Machine Translation: A Threat to Human Translation?
It is widely believed that information technologies do wonders, and the translation business will rely on them so much that soon live translators will no longer be in demand.
Is this true? Well, the answer largely depends on what clients expect from the translation and how translators feel about the machine translation. According to the Wikipedia website, the machine translation «is a sub-field of computational linguistics that investigates the use of computer software to translate text or speech from one natural language to another». Therefore, the machine translation is an investigation of «how to become equal to human translation».
The Wikipedia website defines two phases of the translation process:
1. Decoding the meaning of the source text, and
2. Re-encoding this meaning in the target language.
In the translation process, a computer as opposed to a human translator is unable in many cases to cope with any of these stages. Decoding as part of the translation process means obtaining the «right» meaning in a particular context, which may be incomprehensible for the machine translation. Re-encoding means the right interpretation in the target language of the meaning obtained, which is selecting the «right» words to properly interpret the original ideas of the source text. However, a machine translation can pass through both stages only straightforwardly. When it comes to imagery interpretation (the «right» decoding and re-encoding), the machine translation is worse than the translation done by a mediocre translator.
That is why a machine translation, no matter how sophisticated, will never be able to compete with a human translation.
However, a machine may be able to provide a quick «rough» translation which may be edited by a professional human translator keen on the subject. Therefore, a machine translation can help to speed up the translation process.
Nick Bobroff
Member of the Russian Translators Union