You've got questions.
We've got solutions.


Google Translate Steps up Its Game – Is It Enough?

While Google is no stranger to the translation application game, its newest updates help ensure that it remains at the top of the list for offered features and usability. Google Translate was updated very recently with two new features: it translates street signs and other images just by pointing your phone’s camera at the image (this feature is even available without a mobile data connection), and its conversation mode was updated to include automatic language detection, improving the flow of conversation between the individuals using it.

These features are not completely new to the application, but both have made significant improvements. In previous versions, the image translation software needed an Internet/data connection, and required the user to upload an actual photo of the image or sign. Google is now using Word Lens technology to instantly translate it just by pointing your camera at the image, no Internet connection or photo storage required. This feature is available in seven languages, but Google hopes to expand this list soon. Currently available: English ↔ French, German, Italian, Portuguese, Russian, and Spanish. For users who do not have an international data plan, this is especially convenient, allowing the user to retrieve the translations in real time without running up data roaming charges on the phone bill. Google Translate’s conversation feature is also improving. Previous versions required the users to manually select the language between each use. However, this version automatically detects the spoken language during the conversation, which improves the fluidity of the conversation. Participants are now able to speak more freely with limited interruptions/delays from the application. This feature is currently available in 38 languages, with plans to add more. While Google Translate’s updates are certainly impressive, don’t expect it to replace human translation anytime soon. Reviewers agree that the application is certainly moving in the right direction, but it is not without error. Conor Dougherty with the New York Times agrees that it works fairly well, but that having to transfer the phone between two people speaking can be somewhat awkward, and says that it works best when the speakers use short, jargon-free sentences. Paul Reigler from the Frequent Business Traveler mentions that the app is easy to use and got the main message across, but indicated that the translations themselves were “reasonably accurate, although often not something a native speaker would have said”. While it seems as though Google’s updates are taking on-the-fly machine translation to the next level with its Word Lens technology and improved conversation features, professional translations should still be done by professional, human translators. While travelling abroad, however, Google Translate may certainly help you get around the city by translating maps and road signs and assisting you with asking a local for directions. It certainly seems worth the download, but take each automatic translation with a grain of salt!

Common Misconceptions about Translation

You may have heard the phrase, “Translation is an art.” It may have even crossed your mind to wonder just how in the world translation could be an art. Isn’t it just replacing words in one language with words in another? Simple, right? And hardly an art form. The fact is, translation is serious art, and can have definite ramifications if done poorly. When considering translation, bear in mind the following misconceptions:

Being bilingual makes you a good translator

The idea that someone who speaks two languages would make a good translator makes sense at first blush. However, being able to communicate in a second language does not necessarily imply that you have the skillset to translate into or from that language. A skilled translator possesses a great command of both his/her native tongue and the foreign language, picking up on subtle nuances and cultural subtleties throughout. He or she should also be a skilled writer, having a strong grasp of the rules of both languages.

Translation is just replacing text word for word in another language

Translation is almost never so literal as to replace text in a word-for-word fashion. Every language has different grammatical rules and sentence structures. If we try and translate something word for word, it will likely lose its intended meaning, and often results in something quite comical! A translator will have a good grasp of these rules in both languages to ensure that it is well-written and not overly literal. He or she will also recognize local idioms and their equivalents. While “Break a leg” is well-known to mean “Good luck” in English, a literal translation into another language could imply that you wish someone harm! A skilled translator will be able to take the local idiom and convey the same meaning in the target language.

Machine translation can replace human translators

While many computer programs are able to translate on a basic level, no program is without error. A human translator is able to distinguish between words and phrases with multiple possible meanings, to take on the complexities of languages that are ever changing, and to ultimately ensure that the meaning of a text is conveyed in its proper context. While a good machine translation can provide a rough idea of what the text is about, it cannot replace a human translator when it comes to needing a quality translation.

Translation is quick and easy, so turnaround times should be short

Depending on the nature of the text and the language into which it is being translated, an average translator is able to handle about 2,500 words per day. For a quality translation, you should also factor in time for proofreading and/or revision. If your document is 5,000 words, for example, it is not realistic to expect a quality translation to be returned the next morning, unless you agree to have a team of translators work on your project. However, extra steps should be taken to ensure consistency and quality.  

Do you have a misconception you would like to add or a question about the translation process? Comment below or visit our website to contact us!

What Machine Translation Cannot Do

Machine translation (MT) can be a handy tool when used properly. The key is understanding when to use it, but also understanding its limitations. To know when to use MT, it is important to know what it is. In its simplest terms, MT occurs when a computer translates a text without a human translator’s involvement. Current translation software tools often have a form of MT embedded within them, so many translators do use this functionality to begin their translation work, however it is merely a tool and should not be treated as a solution.

MT can do various things. It has the ability to scan common grammatical rules and even specialized dictionaries related to the text type. Another type of MT utilizes corpora from numerous texts already translated and available. It basically pulls from these documents and tries to detect patterns associated with the various sentences in that particular text. Once it detects these patterns, it “guesses” what the most accurate translation should be.

You might ask, “If such an intelligent software exists, why hire a translator at all? Doesn’t the software do all the work?” Quite the opposite. Machine translation is a great starting point when used correctly. However, every language has its own nuances and can be ambiguous in nature at both a syntactic and lexical level. The use of MT increases the risk of missing the nuances and ambiguity. A human translator is fluent in both languages and therefore understands the importance of maintaining grammatical and lexical accuracy while still ensuring no meaning is lost during the translation. While MT can extremely helpful during the translation process, using it alone is not a solution and often results in a subpar translation, if not a complete mistranslation, at times.

One of the benefits of MT software is that when a translator uses MT in the translation process, it often leads to faster delivery. Since some of the segments not already translated by the translator’s translation memory (TM) can be pre-translated by the MT software, it allows the translator to review these segments and make changes as necessary, speeding up the process. The result is that your text can be delivered back to you quicker than if the translator were working without the additional tools, and you can rest assured that the translation is accurate and conveys the intended meaning of your text.

The automatic translation of a web page is a great example to use when displaying both the strengths and weaknesses of MT. If you’ve ever read an article online that was originally written in another language, your browser may have suggested that you translate it to English. Google Chrome, for example, uses Google Translate to do this. Google Translate uses a version of MT to automatically translate the web page to English so that you are able to read it. Most browsers have a similar function, though they may use different translation software to achieve the end product.

While some sentences may be surprisingly accurate, one can easily notice that web pages translated with these tools alone often display poor grammar and sentence structures. Machine translation can convey the main idea of an article, but it may not deliver it with precision. In a recent Chinese news article, for example, a translated sentence reads, “Xi Jinping run the test area for great concern, repeatedly giving directions and instructions.” Combining the sentence with the others around it, we are able to get the gist of its meaning through context and inference. However, we can also see where the translation contains poor grammar and word choice with “Xi Jinping run the test area for great concern.”

While MT can help provide you with an idea of what is being said, it’s important to remember what it cannot do for you:

  • MT cannot proofread the web page to make sure the translation is accurate.
  • It cannot always accurately choose between two meanings when the original word has a homonym.
  • It may not recognize set phrases and sayings, resulting in a translation that could be too literal.
  • It will not recognize many nuances of the original language, often losing some of the meaning in the process.

So, while reading through a translated web page, it would be best to fact check something that seems odd or out of place; it may be an inaccurately translated sentence. And if your company is looking to translate a text, we highly recommend hiring a professional translator or translation agency to ensure a correct and precise meaning is conveyed at all times. After all, your brand is on the line.