It has to be acknowledged that automatic translators are all the rage. Between the new developments in Google Translate to improve its results and Skype’s simultaneous translator, it looks like machines will soon take over the world of translation.
However the rise of automatic translators means that some translations have been made on the basis of the results of these programs and applications, leading to a final work that did not have the expected quality.
As we showed you in our latest infographic, human translation always adds quality. This quality is Okodia’s raison d’être.
- How does an automatic translator work?
- Mistakes of automatic translations
- Google Translator, does it help?
- Pilot, the automatic translator
How does an automatic translator work?
In order to understand the mistakes they make, you’ve got to understand how automatic translators work. There are two main types of automatic translation:
- Rule-based automatic translation: this is based on a vast number of linguistic rules and millions of bilingual dictionaries per language. The program parses the text and creates a representation from which it generates the text in the translated language. These translations are built on the basis of huge dictionaries and sophisticated linguistic rules.
- Statistics-based automatic translation: this is a paradigm in automatic translation, where translations are generated on the basis of statistical and information theory models whose parameters are obtained from the analysis of huge corpora of bilingual texts.
As you can see, these processes are very elaborate, automated, and very far (for the time being) from the quality and perfection of human translation.
Google Translate purports to know more than 80 languages. It uses a statistics-based approach and gathers information from translated texts in its database. Using these, it fits the pieces of the puzzle together. But, as is often the case, statistics is useful for many things, but certainly not with regard to languages. Be careful and don’t fall into the temptation of using Google Translate—the consequences can be very severe.
Mistakes of automatic translations
- Literal translations: all languages are enriched by phrases and double meanings. Professional translators know and translate them, or replace them by phrases with the same nuances in the target language. However, an automatic translator cannot recognise double meanings, so it will always translate that phrase literally. Automatic translation tools have no sensitivity, nor imagination.
- It does not recognise typos: if there is any mistake in the text to be translated, the automatic translation cannot recognise it and will not translate it. A professional, however, will notice the mistake and correctly translate the term or phrase. In an advert for a forthcoming a concert of a well-known Spanish singer, whose first name had been translated into Catalan as It seems that, in the original text the name was misspelled and appears (Alejando, instead of Alejandro) and this gerund –“ando” in Spanish is the same as “ing” in English– hence, became the star of the show
- It leaves some parts untranslated: sometimes automatic translators leave part of a sentence untranslated, such as articles, conjunctions, etc., which means that the meaning of the sentence is lost. A professional translator will translate everything, conveying the suitable meaning in the target language.
- It does not translate acronyms: many acronyms, such as UN, are not translated and are spelt as such, while a human translator will translated them into the language in question.
- Translating names that shouldn’t be translated: professional translators would never translate Arthur More as Arthur Más – to remind us of one of the first linguistic blunders that automatic translation presented us with–, and please be aware that proper names of people and place in the terrific soap operas, with a few exceptions, do not need translating either.
- Translation inconsistencies: due to the literal nature of automatic translators, a number of inconsistencies arise that render a translated text incomprehensible. The translation of prepositions is one of the obvious signs that the text has been handled by the Frankenstein of translation. The translation of sepia a la plancha as “sepia to the iron” is sadly notorious. Iron, in the same nominal phrase would also be another example of its prowess.
- Nuances, metaphors, accuracy: these elements of texts can only be detected by professional translators; an automatic translator will be unable to translate them and convey the right meaning.
- Words with more than one meaning: when a word can have several different meanings, the automatic translator can make mistakes when deciding which meaning to use. A very famous example was the translation of “casco histórico”, which means old town or historic centre, as “historic helmet” in the FITUR stand for Santander several years ago (since “casco” can also be translated as helmet).
Google Translate, does it help?
We cannot deny that applications with automatic translation can be very useful in certain moments. For example, when travelling overseas it is good to have an application on your mobile that allows you to order a sandwich or find a hotel to sleep in without dying in the attempt. If you have had a product that you have purchased online, the translator can help you to communicate with the seller, as long as the subject matter is not very complicated. However, this does not get away from the fact that for professional translations you need a professional translator. Remember that it is an investment that will create an unbeatable first impression with your customers.
Another option: Pilot, the automatic translator ‘makes you stupid’
Pilot has been a very important revolution. There are people who want to remove the need for professional interpreters and translators, and don’t know how to do it. So much so that they have had to resort to technology. But be careful, because technology ‘makes you stupid’, according to several studies. The day you start using Google or a calculator, a tiny part of your brain dies. The same thing happens with Pilot, the automatic translator that you place in your ear, and let it take care of all the ‘dirty work.’ Without mentioning other complications, in that technology is designed to fail, however automatic the translator may be it, will never be able to translate text with the perfection, nor the skill and mastery of a professional, who is also specialised in technical translation, webpage translation, scientific translation…
Can you imagine holding conversation with a mathematician that has deciphered Hilbert’s unsolved problems? Do you really think an automatic translator will have such a high level of specialisation? It’s likely to turn into a conversation at crossed purposes, where the automatic translator neither puts things into context nor knows the meaning of certain terms or worst of all, doesn’t use them correctly.
In short, our strong opinion is that human translation has a quality with which automatic translation cannot compete, as shown in our infographic:
As our elders usually say, there’s no such thing as a free lunch and using a free automatic translator can end up costing a fortune. Are you sure that you are really making a saving?
Related links
The worst mistakes in the history of Google Translator (or some of them)
