10 differences between Japanese and English
Potential issues in Japanese to English translation
Japanese and English aren't just different languages, they also come from cultures that deal with things differently — even some concepts that you might consider universal.
Here are 10 important differences to bear in mind for localisation projects.
1. One and many: inferring number from nouns
English has singular and plural nouns: one book, two books.
Japanese doesn't. No matter how many objects there are, the noun remains the same.
This means that unless a number is explicitly stated or given elsewhere (e.g. in an illustration or table) it won’t be clear from the Japanese text alone whether there is one item or many.
Without the ‘clue’ of a singular or plural noun, the translator will have to make a best guess based on the context, or be consistent and always translate the same noun as singular or plural. Without an external source of information, singular/plural nouns may end up being incorrect or inconsistent in the translation.
Solutions
Before translation: Provide as much reference material as possible: diagrams, figures, tables, user manuals, catalogues, etc.
During translation: Expect queries as the translator works and give them a contact who has, or can get, the answers.
2. Ladies and gentlemen: inferring gender from pronouns
English uses gendered pronouns to refer to people already mentioned: he mentioned, she disagreed.
Japanese has some equivalents, but they’re rarely used (it’s normal to continue to refer to people by their last name) and a text may not give any clear indication of the gender of the people involved.
Not knowing how to refer to a person can force unnatural translation choices (such as copying the Japanese convention of referring to people by name) or cause embarrassment by using the wrong pronoun.
Japanese forms or registration processes also explicitly ask for a person’s gender, where English gets that information from their salutation (Mr / Mrs / Ms / Miss).
Solutions
Before translation: Provide reference material such as biographies or photographs.
Check forms and other places where personal data is collected to see if the user is asked for their gender and consider asking for a salutation instead.
3. What’s in a name? Reading Japanese first, last and place names
Pronunciation of names in English isn't always simple or consistent.
Japanese has an additional complication: the pronunciation, or reading, of the characters used for names is not fixed.
For example, 光 can be read Aki, Akira, Hikari, Hikaru, Kou or Teru.
The reading of Japanese names have to be rendered in English in translation, but it’s impossible to determine the correct reading from the Japanese characters alone.
There are dictionaries of name readings, but ultimately only the person in question or someone who knows or has access to them can confirm how to read the name. This applies to first names, last names and sometimes place names. (Japanese has no concept of middle names.)
Solutions
Before translation: Provide reference material such as personal biographies.
Confirm the readings of names that appear in the text.
Confirm the reading of place names where possible.
Article: What’s in a name?
4. On this day: converting between different date systems
Dates might not come to mind as something that needs translation. English uses the Gregorian calendar, divided into BC (Before Christ) and AD (Anno Domini).
Japanese sometimes uses the Gregorian calendar, but also expresses dates in the years of the reigns of the Japanese emperors.
For example, 2016 is also Heisei 28, the 28th year of the reign of Emperor Akihito.
Some QA processes or systems (human or automated) expect numbers to be identical in the source document and the translation and flag any differences as an error. However, as era dates are not widely understood outside of Japan they should be converted to the Gregorian calendar in English — the numbers need to be different.
Some documents may also give two versions of the same date (Gregorian and era) but only the Gregorian one should be kept in translation.
Solutions
Before translation: Make sure that numbers are not removed or made uneditable in the text.
After translation: Build flexibility into the QA process to handle correct localisation of dates.
5. Complex numbers: not a universal language
Another potential surprise is that numbers themselves may need translation. Japanese and English use different numbering systems.
Numbers in English are delimited with commas for thousands, millions and billions. Japanese sometimes uses the same system, but also has an alternative system based on multiples of 10,000 with units written in characters, or a mixture of both.
For example, １千 is 1,000, 1万 is 10,000, and １億 is 100,000,000.
As above, this may cause problems in QA if numbers are expected to remain the same, but as this numbering system is not used in English the numbers need to be converted.
Solutions
Before translation: Make sure that numbers are not removed or made uneditable in the text.
After translation: Build flexibility into the QA process to handle correct localisation of dates.
6. Shifting sounds: the gap between the sounds available to languages
The most obvious difference between English and Japanese is script — English uses only one while Japanese has three (two phonetic and one pictorial). Another difference that isn't immediately visible is the different sounds available to the two languages, and how to convert one to the other.
Japanese uses a simple set of sounds laid out on a 5 x 10 grid. There are several systems for converting the Japanese phonetic alphabets into English, and differences between them can cause discrepancies in spelling, especially for names.
For example, 松本 can be rendered as Matsumoto or Matumoto, and 石田 as Ishida or Isida.
The Hepburn system is widely used because its spellings are closer to English sounds, but individuals or official documents may use another system where some sounds are given different readings.
This can lead to one spelling being used customarily (e.g. Matsumoto) but another being used on a passport (e.g. Matumoto).
There is also some artistic licence in choosing a particular spelling for the English version of a Japanese word, including flexibility between l and r (the source of many a joke) and c and k. Macoto Tezuka (son of legendary manga creator Osamu Tezuka) spells his name with a c although it is normally spelled with a k.
Solutions
Be aware that Japanese names may not be pronounced as they are written.
Before translation: Always check spellings on official documents, such as passports, especially when doing something like applying for a visa.
Article: How are Japanese words written in English?
Article: What sounds are used in Japanese?
7. Being unfaithful: balancing closeness to the original with readability
Translation involves an inherent tension between faithfulness to the source text and deviation from it to to create something that reads naturally in the target language.
A translation can be too faithful. Accuracy is important, but staying too close to the Japanese (e.g. replicating sentence structure) results in stilted English that fails engage the reader and creates a negative impression.
For example, the title インストールについて can be translated as “About installation”, but it’s more natural if you take out “About” and use just “Installation” as a title. Omitting part of the Japanese source creates a better English translation without losing meaning.
Translation focuses on conveying the meaning of the original — not mimicking the structure at the sentence or word level. An over-emphasis on accuracy, to the point of recreating Japanese grammar in English, results in poor English text.
Solutions
Before translation: Use a style guide or sample English documents to guide the style of the translation.
Make sure the translator knows the subject and has enough time and other resources to do the job properly.
Article: What are source and target languages?
Article: All about titles
8. The anatomy of a Japanese typo: Japanese input systems and how they can go wrong
Typos in English are easily resolved — there are a limited number of possibilities for what a word should be, especially in context.
Japanese input involves typing the pronunciation of the word and choosing the correct characters from a list, so a slip of the mouse or finger can give a very different word to the one the writer intended.
For example, all of the following terms are read “tenki”:
天気 (weather), 転機 (turning point), 転帰 (natural death), 転記 (transcription), テンキ (dune grass).
Errors in the source text can cause confusion and misinterpretation in translation, particularly if a typo results in a different but plausible term.
Solutions
Before translation: Make sure the text for translation has been checked and proofread.
During translation: Expect queries and as the translator works and give them a contact person who has, or can get, answers.
9. Staying on topic: inferring the subject
It’s hard to have a sentence without a subject in English, where it’s normally explicitly stated.
Japanese is a topic-based language which frequently omits the subject if it has already been established.
This means that translations from Japanese can end up using the passive tense because the subject isn’t given, even when it’s not appropriate. The translator can insert a subject to give more idiomatic English, but it has to be the right one. Machine translation often inserts an orphan “it” with nothing to refer back to.
Solutions
Before translation: Make sure the translator knows the subject and has enough time and other resources to do the job properly.
During translation: Expect queries and as the translator works and give them a contact person who has, or can get, the answers.
10. Dear reader, dear writer: saying what has been left unsaid
High-context and low-context are anthropological concepts describing how cultures communicate.
The USA, UK and other English-speaking cultures tend to be low-context, where the writer or speaker supplies context, where Japan is high-context: the reader infers the context.
These differences may require additions to the translation for it to make sense or have the same impact in English.
High-context cultures rely on shared understanding rather than stating things explicitly in words.
やる and 差し上げる both mean “to give”, but the first implies the recipient is below the giver (e.g. a plant or animal) while the second implies the receiver is above the giver (e.g. a customer or someone respected). The difference encoded in the Japanese has no English equivalent.
Solutions
Be aware that cultural differences will need to be accommodated.
Before translation: Make it clear to the translator that you expect them to raise potential issues.
Make sure the translator knows the subject and has enough time and resources to do the job properly.
During translation: Expect queries and as the translator works and give them a contact person who has, or can get, the answers.
Summary
Make sure the the text to be translated has been checked and proofread.
Provide as much reference material as possible: diagrams, figures, tables, user manuals, catalogues, photographs, biographies.
Confirm the gender and pronunciation/reading of first and last name of people mentioned in the text, especially for official documents and visa applications.
Expect queries as the translator works and give them a contact person who has, or can get, answers.
Use a style guide or example English documents to guide the style of the translation.
Build flexibility into the QA process to handle localisation of dates and numbers.
Make sure the translator knows the subject and has enough time and resources to do the job properly.
Be aware that cultural differences will need to be accommodated and ask the translator to raise potential issues.
Find me on arlinelyons.com, on Twitter, on LinkedIn or on Google+
