The good, the bad and the ugly—ensuring quality in Japanese‑to‑English translation

The Good, the Bad and the Ugly—ensuring quality in Japanese‑to‑English translation

(Since the terms “translation” and “interpreting” are often confused, it is worth reiterating that “translation” refers solely to the rendering of the written word into another language. The terms “source language” and “target language” refer respectively to the language of the original text and the language into which it is translated. The following abbreviations are also used: J2E for “Japanese‑to‑English”, NSE for “native speaker of English” and NSJ for “native speaker of Japanese”.)

What’s up with that, Japan?

Something is rotten in the state of the J2E translation market in Japan. In Europe it is axiomatic that a translator should only translate into his or her own native language. Thus, only a native German speaker will undertake an English‑to‑German translation. The rationale is that even a person who has studied another language extensively will inevitably have a greater “active” command of his native language than of his second language. The term “active” denotes the ability to identify the proper word in a given context, to use technical terminology and idiom, and to accurately craft natural‑sounding sentences. (Correspondingly, “passive” facility denotes the ability to comprehend text or speech in the language.) A true bilingual may have an equally well developed active facility in two languages, but such people are extremely rare.

Few people write well in their own native language … The number of people who can also write well in a second language is much lower—the large majority undoubtedly being people working in pairs of languages that are structurally similar, such as Japanese and Korean. For a language pair as structurally dissimilar as English and Japanese, people with strong writing ability in both are virtually non‑existent.

Translation is essentially a two‑stage process—thorough comprehension of the source text followed by precise rendition of the meaning into the target language in written form. Only the second stage—writing—requires an active command of the target language. Thus a translator ought to be a native speaker of the target language, not the source language. Moreover, few people write well in their own native language (let alone in one they have acquired). The number of people who can also write well in a second language is much lower—the large majority undoubtedly being people working in pairs of languages that are structurally similar, such as Japanese and Korean. For a language pair as structurally dissimilar as English and Japanese, people with strong writing ability in both are virtually non‑existent.

In Japan, however, with its huge J2E translation market, the “target‑language native speaker only” caveat goes largely ignored, and much of this J2E material ends up being translated by NSJs—often with dire results. There are several reasons for this.

What’s the problem?

First, there is a shortage of proficient NSE translators, since Japanese is still not studied much in English‑speaking countries (except for Australia). Hence, relatively few NSEs have attained a level of Japanese reading ability sufficient to enable them to translate effectively. (The increasing prevalence of software dictionaries, which spare the translator the need to memorize kanji readings, has alleviated this situation somewhat.) Hence, a company requiring J2E translation may struggle to find a suitable NSE translator.

Second, some companies with no foreign staff attempt to save money by assigning translation tasks to in‑house NSJ staffers—no matter how limited their English ability.

Third, there is a perception in Japan that Japanese is so complex and subtle that only NSJs can understand its nuances well enough to translate from Japanese into another language—and therefore that no NSE could ever comprehend Japanese well enough to adequately render the meaning of a source text into English.

Fourth, some Japanese managers believe that the best way to manage J2E translation is to have an NSJ familiar with the subject matter translate the source text and an NSE checker edit the resulting English text to produce a polished final version. This assumes that although the translator’s English may not be perfect, his output will be clear enough for the editor to easily divine the correct meaning and turn the text into natural English.

Fifth, NSE J2E translators are often less adept at speaking Japanese than at comprehending written Japanese (two quite distinct skills), and this hampers their ability to convince potential clients to entrust them with J2E translation projects.
Sixth, some English documents are intended purely “for show”—not to be read or understood. Hence, the quality of the English is immaterial.


Experience has consistently demonstrated, however, that the rationales for using non‑NSE translators for J2E translation are fundamentally flawed. Twenty-five years ago Japan was universally admired for its consumer electronics devices, but ridiculed for the English contained in the accompanying user’s manuals, which provided a stark illustration of the need to avoid non‑native writing in the commercial arena. The manufacturers have largely rectified this problem and now produce manuals of a quality commensurate with their products. However, NSJ‑authored English is still prevalent in Japan—in company reports, in corporate promotional literature, on company websites, and on barely comprehensible signs in airports, conference centres and sports stadia throughout the country.

Two reasons cited above for reliance on J2E translation by NSJs merit further discussion.

The belief that only an NSJ can appreciate the subtleties of Japanese well enough to render the meaning correctly in English is rubbish, and denigrates the intelligence and diligence of numerous non‑Japanese students of the language. As stated, the translation process requires only a passive understanding of the source language, and the diligent student can match an NSJ in this regard.

The notion that acceptable J2E translation might be produced by an NSE checker correcting the output of an NSJ translator is not so easily dismissed. (The arrangement may indeed be workable under ideal conditions, with the NSE checker having unlimited access to the NSJ translator for clarification of questionable points; however, this is rare in practice.) The operative word here, however, is “acceptable”. Editing non‑native English rarely results in text that reads smoothly. More importantly, the individual or company commissioning the translation will often seriously over‑estimate the translator’s English ability and under‑estimate how difficult it will be for the NSE checker to divine the intended meaning. Thus, in practice misunderstandings are likely to occur between the translator and checker, resulting in errors of meaning and nuance. Hence, the final text will be prone to imperfection—in both accuracy and style.

Advocates of the NSJ translator/NSE checker combination for J2E translation often buttress their claims with the argument that an NSJ translator with specialist knowledge of the field will do a better job than an NSE who lacks such knowledge. This is a complete red herring, however. Knowledge of the subject and of the technical vocabulary and phraseology used in the relevant literature will naturally confer a significant advantage. However, this alone will be insufficient to overcome the deficiencies inherent in using a non‑native translator. High‑quality J2E translation of specialist material certainly requires an NSE translator with sufficient understanding of the subject and adequate knowledge of the attendant terminology. However, this is entirely separate from the issue of the translator’s native language.

How to get it right

How, then, can top‑quality J2E translation be achieved? Experience demonstrates that consistently excellent results can be obtained as follows: first, a qualified NSE translator with knowledge of the subject matter translates the source document; second, an NSJ checker with an advanced level of English ability checks the translation against the original source text and points out any outright errors or mistakes in nuance; third, the NSE translator incorporates the NSJ checker’s corrections into the translated text; and fourth, an NSE editor performs a final stylistic edit and check for consistency. The two NSEs thus make maximum use of their active facility in English, while the NSJ brings to bear his likely superior passive facility with Japanese. This use of high‑quality personnel and three pairs of eyes is a proven method of achieving a near‑perfect translation that will read as smoothly as an original document written in English.