Ruminations of a Mediocre Translator

9 August 2019 (latest update 29 April 2020)

I have been engaged in translation-related activities in some form or other for about 50 years. These are a few of my ruminations on translation, what makes a translator good or mediocre, and the nature of translation.

An aid to language learning

Like many people my age, my first encounter with translation was at school, where artificially contrived translation exercises were still a part of foreign language learning. This naturally encouraged a tendency to word-by-word translation, that is, making sense of a passage by first figuring out each word and using known grammatical stuctures to somehow produce something coherent.

Whatever its drawbacks, this bottom-up approach was quite useful in learning a language. It allowed me to gain a rough grasp of grammatical structures in foreign languages and provided templates for creating sentences (if rather stiff ones) in the language in question. At high school I studied German by reproducing its strange word order in English until it became second nature. A similar approach helped me study Japanese at university, where I spent far more time than my fellows parsing Japanese texts to understand how they were built up. Needless to say, this also involved transferring Japanese constructions into twisted English.

Unfortunately, I don't appear to have lost this habit, which became a hindrance in developing as a translator.

My first real translations

My first serious attempt at real translation, for my BA dissertation, involved translating two pieces on linguistics. One was by the postwar structuralist linguist Hattori Shirō. This wasn't too difficult, apart from maintaining terminological consistency.

The other was by Hashimoto Shinkichi, a Japanese grammarian from the 1930s. While Hashimoto is well known as the founder of modern Japanese school grammar, I had never before encountered language and argumentation that were so awkward and plodding. In my struggle to translate it I fell back on my old standbys: parsing and direct translation. While I managed to parse it, I had a much harder time making it sound either intelligible or intelligent in English. I still shudder when I look at my translation.

My first job as a translator

My first job as a real translator was at the Australian Embassy in Tokyo, where half my day was spent translating Japanese newspaper articles into English, and the other half translating other miscellaneous materials. The newspaper translations were for the benefit of Embassy officers, who apparently used them as reference material on reports sent back to the Department.

It was a point of pride at the time that the press translations produced by the Australian embassy were far superior to those of the U.S. embassy. The U.S. Embassy translations were barely translations at all; they merely transposed words and constructions in a way that closely mimicked the original. But there was a rationale behind the American approach. When reading a translation, an American officer who knew Japanese could gain a fairly accurate idea what the original Japanese said. This is what the U.S. Embassy wanted, not natural-sounding translations that obscured the original.

Although the Australian Embassy translations were far more readable than the American, they had their own quirks. One of our translators insisted on translating 豪ホーク首相 Gō Hōku shushō 'Australian Prime Minister Hawke' as "the Prime Minister of Australia, the Hon. Robert James Lee Hawke".

My own personal aim as a translator was to produce natural English while reflecting the original Japanese as far as possible. Since the writing style of Japanese newspaper articles was fairly stereotyped and I'd had plenty of practice reading them, my translations were, I like to think, a cut above my previous efforts.

I had two particular concerns when translating.

The first was ensuring that technical terms, names of organisations, etc. were consistently and correctly rendered. Glossaries that colleagues had built up and bilingual (Japanese-English) terminological dictionaries of the time were indispensible tools.

A simple example of an organisational name was the then 通商産業省 tsūshō-sangyō-shō, universally abbreviated to 通産省 tsūsan-shō, which was officially known in English as the Ministry of International Trade and Industry (MITI). There was no room for off-the-cuff translations, such as 'Department of Commerce and Industry'. If you came up with your own translation, people would automatically assume that you were referring to someone else.

(MITI has now been renamed METI — the Ministry of Economy, Trade and Industry, or to give it its new Japanese name, 経済産業省 keizai sangyō-shō, abbreviated to 経産省 keisan-shō.)

An example of a technical term is what the Japanese call 一般炭 ippan-tan, literally 'ordinary coal', the type used by power stations to generate electricity. This must be rendered as 'steaming coal' (the name by which it was generally known in Australia) or 'thermal coal'.

While this might seem like commonsense, not all people appreciate it. Many years later I had occasion to listen to a hapless Chinese interpreter — possibly an employee roped in to the job at the last minute — translating for Chinese and Japanese business partners on the issue of thermal coal. While I don't remember the specific terms he used, he was essentially translating Japanese 一般炭 ippan-tan 'thermal coal' as 一般煤 yībān-méi 'ordinary coal' in Chinese, and Chinese 动力煤 dònglì-méi 'thermal coal' as the non-existent term 動力炭 dōryoku-tan ('motive power coal') in Japanese, blissfully unaware that Japanese 一般炭 ippan-tan and Chinese 动力煤 dònglì-méi are exactly the same thing!

As a result, a hypothetical conversation that was supposed to go like this:

actually went like this:

Luckily the two parties were familiar enough with the industry and the market to communicate despite the garbled interpreting.

The second concern was finding ways of smoothing the rough edges that result from sticking too closely to the structure and vocabulary of the original.

As I'd learnt from Hashimoto Shinkichi, Japanese prose doesn't always lend itself to easy translation into English. Ungainly and unnatural sentence structures need to be polished and prose needs to be remoulded to follow English-language patterns.

For example, at the time one sticky problem in translating newspaper articles was translating the lead. In those days, a Japanese newspaper article might have started with something like this:

'With regard to the issue of beef and oranges, which has become a major sticking point in U.S.-Japanese relations, the Japanese Minister for Agriculture said today....'.

One highly experienced colleague would translate such a sentence as:

'The issue of beef and oranges has become a major sticking point in U.S.-Japan relations. The Japanese Minister for Agriculture said today...'

Unfortunately this violates the cardinal rule of English-language journalism: the first sentence should summarise the news story. In this case the story was the Agriculture Minister's statement, not the fact that beef and oranges were an ongoing issue.

It was also crucial to find pat English equivalents to Japanese buzzwords, equivalents that could be retrieved instantly without thinking. Dictionaries of current terminology were also helpful in this regard. One example was 合理化 gōrika, literally 'rationalisation', in a business or bureaucratic context. For speakers of U.S. English in particular, this use of 'rationalisation' sounded strange and awkward. Terminological dictionaries suggested 'streamlining' as a more natural equivalent. In this case, however, Australian usage actually tended towards 'rationalisation', so that is what I went with.

In those days 空洞化 kūdō-ka, 'hollowing out', the tendency for companies to send manufacturing offshore, had become a hot topic in Japan. One of my colleagues suggested 'de-industrialisation' as a natural English equivalent. It was a very apt suggestion, but 'hollowing out' appears to have begun life in English, and Japanese was just following suit. 'Hollowing out' was the translation of choice.

At one stage I attended translators' informal 'study sessions' where I learnt some of the little tricks that a translator needed to know. One that sticks in mind even now was Chris Holmes' humorous admonition to "ignore mo (も) when translating into English and it will love you mo' and mo'". [も mo is a particle meaning 'also' or 'even' in Japanese. It is frequently used in Japanese but rendering every occurrence as 'also' does not make for good English.] The point of many of these tricks was, of course, to steer translators away from the traps of word-for-word translation.

A rude shock

It was therefore a rude shock when the Embassy's Labour Counsellor pulled me aside one day and told me that my translations were like a mixture of good English and Japlish. The Counsellor then went through and pointed out all the awkward phrasing in my translations, particularly terminology. The article in question was a Japanese report on Australian industrial relations (labour relations). The Counsellor jabbed a finger at an expression in my translation and told me, "The English should be 'the path of industrial harmony'".

I don't remember what the Japanese said nor how I translated it. I might have used something like 'the policy of peaceful labour relations'. Whatever it was, it was clearly unacceptable to the Labour Counsellor. In hindsight, I realise that he wanted the report to reassure Australian readers that the Japanese reporter had got the message. In fact, he had probably been instrumental in facilitating the reporter's visit to Australia. In those days, Australia's fractious industrial relations record was a source of great concern to Japanese investors. To ensure that he was given kudos for a job well done, it was essential that people back home shouldn't have to trip over outlandish translations.

Although one reason for my poor translation was my tenuous familiarity with Australian jargon, equally important was my ideas about the aims of translation. My desire to acquaint the Australian side as far as possible with the connotations of the Japanese was quite at odds with the Counsellor's thinking.

And so I came to a realisation: my own approach wasn't really too far removed from that of the U.S. Embassy. I was all for polishing translations to make them more readable, but I wanted to reflect the original as closely as possible to indicate how Japanese writers understood the meaning. It seemed important that an Australian reader should get a feeling for the original Japanese, not a nicely manicured version that reassured him or her that everything met Australian expectations. I was wrong.

The coup de grâce to my faltering self-esteem came when I met a Japanese teacher at my alma mater:

Other experience

Apart from work in the embassy, I did gain some other experience at the time. One involved a part-time job for a Tokyo patent attorney. While I never did any patent translation myself, I became aware of the exacting conventions for translating patents. A translation that varied at all from the original was unacceptable because a patent was a legal document. Getting it wrong could have costly consequences.

I also did some work for a TV station translating the Japanese news into English and reading it over the air. When it is broadcast, news must be easy to read and sound natural. I was humbled by the abilities of a Japanese colleague who was exceptionally adept at putting Japanese news items into natural English with almost effortless ease. By contrast, I found myself still unable to completely shed my habit of tracking the Japanese.

Departing for the real world

After some years I decided to leave the Embassy and get a 'real job'. I found myself a position as a research executive with a market research company in Tokyo. Apart from research reports, which require adherence to a particular style of language, I also did reports on Japanese politics, economics, and society for a client. This also involved translating articles, but this time I was freed from the need to convey exactly what the Japanese said. Instead, the aim was to elucidate different aspects of Japan that the client might be interested in. I like to think this led to an improvement in my translation skills, but I can't be sure.

China

A few years later I decided to depart Japan's fair shores for the broad horizons of China. Although I never returned to professional translation, many of the jobs I held in China (often in the role of 'assistant' of some kind) involved some degree of translation work, mostly from Chinese to English or Japanese to English, but also from English to Chinese or even Chinese to Japanese. Some of my superiors were Chinese, some were Japanese, and one was Pakistani Irish.

On sundry occasions I was asked to translate contracts. These needed to be professional-sounding and convey the meaning correctly. Luckily, since my translations were mostly for internal reference, I was not held to the stringent requirements of translating legally-binding documents.

In a PR position I held at a hotel, I used to translate publicity and promotional material, explanations to be read by hotel guests, and even restaurant menus. Publicity and promotional materials needed to present the hotel and its services in an attractive light to entice customers and guests to use them. Menus needed to lead guests to sample restaurant dishes; instructions (on how to open a safe, for instance), needed to be clear and easily understandable. This was not just 'translating words'; it was translation for a purpose. A direct translation of the original was not necessarily appropriate or adequate.

In another job I was required to interpret between Japanese and English for foreign politicians or former politicians. This was challenging work because the nuances of words were difficult to capture. I remember one politician using the term 'entrenched attitudes'. Now, what the person was referring to was (if I remember rightly) xenophobic or exclusionary community attitudes that tended to emerge in certain political situations. Google Translate gives 定着した態度 teichaku shita taido ('established attitudes'), which is a good enough fit although not necessarily the most appropriate for the context. I think I used 先入観 sennyūkan, meaning 'prejudice, preconceived notion'. Given that the speaker was probably consciously avoiding the word 'prejudice' in English, this was perhaps an unfortunate choice, but coming up with a suitable Japanese euphemism for 'entrenched attitudes' on the spur of the moment was a tough call.

Japanese tends to use anodyne formulaic language in discussing politics, which can differ from the style of foreign politicians. On one occasion I did a back-translation from fairly anodyne Japanese back into fairly anodyne English for the benefit of the same politician to confirm that I was conveying the right message. She said that that was "not how she would have put it" but grudgingly admitted that it was roughly what she meant. Translating for politicians requires great sensitivity to the nuances of their message. I would not like to be in the shoes of people who have to interpret for international leaders.

A different type of translation

In a later job I was tasked with gathering industry news (including political and economic news as necessary) for the head of the trading company I worked for. This involved translating from English and Japanese into Chinese, which was challenging. I ended up making heavy use of Google Translate and spending a lot of time fixing up its poor translations. I am not particularly proud of the results. Compared with the smooth, polished translations that some of our student trainees produced, my translations were extremely poor.

But I found that I had two advantages over less experienced Chinese translators.

1. I understood the English almost perfectly, unlike even the most talented students. I remember one student translating 'wet weather' as 'damp weather', not realising that it referred to rainy weather. While this is a fairly elementary example, I discovered that even translations from major newsagencies at times contained errors of interpretation.

2. As time went on I came to realise that much of the information in news reports is background information or filler. When translating into Chinese, I came to omit such superfluous information as 'the State of Queensland, located in the northeast of Australia' since this was common knowledge in the industry. In this way, my translations were stylistically poor but accurate and to the point.

One thing I did learn from this experience is that translating into a foreign language does not contribute greatly to your translating skills because you are always working within the constraints of your existing knowledge. Translating out of a foreign language affords far more opportunities to expand your vocabulary and expressive abilities in that language.

Failures

There have, of course, been failures over the years. I am not a great interpreter as I tend to get flustered and make mistakes. And there have been less than happy translation experiences, too.

On one occasion I was privately asked to translate the preface to a volume of work by an artist. I was given both the Chinese original and a Japanese translation. The Chinese original seemed reasonably clear in its meaning, but I had no idea how to translate it in an appropriate style. The Japanese translation (by the artist's wife) seemed to me roundabout and unsatisfactory, full of clumsy constructions that strained to bridge the gap between Chinese and Japanese. I gave them my translation and heard no more about it. Later I heard that the artist's wife was highly dissatisfied with my work. I suspect that my fairly straightforward translation was found to be inartistic. Unfortunately I will never know. The lady provided no feedback and I failed to grow from the experience.

More recently I did two small jobs for a Japanese client about the challenges of providing sewerage facilities after an earthquake. The first was for a speech to be given by the client. I did a translation that was fairly close to the original in language and structure, with which the client was satisfied. The second covered similar content but was for a public poster in English. Given that public posters have limited space and need to be easily understood, I streamlined and improved the language. The client was most upset and spent all night trying to restore the original language. Perhaps I had finally mastered the art of shaking off slavish adherence to the original text, but ironically I failed to satisfy the client, who obviously believed in the very style I was trying to avoid!

What I have learnt

I would like to be able to say that as a result of my long experience I am a font of wisdom concerning translation. That, unfortunately, is not the case, but below are some of the things I learnt or realised from doing such workaday translation over the years. Much of this is commonsense, but it may have some modest relevance to the world of 'Translation Studies'.

1. Translation is a goal-oriented activity, and can only be understood in terms of the goals it tries to fulfil. There is no ideal translation: only a translation that is fit to purpose. Translation is often divided up into 'legal translation', 'literary translation', etc., but these labels only make sense in terms of the requirements of the finished product.

If the U.S. embassy wants transparent translations that are directly relatable to the original, that is what is required;

If a translation is required to persuade or impress people, the translation must be smooth and persuasive;

Translations done with the aim of attracting or enticing customers must present the product or service in an attractive light;

If a translation needs to satisfy political requirements, it should be written to do so. (One of the problems of Chinese overseas information services is that they are meant to impress overseas readers while satisfying political requirements back home, an impossible task);

A legal translation must be in the style of a legal document (in order to be taken seriously) and must be legally watertight;

Translations for politicians must be highly sensitive to their political concerns;

A diplomatic translation must follow the peculiar forms and language of diplomacy;

Translations for people requiring key information need to be accurate and useful. A beautifully polished translation that gets figures wrong is worse than useless;

Literary translation must deliver a literary experience. People expect high literature to read like high literature, detective fiction to read like detective fiction, and comic stories to be comic;

The expectations of literary translation are not frozen in time. Pope's rendition of Homer in heroic couplets was appropriate for his time but not for ours, which expects the words to speak directly from ancient times.

A translation as a personal exercise and a translation for public release have different 'audiences' and therefore different requirements.

Academic translations (often with footnotes) are not expected to be literary, nor are literary translations expected to be academic, although it is a bonus when the two coincide.

The requirements of dubbing and subtitles for movies are totally different. The main restraint on dubbing is the time frame for delivery. On the other hand, because viewers need to read the subtitles while watching the action, there are strict requirements on the number of words that can appear at one time and the speed at which they appear. The translator must be able to render the message as succinctly as possible without losing too much information.

A story like The Little Prince might be translated conventionally, or it might be translated specifically as a story for children, with important changes in the text to accommodate this goal.

In the real world, regrettably the goals may be less lofty... like satisfying the demands of a Japanese client who believes that a good translation is what he learnt to do at university.

I guess you get the picture....

2. The corollory is that, although we may one day have machines to do it for us, translation is a human activity. Most importantly, a translation is a result of the choices that a translator makes. These are what will determine the overall style, tone, and accuracy of a translation.

Many factors on the translator's side will affect the quality of a translation:

The translator's understanding of the original. It is truly surprising how often misunderstandings and misinterpretations mar published translations.

The translator's experience of both translation and the field in question. Translators spend a lot of time getting a grasp of the subject matter and learning the appropriate style and vocabulary.

The translator's sensitivity to words and their use. Some people are good writers because they have a good writing style, a rich vocabulary, a feeling for words and their collocations, and wide exposure to different kinds of writing. Others less so.

The translator's quirks. Different translators will have different ways of tackling a source text and features of that text. Like my colleague who had his solution in translating the leads of newspaper articles, all translators have their own approaches and techniques. This can have a major impact on the style of the translation.

The translator's background. I strongly suspect that some mechanical translation habits can be traced back to the very start, when the translator was sitting in the classroom learning the foreign language. Such habits can die hard.

3. Most contemporary translation is at base word-for-word.

Despite the call for 'dynamic equivalence' — that translations should not be word for word, but should as far as possible create the same effect in the reader as the original did — the translator is ultimately working with words on the page. The choice of words and the order in which they are placed are the starting point for all translation.

It is not only the inexperienced translator whose first instinct is to look at a text and consider how each word and construction might be rendered. Experienced translators do so, too, because that's what texts are made of. To improve his or her skills, a translator must strive to avoid excessive concern with the exact form of the original. This does not happen overnight. It might start with something as simple as consciously avoiding the rote translation of specific structures (such as the unthinking translation of Japanese も mo into English), or the avoidance of overreliance on fixed equivalents. Some people, of course, have a natural talent for writing and find less difficulty producing sparkling prose than others. But like all writing, there is seldom anything effortless about translation.

Corpus-based studies, with their emphasis on specific features of comparable translations, hold promise in elucidating the specific ways in which translations differ with respect to the source text, and thus the skills, habits, and idiosyncrasies of translators.

(Note: An exception to the above characterisation is certain types of oral interpreting, especially in talks or on-site visits, where the interpreter is concerned with conveying the message rather than each individual word. This is one reason why experience with interpreting can be beneficial to translators by taking them away from the habit of concentrating on individual words and structures.)

4. People tend to overlook the huge role of standardised international vocabulary.

A good portion of modern so-called 'translation' means knowing the conventional equivalents for standard vocabulary. A translator working in a major language simply has to plug in ready-made words. For example, the word 'economy' or 'economic' has routine equivalents in Japanese (経済(的) keizai-teki), Chinese (經濟 / 经济 jīngjì), Mongolian (эдийн засаг ediin zasag in Mongolia, аж ахуй aj akhui in Inner Mongolia), or any other major language.

Just 150 years ago people translating these terms did not have this luxury. Apart from European languages, most languages did not have a single word that corresponded exactly to 'economy' or 'economic'. Translators in the 19th - 20th centuries had to create such vocabulary from scratch as part of Westernisation / modernisation, laying the basis for what we have today. For the translator, this makes the difference between asking "What is the Swahili word for 'economic'?" and "How should we express the concept of economics in Swahili?" This kind of standardisation now covers vast fields of science, technology, and even sport, making a lot of translation an exercise in memory or dictionary lookup rather than brainstorming.

The field of Translation Studies has in recent decades seen an efflorescence of new approaches, often based on linguistic corpora. These are refreshing developments, but many appear to be concerned with what might be called the 'ideal translator', on the model of Chomsky's 'ideal speaker-listener', translating an ideal text in an ideal way. The problem is that there is no such thing as an ideal translator nor, in many cases, an ideal way of translating texts. Much depends on the reason for doing the translation. Any approach to translation must take this into account.

Sorry, I'm now using Disqus for comments. If you'd prefer not to use Disqus, please send me an email and I'll list your comments separately. Thanks!