MT: A Positive View

Details

Written by Nadia Sabbah

Three years ago, my brother got married to a lovely Chinese lady. Beside all the joyous emotions, there was a huge concern within the realm of all events: Language! My new sister-in-law, Mai, can only speak Mandarin; a major Chinese dialect. In this day and age, I thought, this should be no problem.  What’s the good trusty Google Translate for if, not in such circumstances! Then Mai asked for my personal advice on weight reduction. I immediately resorted to the instant translation engine online and inserted the following text; “Dear Mai, you need to avoid carbohydrates and sugary diet. Use a lot of fruit and vegetable intake.” The target text in simple Chinese was : “亲爱的麦，你需要避免碳水化合物，水果和蔬菜摄入含糖很多.” Prompted by my intuition, and past bad incidents, I decided to reverse the translated text, and see what the English output would be, before I sent my dear sister-in-law any advice! The Chinese-to-English translated text came out as: “Dear Jimmy, you need to avoid carbohydrates, fruit and vegetable intake a lot of sugary diet use.” I believe it goes without saying that I turned to my brother to carry out the authentic translation job!

Reading my introductory anecdote, one might think I am about to list all limitations of online machine translation systems. I, however, intend to do just the opposite! There is no argument that online machine translation systems have their drawbacks. They are prone to errors of all kinds, ranging from funny to even fatal mistakes! Nevertheless, I am rather inclined to discuss the ‘glass half full’ side of the equation and bring up some success stories in using machine translation (MT) worldwide.

Machine translation came a long way since its inception in the early 1950’s. Back then, the innovative instant translating systems fired up all fantasies of science fiction novels and films. People would be speaking their native language while traveling through space, having instant lingual renditions to all species despite all barriers. Such wild expectations exceeded the rather slow-paced-developing machines. Yet, there are certain factors that can contribute remarkably to the betterment of such systems. One very important factor is the use of a control language; known in the field of computational linguistics as “sublanguage”. Texts of more technical genre can attain better results when such instant translators are used. A most common example is the Canadian Météo Machine Translation System. It was basically designed for translating daily weather forecasts. The nature of such redundant technical texts was so boring for human translators that many had to leave their job after only a couple of months. From 1981 until 2001, Météo was automatically completing all necessary translations without a need for post-human editing. In the province of Quebec, the bi-directional translation was between English and French. “Since then, a competitor program has replaced Météo system after an open governmental bid. The system was often mentionedas one of the few success stories in the field of machine translation.” (Wikipedia, 2014).

Frequency and consistency in term-use of any subject matter can lead to satisfactory results when utilizing a machine translation system. Another example of successful machine-translation systems is METAL. It is used at a number of European companies. The European Union also used the Systran very well-known system for its translation services. (Volk, 2008) It is a different version from the one that is free online. “The Pan American Health Organization is another international body that has long used MT for publishing purposes.” (Munday, 2009)

Film subtitling is another area that proved certain machine translation systems can contribute greatly to facilitate the translator’s job; consuming less of the human translator’s time.

In the Arab world, the interest in such systems grows rapidly by the day. The common conception, however, revolves around the deficiencies of free online engines to deliver decent translated versions of a myriad of contexts. Yet, many are unaware of the breakthroughs in the area of machine translation. Misconceptions prevail due to many incompetent free online instant translator systems. Not many can relate to the fact that MT systems can be upgraded to a satisfactory level of performance. In her MA thesis, Fatin Al Mutawa, clarifies that machine translation awareness and utilization is moving up hill. But SAKHR is the only company in the Arab world with real interest in MT and its progress. It opts for rapid and accurate translation. (Al Mutawa ,2012)

In their study on the subject, Zughoul & Abu-Alshaar indicate how MT systems are receiving further interest sparked by globalization. Countries keen on political and security issues, the assistance of the deaf and blind and many other concerns, are paying more attention and taking wider strides in adopting and developing competent MT systems. The US, for instance, is highly concerned about machine systems that can cover some military and political contexts accurately when many languages are involved, most importantly among which are Arabic and Pashtu. (Zughoul & Abu-Alshaar , 2005)

Tarjim, Al-Misbar, Al Wafi and Ajeeb are some online sites for the bidirectional Arabic and English languages. Other systems that are still under constant attention and advancement are: Apptek, Al Mutarjim Al Arabi (by ATA), المترجم العربي, and a multilingual system: Al Naqil Al Arabi النّاقل العربيّ. The latter systems are not free online systems. (الحطّاب ,2008)

Advancements of MT have undoubtedly been gaining greater attentiveness in China. Qun Liu believes “the machine translation researches and applications will have a better prospect in the future in China, along with the growing of the Chinese economy and the exchanges between China and the world.”(Liu, 2011) In Beijing 2008 Olympic Games, machine translation technologies were also used to provide automatic or computer-aided translation services.

On the official website of the International Association of Professional Translators and Interpreters, researches and forums come out usually with the outcome that “Free or commercially available machine translation (MT) software cannot be expected to render with accuracy the meaning of a foreign-language text, and thus MT systems cannot guarantee that their output is suitable for any purpose other than obtaining the gist of a text.” (IAPTI) Nevertheless, computerized instant translators are getting more advanced and becoming more accurate and cost-effective by the minute, for quite some time now! Whether free online, or purchased versions, organizations are keen on developing such systems to stand against the wind of competition. The giant engine: Google has been working relentlessly on upgrading and increasing its online services, most important of which is Google Translate. The frequency of erroneous renditions is noticeably decreasing.

 Translation is not a matter that can be taken for granted. Translation studies became a major academic discipline, according to Susan Bassnett, a prominent professor in the field of Translation Studies, around the 1970’s. “Indeed, after a period in which research in computer translation seemed to have foundered, the importance of the relationship between translation and the new technology has risen to prominence and shows every sign of becoming even more important in the future.” (Bassnett, 2002). Machine translation is a major area in Translation Studies. Forums all over the world are held annually or biannually to discuss all updates in the area in question. The non-profit European Association for Machine Translation (EAMT) hosts a yearly conference to present all issues pertaining to the field of machine translation. The Association for Machine Translation in the Americas (AMAT) organizes “conferences every other year to create a forum where people can exchange experiences and ideas to improve these technologies and use them more effectively.” (AMAT, 2014)

Pre-editing and post-editing intervention by translators to check work produced by machines is still required if the output is to be of publishable quality. Electronic tools are rapidly progressing along with all other technological advancements. Human professional translators should always be on their tiptoes, for the market has never been more challenging than it is nowadays.

Works Cited

Al Mutawa, Fatin. MACHINE TRANSLATION IN SAUDI ARABIA. Thesis. American University of Sharjah, 2012. Sharhjah: n.p., 2012. Web. 12 Oct. 2014.

1. AMTA. N.p., 2014. Web. 19 Oct. 2014. <http://www.amtaweb.org/>.

Bassnett, Susan. "PREFACE TO THE THIRD EDITION." Translation Studies. Third ed. London: Routledge, 2002. 15. Print.

LIU, Qun. Machine Translation in China. N.p.: n.p., 16 Nov. 2011. PDF.

Munday, Jeremy.The Routledge Companion to Translation Studies.London,2009. Print.

"METEO System." Wikipedia. Wikimedia Foundation, 16 Oct. 2014. Web. 19 Oct. 2014.

Zughoul, Muhammad, and Awatef Abu-Alshaar. English/Arabic/English Machine Translation: A Historical Perspective. Erudit. Les Presses De L'Université De Montréal, 2005. Web. 14 Oct. 2014.

وحلولالحطّاب, مأمون. "التّرجمة الآليّة للُّغة العربيّة/ قضايا وحلول." الترجمة الآلية للغة العربية-قضايا    . Jordan, عمان-الأردن. 2008. مجمع اللغة العربية الأردني. Web. 19 Oct. 2014.