How Machine Translation Works With The Complex and Wonderfully Unique Japanese Language.

Written by:
Hiroshi Gunji

All languages are unique. Developing over thousands of years and reflecting cultural, geographic and other influences, languages occupy an incredibly important and unique place in human development and interaction with other humans.

In this blog, I would like to examine some of the features of Japanese that have made it not only unique but as anyone who has tried to learn the language knows, extremely difficult.

While spoken Japanese was developed in Japan by and for the Japanese, over a long period of time it has also been very strongly influenced by regional and cultural trends that are quite unique.

One major influence on Japanese was the importation of many ideas and philosophies from China, including the very script, kanji, for writing the language. The wholesale introduction of the Chinese writing system created some of the many difficulties that non-native speakers encounter.

For example, there were of course many perfectly useful Japanese words in use before Chinese ideographs were introduced, but often, at the same time as the Chinese kanji were introduced, so was the Chinese reading. For example, the kanji for mountain has 2 separate readings, “yama” and “san”. Same kanji but 2 readings depending on the context. The “yama” is the original Japanese word and “san” is the imported Japanese reading. Hence Fuji, the mountain can be referred to as “Fuji yama” or “Fuji san”. I remember reading where a foreigner who obviously did not know much about Japanese kanji said that the Japanese so love Mt. Fuji that they call it Mr. Fuji. You see, to make this story even more complicated, it turns out that ‘san’ also means Mr. or Mrs. etc.

As a consequence, when learning kanji, the non-native speaker must learn multiple readings for the same kanji, sometimes as many as 4 or 5. So although the meaning of most kanji is relatively simple to understand their reading often changes depending on the context.

Another interesting point in Japanese is that pronouns are used as little as possible. So we would say “Does your brother like school?” In Japanese, this becomes “Oniisan gakko suki desuka?”. And the answer would be “Suki desu” which means “Like”, just that. But that’s not all, as a further complication, there is one word for big brother if my own brother and that is “ani” whereas if it is your big brother then I must say “oniisan”. If it is my young brother then I say “otooto”, while your young brother would be “otootosan”. Japanese also uses long vowels hence the double “i” in “oniisan” means the vowel is lengthened as with the double “oo” in “otooto”.

Now to make it even more complex, Japanese uses 3 different scripts. First, there are the kanji imported from China, then another script that is phonetic and used principally for grammatical functions such as verb tenses etc. and finally there is one more script which is used for foreign words.

These are just a very few examples of the complexity of Japanese that make it one of the most difficult languages to learn by humans. So, imagine how much harder it is for machine learning to conquer translation between Japanese and English. And I have not yet even begun to discuss the various levels of politeness depending upon to whom you are speaking, and another separate level decided by whether you are referring to your family or someone else’s.

A great deal of work is being done in the field of machine translation (MT), and while some of that work on MT can translate Japanese into English and English into Japanese effectively; there is still a lot of work that needs to be done to perfect MT when it comes to the uniqueness and complexity of the Japanese language.

This is why, at Straker Translations, we strongly recommend translations into and out of Japanese are edited and reviewed by skilled and experienced Japanese linguists, who can help to ensure that the content is contextually correct and reads as if it was produced locally by a native speaker.

While machine translation hasn’t mastered the Japanese language just yet, MT is still making progress with other languages and depending on your industry and language pairs, machine translation can be beneficial to your translation project.

We recommended that you work closely with your language service provider to find the best solutions to your language needs and to ensure that you get the best outcomes.

You can read more about what exactly machine translation is and the four different types of machine translation here.

Looking for a language service provider? Get in touch today!