With increasing digitalization and increased internet penetration, the usage of computers has also increased. A few decades back, only a handful of households had computers. Today, the size of a computer has shrunk to the palm of your hand. Anyone with a smartphone and an internet connection can learn new languages, find their way around a new country, and even establish a dialogue with complete strangers.
The technique of using computers for translation took over the world of translation by storm. Upon its introduction, companies, and businesses were quick to adopt computer-assisted translations. Along with providing a host of benefits, computer-aided translation has made the task of translation an easier one.
But are computers really capable of delivering accurate translations? Are there any languages that are particularly hard for computers to translate? What about Japanese translation by computers? Let us find out.
Why Translation by Computers is Hard?
The task of translation is a difficult one in itself. Although there are millions of translators online to make the translation process easier, computers are now widely used for translation. But there are certain limitations to the translation by computers. Here are all the reasons why translation by computers is difficult:
The translation is a “difficult” task
It goes without saying that the task of translation from one language to another is a difficult task. The translation process is a multi-faceted skill that is greater than simply converting one language to another. The job of a translator is to take a text in one source language and produce the text in another target language. This conversion works on the principle of “equivalence”--the translated text must somehow be equivalent to the original text, in meaning and context. Establishing that equivalence is difficult for a computer, especially when no human intervention is present.
Human translation is needed alongside computer translation
Let us be honest; the task of a translator goes well beyond just picking a text and translating it into another language. Human translation is executed successfully when certain criteria are fulfilled; the nuances and background of the original text are understood fully, and the translator has knowledge as well as expertise in the target language. In the process of translation, a human translator often needs to produce a text which is clear, unambiguous, interesting, persuasive, and gripping. Although these are understandable, they are quite unfair as well, especially when automation is on the table. A computer can only be expected to produce a translated text that is somewhat “equivalent” to the source text. However, it cannot be asked to be “creative” when translating the text. This means that anything the computer translates will have to be post-edited for “creativity” and “persuasiveness”. A human translator can only add these qualities and a creative element to the translation. Only a “draft” quality translation can be expected from a computer–one that is a starting point for a polished translation.
All languages do not have precise “equivalents”
The task of translation involves converting a source text to a target text while maintaining the original text’s tone and style. In this process, the target language should have an equivalent in the source language to ensure the uniformity of the message. However, not all languages are alike, and there are many cases where there are no exact “equivalents” of terms in target languages.
Take the example of English and Japanese language translations. In English, a person can be vague about the gender of a friend, by using the term “they”, without being evasive. On the other hand, it is harder in Japanese, where a person has a choice between terms for mail (otoko) and female (josei). This means that a professional Japanese translation service provider will have a hard time finding the exact equivalent of otaku and josei in the English language.
The translation is a “creative” task
As mentioned earlier, translation is a difficult task, even for human translators. These human translators are not only expected to deliver accurate translations but they are also expected to be “creative” in the process. The task of translation is creative for two reasons. Firstly, the translators need to provide translations of new or novel terms that are present in the source text. Secondly, translators are to play the roles of cultural mediators, with the responsibility to convey to the readers of the target language the exact meaning of the source language. Without taking the culture into perspective, it will be difficult for the translator to convey the exact meaning of the source term in a precise way.
Owing to this “creative” responsibility, computers cannot deliver the precise and most accurate translation of source text, particularly for subjects that require creative thinking.
Computers cannot “reason” while translating
Although computers are now being used to perform pretty much every task, the case of translating one language to another is different. Computers are basically devices that follow rules–both mechanically and literally. This rule-following can also give rise to creativity, but this is not the same creativity required for precise translation. Coming up with a new term during translation can be more of inventing a rule. Although it is true that cultural mediation requires extremely sophisticated reasoning, one must be conscious about what meaning a potential reader will get.
Machine translation is not enough
The use of machine translation is a different story in the translation process. Since the markets are facing a shortage of human translators, and machine translation, companies may adopt machine translation.
Although many people view machine translation as an obvious solution for the shortage of human translators, it may seem perfect for speeding up the localization process. This happens when a Japanese translator wants certified Japanese translation quickly. The source language is fed into a computer and the target language emerges on the other end moments later.
However, one cannot deny that despite having these amazing translation benefits, a machine is a machine and would ultimately give out robotic-sounding sentences. Machine translation may be fully capable of handling complex technical documents and providing Japanese-to English-document translation, but it cannot add the “creative” element to the translation.
Machine translation is often programmed with collections of written texts (called corpora) which train algorithms to seek the “best possible” translation for a given term. The previously translated texts may leave undefined patterns and these patterns may be generated during translation. This means that the MT software you are using to handle English-to-Japanese translation will contain a corpus of similar translations. It will become easier to translate complex terminology related to Japanese or the English language, based on these corpora. This also means that MT relies on these corpora for delivering accurate translations.
The Case of Japanese-to-English Translation
The English language comes second in the most spoken languages in the world. However, this translation of Japanese to English content cannot be carried out with computers or machine translation. Japanese is a minor language that requires careful precision and accuracy when translated into other languages. The translation of Japanese text is difficult owing to the lack of open-source human translation for the corpus. In addition, the Japanese language is a complicated one and there are exceptions to grammatical rules. This makes it almost impossible to give these machines large database for a “perfect” translation.
The Japanese language is a high-context language, which means that the defining information is often not clear. Language, a number of objects, and gender are left out when producing the language. A simple sentence like “someone ate my pudding” in Japanese could be translated to “my pudding was eaten”.
The idiomatic phrases, slang, innuendos, expressions, and regional dialects present an even greater exception to grammatical rules. This means that is a difficult task to make computers learn and handle these properly. For example, take the Japanese phrase (shikatat ga nai). In English, it basically means “it could not be helped”, but depending on the context, it could be translated to “tough luck” or “it was bound to happen”.
What to Do for Japanese translations?
Some languages like Japanese may be too tough for a computer to handle. This is the reason why translators have come up with MTPE (machine translation post-editing). This MTPE denotes a situation where a machine translates a language and a human translator checks and tweaks the final result. Furthermore, some companies claim that MTPE improves the productivity of human translators, can increase speed, and can also contribute to bringing down costs. On the other hand, translators are of the opinion that it is best to handle complex languages like Japanese through human translation, as there are much fewer chances of errors with human translation.
Key Takeaway
The Japanese language is a complex one, requiring careful scrutiny and precision during translation. In an attempt to make Japanese translation easier, computers can provide the translation. However, a computer may fail to accurately deliver the correct translation in the Japanese language due to some reasons. This is the reason why translators recommend handling such languages themselves or adopting the MTPE approach after machine translation has been performed.
简体中文