Machine Translation

With the development of AI and its use in various areas, machine translation has been one of the topics at the center of interest. There have been a lot of debates over whether it can replace “human” translators. Even though lots of advanced machine translation tools emerged over time, it still seems far from taking over the job of translators. In this article, I will touch upon the uses of machine translation and why it seems unlikely to take over the jobs of actual translators.

Machine translation or MT, is defined as the translation done by suitable computer software, analyzing a text in the source language and reproducing it in the target language. This is often called raw translation. It also should be distinguished from other translation technologies, including computer-aided tools such as translation memory. While machine translation has no human intervention, computer-aided tools are used by professional translators to help them, aiming to increase their efficiency. The raw translation is distinct from the works done by professionals. While this does not necessarily mean that MT has no use in many areas, one should be aware of its specific applications in specific contexts where it can be effectively used.

There are various challenges and limitations MT faces, and one is ambiguity. Due to the ambiguity (whether it is lexical or grammatical) and complex structure of languages, a text is open to different interpretations. When MT attempts to translate a text, meaning might be compromised or there might be deviations. To produce an appropriate translation, MT is often required to choose between the possible interpretations. Unlike people who can use their knowledge and identify the context, MT relies on a certain amount of data to resolve ambiguity and process information. Therefore, if it fails to correctly determine the context, MT might not deliver suitable translations. Another obstacle is the relationship between content and form. Meaning can be enunciated in various ways, and languages often use different structures to convey the same meaning.

Despite the issues listed above, machine translation has two main uses: Assimilation and dissemination. Assimilation refers to the use of machine translation when a person does not understand the source language in which a text is written, allowing them to have a general idea of the content of the text. Translation errors are often neglected provided that the MT system delivers a text that is coherent enough (in terms of meaning) to be understood. Dissemination, on the other hand, refers to the use of machine translation of texts which will be edited afterward before publishing. Post-editing of raw translation is revised and done by a skilled translator.