’t think I have talked about before: accurate self-evaluation of your translated text in the target language.
Let’s take the example in the Twitter thread about people using “It can’t be helped” as a candidate translation for “仕方ない”. The magic process of iteration only works if you clearly understand and feel your source text. What nuances and atmosphere does it have? Does it flow well? Does it sound natural in the sense of something a real person would actually say, or something you would read in a well-respected book of literature (or comic, etc.)? Note that this assumes the source text feels natural in the source language, if it is, for example, (purposefully) awkward in some places, you should try to convey that awkwardness in your translation.
You can find a phrase like “It can’t be helped” in two places: dictionaries and (bad) translations. But you’ll rarely hear that in the real world, and probably never see it in a (non-translated) comic book or novel. While I think there can be an element of subjectivity when talking about whether an expression is natural or commonly used, to me “It can’t be helped” is clearly an awkward, rarely used phrase. If you disagree, then I would like to politely suggest that perhaps you need to read more books/comics originally written in English, and start paying more attention to nuances and what phrases are actually used and which aren’t. (Ultimately I think using the target language in your day-to-day life is an important step, though surely many translators can’t afford to do that for one reason or another. This is why people often translate to their native language, since after all they already have much of that natural experience.)
To clarify, I am not saying “originally written in English” because all translations are bad. In fact, there are many good ones; but there are still bad translations out there, so I would recommend avoiding the risk of “poisoning” yourself with phrases like “It can’t be helped” unless you have some level of confidence about the quality of the translation. If your primary reading material is (unprofessional) fan translations, then you will likely pick up such unnatural phrases and eventually start using them yourself. (Again, I am not saying all fan translations are bad, I’m sure there are some awesome ones out there.)
Judgment of whether a piece of text sounds natural can be a pretty subtle, and a pretty tricky thing. Even those who normally feel they have strong language skills can be biased, for example when someone asks you “Does this sound natural?”, in which case you are more likely to think it sounds strange. On the other hand, just reading through text in the context of a story (when you are trying to enjoy and not find fault) may actually desensitize yourself to subtle issues. I’ve also noticed seeing subtitles in English when the audio is also English can make me feel like the language is unnatural.
One really important thing is to give yourself time in between revision sessions, since once your brain has forgotten about all the extra stuff (especially when you just translated) you will be a better judge. While it isn’t usually feasible for projects with a tight deadline, if possible I suggest waiting a week, or even a month, and then see on a reread how you feel differently about a certain translation you have performed.
You can also use specially-crafted online searches to help gauge the naturalness and commonality of a short phrase. If you get only a few hits of mostly (what you feel are) unreliable or bad translations, then that’s a hint you may have found an especially cheesy or unnatural way to translate something.
Another option is to find someone to proofread your work and give you feedback. Finding a person experienced enough to compare a translation line-by-line may be difficult (unless you pay them enough), but you may have a little more luck searching for a person just to proofread your translation without worrying about the source text. Feedback can be extremely helpful if it comes from a person with a refined sense of language who can accurately (and politely) express their opinions.
Japanese has a host of things that make translation to English difficult, and I could write a bunch of articles on the technical details to give you some suggestions and get you thinking (actually in a few cases I have), but my point here is that for the long term you want to get off these crutches and be your own boss when it comes to translation. Hone your language sensibilities, especially your ability to distinguish text that sounds awkwardly vs. naturally written, and use iteration to gradually refine your translation until it’s a shining thing of beauty.
But regardless of how much someone tells you “practice makes perfect”, what do you do when you come across a phrase for which you have no idea how to translate? You could skip it and come back to it later, but as a rule I always try to put down at least a guess at a translation, no matter how uncertain I am about it (just make sure you leave a note for yourself to follow up on that in subsequent editing iterations, for this I usually use high highlighting in the first few iterations).
There are usually two causes, one is that you don’t understand the original text well enough to even attempt a confident translation. Usually in that case, looking through enough dictionaries and other reference sources will give you enough info to get a good feel. This may include also looking at more instances of the word used in natural texts of the source language (web articles, books, etc.), to see what kind of context it is frequently employed in. If you are truly stumped, you can always try to ask a native speaker or post to an online forum.
As for the case where you feel you have a good grasp of the source text, but still are clueless as to how to render it in your target language, creativity is key here. Sometimes trying to find the perfect translation can limit your options, so if you are completely stumped I would just try and freely write down at least 5-10 attempts at a translation, and in that brainstorming process you may come up with more than you expected.
Also, there is nothing wrong with trying to find existing translations that already use the word or phrase in question. For this, you can try web searches with the word or phrase in question, plus “English” or some other English word(s). In my experience, often those found via this method will usually be either awkward translations or those where the context doesn’t fit what you are translating, but if you search enough you may find something that just clicks for the situation you are looking at. Again your language sense is what is key here, otherwise you will end up cutting and pasting phrases that are unnatural or don’t fit the context.
I want to stress that for the initial translation, don’t try to get things perfect since you have too much on your mind then to really make the best translation. It’s not uncommon for me to think up new ideas to render an expression in a natural way after a few editing interactions. Usually I don’t know where they come from; it just seems my subconscious mind provides an answer after going through the text enough times. This is another reason I can’t help but feel that translation is truly a creative process.
A final trick I wanted to give (and this is something that has helped me immensely) is to break things into manageable pieces. For example, if I come to a long paragraph that has, say, 20 long sentences, it can feel like a nightmare. In cases like that, I often break it into smaller chunks so I can fill up only half of the screen (just don’t forget to rejoin them later!), and that relieves much of the stress. You can also do this with long, complex sentences; pick bite-sized phrases and translate those, and then rearrange when you are done and tweak to patch things together. This usually won’t give a terribly natural translation, but it will help you understand the original sentence better and think of different options to render that meaning in the target language.
Before I finish up I wanted to talk briefly about translating 許さない as “I won’t forgive you!”. While this one is not as bad as “It can’t be helped”, there is certainly an unnatural feeling to it. Out of curiosity I did a Google search for “I won’t forgive you” and in the top 10 or so results was a page with someone asking about why anime characters use this phrase so commonly. But whether you are reading someone’s translation or doing your own, especially for phrases in a modern context you usually can just ask yourself “Would I ever say this?” Or, for this case, perhaps it’s better to say “If someone did something horrible such that I would not be able to forgive them, what would I say?” There are clearly a bunch of natural answers here, but probably the simplest is “I will never forgive you!” If we remember that “won’t” is a contraction of “will” and “not”, then the main difference here is swapping “not” for “never”. I am not sure if there is technically a reason that “I won’t forgive you” is not natural, but sometimes natural language doesn’t need to have a logical explanation. Of course, depending on the situation there probably are other translations of this phrase that will be more natural.
If I were to be completely honest, when I saw “I won’t forgive you!” (許さない) my first thought was “What’s wrong with that?” This is perhaps a bias because I saw the Japanese word at the same time, and part of me tried to translate it literally to English. But taking a step back and trying to imagine what I would say in that situation made me realize it is unnatural after all. That’s why you should do at least one (probably a few) iterations where you aren’t even looking at the source text (one problem here is that you might have memorized parts of it after many rereads, which is why you need to put some time between iterations to forget). Being able to refine and be aware of your sense of linguistic judgment is one of the most important aspects of becoming a truly great translator.
Now if you are working on jobs with a tight timeline, then all of the above will feel rushed and you will have to do your best in the allotted time, and probably make a bunch of mistakes. That’s why I suggest trying to fit in time for side translation projects where you do enough iterations that you feel the text ‘settles’ to a stable point, and try to perfect your skills. If you can do a few of those before you do actual paying jobs, even better. But at the same time, time management is an important part of the process, and that too will get better only with practice. Over time, you will develop instincts as to where it makes sense to put in extra research time and when to not.
I hope this article was helpful, especially if you are a translator. To me, just the idea that you don’t need to have a teacher or attend a course on translations to become a good translator (given enough time) is empowering. If there are any areas you want me to dig deeper in on follow-up articles, please let me know.
Also, if you are a beginning translator and have an especially difficult-to-translate short piece of text feel free to put it in the comments (ideally not more than 3-4 sentences). If it’s OK for me to use it in an article, I may write up an analysis of it, including some ideas on how you can translate it.