Japanese Translation: Man or Machine?

Machine translation has proven useful not only in everyday communication, but also for science, education, and business. Yet, there is an ongoing discussion about whether or not translating tools will make human translation obsolete. There have been some very strong arguments to support both viewpoints. We will delve more into this subject by taking as an example one of the most fascinating languages in the world, namely Japanese. The peculiarities of the Japanese language can help us judge fairly the capabilities of modern digital language translators.

Let's dive into the characteristics of the Japanese language and the ability of both human and machine translators in terms of Japanese translation.

Japanese, also known as nihongo, is one of the ten most spoken languages across the world. Currently all of the 128 million native speakers are communicating in Japanese, along with a large number of immigrant communities spread across the world that also speak the language due to historical emigrant migration in countries like South Korea, Taiwan, the Philippines, Brazil and others. Although it is known that Japanese belongs to the Japonic language family, which also includes the endangered Ryukyuan languages, there are considerable disputes amongst linguists and historians as to the true origins of the language.

Therefore Japanese is the only major language whose genetic affiliation is largely unknown, making it impossible to trace the exact roots of the language as there is no compelling evidence associating Japanese to a single family of languages. Some theories suggest that Japanese is part of the Altaic language family that also includes Turkish, Mongolian, Tungusic and Korean, with the closest relationship to Korean. Others believe that the language arose from contact between the language of the Yayoi people, who migrated to Japan approximately 2000 to 3000 years ago, and the language of the Jomon people, who were already living there.

One of the reasons why there is very few evidence of the early history of Japanese is the fact that it was not a written language during the Yayoi period. The absence of a written form of Japanese was also the reason behind the introduction of written Chinese along with other aspects of the Chinese culture in the 4th century, especially after the heavy Chinese influence that the country experienced in the past century. Later Chinese characters began being adopted as the phonetic representations of the sounds of Japanese syllables by Japanese people who were now starting to record extensively their language, as is evident from the earliest examples of poetry and prose dating back to the 8th century.

For the purposes of this article, it is important to note that following some historical events like the Nanban trade in 1543 or the Sakoku period that took place 60 years later, some European vocabulary entered Japanese. The latter was the result of the contact between local traders and missionaries. For example, a significant number of Dutch loanwords entered Japanese during the isolation period of the country, most of which were common everyday words like garasu that comes from ‘glas’, the Dutch word for glass, or kōhī that comes from ‘koffie’, the Dutch word for coffee. The number of loanwords from other European languages began to increase significantly once the isolation period was over and trading between Japan and other countries around the world continued. Many of the new compound words were created with the purpose to represent new Western concepts and also to indicate specific terms used in the academic vocabularies of disciplines likes Arts, Maths, Sciences or Technology.

Japanese Writing: A Complex Mixture Of 3 Alphabets

One of the most notable peculiarities of the Japanese languages is its complex script. As it was really hard during the dawn of the Japanese language to write a language phonetically using another language’s symbols that had no connection to the meaning, the Japanese people decided to develop two simplified alphabets of the Old Japanese. The annotations that helped Japanese speakers read Classical Chinese or Chinese-style texts as though they were Japanese were called Hiragana and Katakana. This was also the precursor of the modern Japanese writing system that continued to develop through time.

As of now, the Japanese language uses three different writing systems that in combination with each other create the unique synthesis of written Japanese. The Kanji were the first logographic characters that were imported from China, as we mentioned previously. Hiragana, on the other hand, is mostly used to write all native Japanese words, whereas Katakana is most often used for the transcription of words from Western languages such as English, French, German or others. Another distinctive characteristic of Japanese writing is the fact that it does not use spaces. Although this might seem as quite a challenge to learners from Western Europe, it is in fact pretty normal for Japanese speaker as they can recognize the meaning of the words just by looking at the alphabets used for their writing.

The Colorful Pallet Of Japanese Dialects

The country’s geographical locations that consist of small isolated islands have fostered the development of multiple different dialects. Standard Japanese today is mostly based on the dialect of Tokyo, yet there are various other colorful regional and local dialects that are spoken throughout the archipelago. Some of them, like those spoken in the southern islands like Okinawa and Kyushu, are incomprehensible even for the native speakers from other parts of Japan.

Therefore, in order for people from the different islands to communicate easily and efficiently with each other, the use of a standard, common dialect is required. The latter is divided into two main groups – Eastern and Western dialect. With the help of modern mobility and mass media such as television and radio, the common dialect has spread outside Tokyo, through most of the regions in Japan. Although this has significantly helped for the communication between native speakers, it has also accelerated the rate of the loss of traditional local dialects that are currently spoken mostly by elder people, working class people, or people in the countryside. When speaking politely, with strangers or in formal situations, native speakers tend to avoid dialectal speech.

The social styles are another peculiarity of the Japanese language making it so intriguing for foreign learners. With its extensive system of politeness and honorific markers, a Japanese speaker must always consider the social standing of the person they are addressing or the person that they talk about. Some of the examples of situations where one should use honorific markers include people that are strangers, older than the speaker or socially superior to the speaker. Depending on social style the honorific markers can appear on verbs, adjectives or even nouns.

Japanese grammar & phonology: How hard it is to learn the language?

In every language there is a basic word order when a sentence is structured. The main elements of this process are subject, verb and object, and in many languages they can switch their order depending on the meaning that is to be conveyed. In Japanese language, however, the rule of thumb is that the verb always comes at the end of a sentence. As this rule is so different than the one of English grammar, for instance, many people tend to believe that Japanese grammar is impossible to learn. However, what they do not realize is that in fact it is not grammar that makes it so challenging to learn Japanese but its phonology.

The basics of Japanese phonology might seem simple at first as it consists only of five vowels, however there are also long versions of these vowels. It is very important for one to learn how to distinguish them as this can significantly impact the meaning of words. In addition, when it comes to pronunciation, another challenging aspect of the Japanese language is pitch accent. Some Japanese syllables must be pronounced with either high pitch or low pitch, which can sometimes indicate a distinction in the meaning of a particular word. For example, the word hashi could mean end, bridge or chop-sticks depending on the accent that one places while pronouncing it. Pitch accent could also differ depending on the dialect of the speaker.

Japanese is undoubtedly a difficult language to learn, however in recent years Japanese has gained tremendous popularity. People are using the language for various different purposes, including for speaking with locals while on a trip to the island, communicating with business partners, watching Anime movies, and more. Learning a language properly, especially a complex one like Japanese, could take many years. This is why nowadays many people prefer to take advantage of online translators in order to facilitate experiences that involve communication in Japanese. Yet, some question the accuracy of a modern device like the online translator when it comes to the identification of different dialects, cultural influences, grammatical rules, and other nuances that accompany a particular language. With respect to the abovementioned characteristics of the Japanese language, we will explore in the following chapter the efficiency of both human and machine translators in an effort to evaluate which one is better.

Studying Japanese has become really popular in recent years with many enthusiastic learners who are interested in learning more than a few useful phrases for an upcoming trip to Japan. Although there are multiple opportunities to learn the language in a classroom in some of the bigger countries, there are still many places worldwide where there is a deficit of Japanese teachers and people must find their own methods of studying Japanese. Therefore when one cannot take advantage of a traditional learning method that involves a person who is a native Japanese speaker or a Japanese tutor, one will often run through online learning. Advanced machine translators help people understand books, movies, songs, dialogues or any other type of written and spoken Japanese.

However, as is evident from the previous chapter, nihongo has many peculiarities in terms of its writing systems, grammatical rules, phonology and dialects. Can therefore machine translators match a native speaker’s touch when it comes to Japanese translation? The answer to this question can be delivered once a comparison between human and machine translators is conducted.

Machine translators have come a long way in a short period of time due to the latest implementations of Artificial Intelligence in every second device that comes out on the market. Used mostly for their ability to translate written texts in the past, modern digital translators can be now applied also for the translation of human speech as well as texts from visual images. The primary method used by computer programmers when a machine translator is being developed are algorithms. When fed to a translator, these algorithms should be able to identify all of the different variables of a language and quantify them into usable data so that an accurate translation can take place. Although it is true that in recent years online translators have been performing decent and meaningful translations between different languages more than ever before, the case with Japanese translations is yet to be further refined until it reaches the accuracy of a professional one.

This has to do with the fact that Japanese is a very complex language and as discussed above, its lengthy and complex writing system can pose a challenge for a machine despite its advanced technologies. Technical accuracy is not the only fundamental aspect of a translation, as there are also other concepts such as style and tone that need to be taken in consideration. For instance, as previously discussed, Japanese has very strict rules when it comes to politely addressing strangers or people that are socially superior. It would be much easier for a person to determine the social style of speech that should be used in a particular situation than for a machine, since although some translators are using AI, there is still a long road ahead until it achieves the processing capabilities of humans.

What is more, as already discussed, native speakers can assign different meanings to the same word through their pitch accents. When a foreigner is trying to learn Japanese with the help of a translation device, the wrong pronunciation of a certain word could highly impact the general context of what one is trying to say. Whereas a human translator who has years of experience in the Japanese language can naturally infer which meaning of a word is being referenced based on the pitch accent of a native speaker, a computer might struggle to do the same. In addition, in Japanese writing all three alphabets are used together in order to construct a sentence. A human translator can quickly determine which one of them is Katakana, Hiragana or Kanji just by looking at the symbols and therefore understand more easily the meaning of the text that is to be translated. Machine translators, however, might not be 100% accurate when they are presented with the task to translate Japanese symbols.

An example of this was observed on the Tokyo’s Toshima-ku district municipality website where an automated service was provided to residents for translations of the webpage’s content from Japanese to English. The website provided general information about the essential aspects of living in the region, such as education and employment opportunities, the local industry, health and welfare and more. Although users are advised that the translations might not entirely accurate, an inaccurate translation could immediately catch the attention of English speakers who wish to view more information on the subject of pensions. The problem here was that the word nenkin, which means ‘pension’ in Japanese, when written in Hiragana, is translated as ‘slime mold’. In this case Google Translate had no context or ideographs to guide it and therefore chose to go with the inaccurate translation providing a service called ‘Information on Slime Mold’ on Toshima-ku’s website.

It would be indeed hasty to say that machine translation will never be able to reach the accuracy and professional approach that a human could provide. However, if such advancement is ever to become true it would for sure happen far into the future. As of now the arguments as to whether human or machines are better to be used for Japanese translation land in favor of humans. Producing comprehensive translations while considering so many aspects of the Japanese language that a real-life person could is simply a too challenging task for computers at this stage.

Furthermore, translation is mainly used by people who are not proficient in a particular language. If one can barely understand a text or a speech, how can they tell one if it was well translated, specifically in the case of Japanese? With such a complex language, it might be better to stick to an interpreter in order for one to avoid misunderstandings, especially for business or formal purposes.