Call us on 020 3286 4122
April 13, 2016
In January 1954, IBM’s 701 Computer also known as the ‘brain’ translated a number of Russian sentences into English. From this point on, the researchers claimed that human translations would now be a thing learnt about in the history books, and after such a great achievement who could blame them for such a point of view. Six decades later and it seems that history is repeating itself with the MIT Technology Review proclaiming that to translate one language into another you only need to find the linear transformation that maps one to another, something that engineers at Google have been implementing with their Google Translate tool.
There is no doubting that simply plugging text into Google Translate and less than a second later receiving a translation is an incredible innovation in the field of translation. Journalists reported that this heralded a new era for language translation and a groundbreaking milestone in technology that has been around for 60 years. The belief that the translation problem, if there ever was one, has now been solved due to the technological advancements and computer translations is one that has holes in its very foundation. Machine or computer-generated translations do serve many purposes and used properly in specific and limited cases, they can be useful for a number of tasks unrelated to the final translation produced. However, machine translations alone will remain as second best to human translations for a very long time, and below are the reasons why.
Human translators will always remain a step ahead of the machines in the same way that bilingual laypeople cannot replace professional human translators. The majority of translation jobs require more than just the knowledge of two languages with the idea of one-to-one language equivalencies being false. Translators aren’t human dictionaries; they recreate the language, crafting beautiful sentences and ensuring that they have the same impact and message as the source. In order to do so, translators, draw upon a lifetime’s worth of industry knowledge and experience and combine it with their understanding of the two cultures. Machines, at least for now, cannot do this and even if they could, they would fail at being subjective.
Currently, Google Translate supports 80 languages which is great until you consider that there are between 6000 and 7000 languages that are around today, of which about 2000 are considered endangered. Estimating that there are approximately 1000 languages holding significant economic importance today, this means that Google has yet to develop another 920 languages. Taking this number of 920, and predicting that Google could add 10 languages to its translation service a year, it would be another 92 years until they provided a fraction of the world’s human languages. Even at this low quality level, the machine translation would not provide a service for the majority of the world’s languages, at least not during our lifetime.
A significant drawback of machine translation tools is its inability to understand and learn about the context of a word, not just in one language but across tow. In Spanish, a single world can have hundreds of different meanings depending on the context and where the word is used within the Spanish-speaking world. In this way, professional human translators are able to not think in word for word translation but instead figure out the meaning based on the context and the words interacting around them. With the number of combinations constantly changing and increasing in number, these are only limited by human creativity and therefore the machines would struggle to keep up.
Words are vital to the perception and voice of a business brand with consumers making choices as to whether or not they can relate to the words being used to market and sell a product. Every human being has encountered a different upbringing, adding to the beautiful diversity of likes and dislikes. With human beings finding it near impossible to accurately predict which words won’t annoy or confuse people, how can we expect machines to fare any better.
Computers will never fully resolve the problems facing translators on a daily basis, such as the subjectivity of a translation. Even in the development and working towards replacing humans with machine translations, the computers will still require significant help and input by human translations. In the meantime, whilst Google and other machine translating companies hold on to the dream of one day replacing human translators, the fact of the matter is, that good quality translation will always need the human touch!
Thanks for visiting the Latin Link blog! Contact us today for your free translation quote!