Translation: Use of Technological Tools


 Analyzing the attitude of translation students towards Computer-Aided Translation (CAT) tools, Caner Çetiner, a research scholar on translation training, translation technologies, and machine translation post-editing, in an article in the Journal of Language and Linguistic Studies wrote that the translation profession had witnessed some technological innovations in recent years, primarily because of  ever-increasing workload. Translation tools came to be seen to ease the ever increasing workload pressure, and though there were limited studies available dealing with the acceptance of those tools by translation students, there is undoubtedly a clear opinion that supports the view that students develop a positive attitude after they are taught the benefits of using computer-aided translation tools.

Perhaps there could be no two opinions about the need for technology to evolve and invent new tools to help ease the workload pressure, particularly the drudgery involved into voluminous works, but technology has still to tread many miles before the tools and techniques it has evolved could come anyway near to the professionally acceptable standards. It is early stages for these tools but their blind use could produce target texts no way near the letter and spirit of the source text.

Recently, I came across an audio interview  in Punjabi language, I was asked to translate its Punjabi transcription into English. Any Punjabi reader making a comparison between machine  and manual  transcriptions of the original audio would tell volumes of the whopping gaps in the machine transcription, and obviously when translated into English, it made hilarious reading to say the least. Surely the English readers would be at a loss to miss the letter and spirit of the original audio. The subject – same sex marriage – itself was much talked about if not controversial. Obviously, I am talking about Punjabi-English pair only, and have no information or knowledge about other languages, but I guess, of such a scenario, there would be more similarities than differences. 

"It is  anybody’s guess as to how machine translation will cope with the subtle changes in meaning brought about by changing stress or for that matter facial expressions or gestures of the spoken word"

It is not to undermine the machine transcription and translation. Technological tools are surely helpful in reducing the ever-increasing workload and the drudgery. These are in fact being used contemporarily to varying degrees depending upon the kind of source text and the resourcefulness of the professional translator, but quality of professional workmanship of the end product i.e. the target text is of prime importance. Machines are fairly good and could achieve fairly reasonable and professionally acceptable standards when it comes to simple, even involved but not very long sentences. Machines have a big advantage – they can be fed with an excellent encyclopedic vocabulary - but the dynamics of a language, spoken or written, will still require machines to catch the subtle nuances of the spoken or written word, for the litmus test has to be catching the letter and spirit of the source text into the target text.

I pointed out the flaws of the provided machine transcription to my client - omissions not just of words but of phrases altogether, linguistically incorrect transcription, incomprehensive and/or meaningless transcription, fails the litmus test of accuracy which is the cornerstone of AUSIT/NAATI guidelines, no way near the 'letter and spirit' of the original audio, and if translated as it was, it would have been  difficult, if not impossible, for the translator to translate without compromising on meaning, and the negative impact on English readers by reading shoddy target text. My client rightly decided not to proceed with the translation of the machine transcription.

I am not an expert on machine transcription technology, but I suppose, the technology must be based on phonology of a language to which it must be adding grammar and vocabulary to produce the end result. If sound is the basis it is well understood that technology may not be equipped with the dynamics of grammar.  The chances of target text missing the letter and spirit of the source text could not be ruled but towards the end. In the given example, I observed there was a large scale cutting and chopping even adding as if the machine had some sort of artificial intelligence that could identify and weed out what was superfluous, repetitive and unnecessary but then a professional translator is not an editor.

The litmus test for machine transcription and translation would be the literary works such as Essays of Elia by Charles Lamb known for its rich, erudite and sublime style by using similes’, metaphors, juxtaposing humour and pathos to delineate poetry in prose. Deep philosophical works, like Aurobindo Ghosh’s Life Divine, known for its profundity and sublimity, both in matter as also manner., The sentences run into half a page if not more. By the time one comes to the end of a sentence, one is often likely to forget the beginning.

It is  anybody’s guess as to how machine translation will cope with the subtle changes in meaning brought about by changing stress or for that matter facial expressions or gestures of the spoken word. For example,  “achha” as a stand-alone in Punjabi means ‘okay’ but with a slight change in stress it could also mean “is it?” or ‘is that so!”, from affirmative to interrogative and speculative.

There are many more ways and examples where machines will find it  difficult if not impossible to come  near  professional acceptable levels, but technological evolution is a continuous process. Who knows, it is not very far when we could see technology bringing rich dividends in the T&I field too.


Author Biography: Gurdev Singh Grewal holds a master's degree in English and passed India's Civil Services Examination in 1980. He worked for the Indian Government for eighteen years in middle management positions, across areas including development, aid and currency, followed by fifteen years in senior management positions in the private sector. Gurdev has been visiting Australia and working as a freelance NAATI ­certified Punjabi>English translator since 2013.

Published: 15/05/2019