From a buyer’s point of view, a very common problem with quality is that it might not be visible or evaluable before the actual purchasing transaction is made. In some cases the quality cannot even be seen afterwards. Unfortunately, it is the lack of quality which becomes visible when least expected, and with high costs. Translation quality makes no difference. Translation quality is extremely challenging to evaluate. To be able to say whether the translation is good or not, one needs a lot of knowledge on the used languages and the source material. Because it is so hard to see the quality, buyers (and service providers) often search for help from secondary quality indicators.
Even highly skilled professional translators and interpreters have to rely on secondary quality indicators when communicating with potential customers. People who are looking for help with translation are rarely experts in the field. References, recommendations, diplomas, reputation are examples of quality indicators which refer to the actual translation quality only indirectly. Every translation task is individual and the success in previous assignments doesn’t guarantee or determine the quality of the next job. This applies not only to translators; it is a real challenge in most service businesses.
Although the link between secondary indicators and upcoming quality may be weak, well-selected indicators can enable building trust between the participants, and trust is indeed a key element between translation service buyer and provider. It is said that every translator will provide a different translation for the same piece of text, and that there isn’t any single correct or right way to do it. When the absolute correctness is such a complex concept, secondary quality indicators can be very handy, like many translators already know. However, it would be on everyone’s interest if those indirect quality indicators had only the attention they deserve. Misleading use of those trust building elements is not profitable in the long run. Long term business relationships are always built on real translation quality.
Secondary quality indicators are an advantage for professional translators as compared to machine translation. When talking about machine translation quality there is rarely any secondary elements involved. Hence, the debate between human and machine translation quality is not always completely even.
Free Version Available: Translate PDF Documents with Multilizer!
Let machine translators translate your PDF files automatically and use your energy only in proofreading. Try this kind of simple and convenient translating with Multilizer PDF Translator!
Translation quality is an interesting thing. It can be easy or impossible to evaluate, depending on your own languages skills. When the text is translated to your own native language, poor grammatical quality is fairly easy to see. When the text is translated to a language you don’t have any knowledge of, you cannot say a word about the translation quality. However, to be able to say whether the overall translation quality is good or not you should not only pay attention to the grammar of the translated text but you should also be able to check and understand the original text.
This all means that no one can truly evaluate the translation quality unless he or she has very good language skills. This dilemma makes me wonder the role of those people who buy translations. They may never get to know what they have purchased. The only thing they can rely on is the translator’s word, and trust becomes a prerequisite in any transaction between a translation buyer and provider.
Based on this observation, I would argue that the problem with machine translation is not the actual translation quality. Machine translation suffers from lack of reliability. If we don’t have the required knowledge to be able to proofread the automatic translation, there’s a risk involved in using the translation publicly. But if we aren’t able to trust the machine, it doesn’t matter whether the automatic translation is correct or not.
Machine translation quality is a hot topic. Most people and companies know that machine translation is available but the common opinion is that its quality is very far from perfect. Although machine translation is rarely flawless, one don’t have to settle for the quality. The quality of machine translated text can be improved in many ways. Here are three common methods:
One commonly used method is a technique called post-editing. In post-editing, a professional, human translator edits the translation made by a machine. This is often a bit faster than translating without machine translation. Faster translation means also lower cost. And the quality will be the same as with traditional, professional human translation. However, although the benefits are significant they are not that dramatic.
Another method is customizing machine translation engine to better translate certain type of text. For example, a machine translator might be customizable to translate text related to medical domain. Customizing includes “teaching” the machine translator about vocabulary and phrases that are common in medical domain. This improves the machine translation quality in medical texts. Thus the translation quality is better and the above mentioned post-editing stage might be skipped altogether. And without post-editing the translation process does not involve human work at all. This results in huge benefits in translation cost, time and volume. However, customising machine translation engines is very costly and requires special knowledge.
The third method is to modify the original text to be more machine translation friendly. It is true that machine translation is rarely flawless but currently most advanced machine translators translated simple text quite well. This method is especially useful if the machine translator cannot be taught or customized. Text with complicated sentence structures and words with many meanings is more error-prone than text with simple sentences and common vocabulary.
Have you tried any of these methods yourself?
Did you know that it is possible to translate PDF documents automatically and edit the translation easily with for example MS Word? With Multilizer PDF Translator Standard you can.
Like you probably know, money has been a hot topic in the European politics lately. If someone has been wondering why Finland is so critical against lending more money to Greece without collaterals, here could be the reason.
The Greek prime minister Papandreou has said: “Υπάρχουν χρήματα”. Because I don’t understand Greek I used machine translation to find out what that means. First I translated it with Google Translate to English and found out that it means “there is money”.
According to Google Translate the situation in Greece is not totally hopeless. But just to double-check, I translated the Greek sentence to Finnish with Bing Translator. This time the outcome was less promising: The Finnish translation “ei ole rahaa” means “there is no money” in English.
Apparently Finnish politicians have been using Bing Translator because obviously lending money to someone who is broke is not wise, at least not without collaterals! Although it seems like it, probably these translations are not Google and Microsoft’s official statements about the financial situation in Greece. All this proofs is that there still are some weaknesses in automatic translation, translation quality estimation is important when using a machine translator and one should use machine translation correctly.
Free Demo: Translate PDF documents automatically.
Understand PDF documents written with foreign languages. Quick and automatic PDF Translator translates it to your language. Free demo available!
While machine translation is improving and becoming well-known tool all over the world, there are still strong opposition running against the tide. They argue that publicly used incorrect machine translation is changing languages too much and specially not in a good way. It is true that only professional translators and native speakers are able to evaluate the quality of the automatically translated text, and others can rely only on their own knowledge or on automatic quality check.
The situation is similar to when SMS became the most popular communication tool. Shortened expressions and word combinations were basically ruining the language skills of the young generation – at least according to the negative force. Whatever the circumstances, development and changes can be scary but usually they are inevitable and even useful.
Languages have been, are and will be developing constantly. If you watch a really old movie you probably find the way the characters speak old-fashioned. Likewise old books and other linguistic material clearly represent their time, not ours. Like human population, languages have a history; an ongoing story with past, present and future. For example linguists and scientists agree that irregular verbs are more or less relics from the past. I’m sure that we all are glad that today we need to study difficult irregular verb inflections because some people stopped the development by refusing to renew their language centuries ago.
So, is machine translation ruining our languages? I don’t think so, but we are definitely changing the rules of languages by using new tools like machine translation. Whether the development is good or bad is not relevant. And if needed, machine translated text can always be proofread by a professional, human translator. Right?
What do you think about the relationship between language development and machine translation?