There used to be a favourite anecdote among translators. Supposedly a true story, it went like this: A supercomputer capable of translating the world’s major languages is to be introduced to the public. To showcase its benefits, a senior executive asks the computer to translate a line from Scripture into Chinese. It is Christ’s observation “The spirit indeed is willing, but the flesh is weak,” and the computer does as it is told, except no one present speaks enough Chinese to gauge the quality of the translation. So the computer is asked to translate the Chinese back into English, and it returns the words “The whiskey is strong, but the meat is rotten.” Causing translators no end of mirth, the story suggests that no matter how cleverly constructed they may be, machines will never seriously compete with humans in mastering the complexity of human language. Yet just as the Trojans ridiculed Cassandra’s warning about the wooden horse not be- cause it was so absurd but perhaps because they were simply not sure and their situation so precarious, the anecdote provides comic relief to the translators’ fear that the computer will one day have done its homework and become as good as them.
On Sustainability in the Translating Business
The promise of computer translating is arguably one of sustainability. If you assume that language functions like a cybernetic organism and follows set rules of grammar, semantics and practice, then artificial intelligence following a similar logic would seem to be up to the task. Ideally, the machine would not only match words and phrases in one language with those of another language, but would do it better, faster and always in the same quality. By combining databases of empirical knowledge with methodology and systematicity it would make translating “sustainable” in an age celebrating sustainability as the answer to most major issues.
This is not to say that a translation can actually be “sustainable.” Yet obviously, the idea underlying machine translations is sustainability in the general sense of “sustaining the status quo.” Any customer will expect a translator to keep delivering translations of unchanging quality, consistent in terminology, register and style. Here, the challenge of sustainability narrows down to quality assurance, and it is tempting to think that a computer-assisted solution would deliver the necessary consistency in demonstrable ways. With mounting processor power, you might even teach the computer’s unfailing memory that “spirit” is not the same as “spirits.” Yet while perfect machine translations may be just around the corner, chances are that they will never take the lead in professional translating, but will remain a sideshow playing to search engines and tourist-guiding gadgets.
Does this mean that translating software is dead in the water? It does not. On the contrary, it is the up and coming thing. Not the kind of thing that would have a machine in the corner do the whole job and substitute whiskey for spirit. Rather, the sustainability issue is being addressed through a combination of computer, databases and – yes – brain. True to the holistic concept of sustainability, professional translation software integrates electronic assets, personal expertise, and client style guides in a unified approach to quality assurance. The computer’s part is to ana- lyse the source text, to identify identical sentences previously translated, and to deduct these matches from the quote. Aside from ensuring the consistency of your translation, it is a good way to reward returning customers with cash savings. Other value-added aspects include glossary, concordance, and import/export features as well as resource sharing and remote desktops. As this trend and the ongoing software development mutually boost each other, translations have become better, less costly, and more sustainable. So, where does the brain come in? For one thing, it handles the actual translating, because that is not what this type of software is designed to do. Also, it would recognise a Biblical quote, and research the existing translation in the target language rather than translate it. And finally, the mind, being the very source of human discourse, is less vulnerable to its stylistic booby traps. It understands allegory, irony, and metameanings, or at least more so than the computer.
The other day, a well-meaning client sent a “pretranslated” source text he had run through an automated translation program to make my life easier. It did not. In fact, we could both have gotten into trouble if I had let it slip, because the computer-translated manual advised buyers of a memory stick to put the device “into a free harbour.” The program had mindlessly chosen the wrong meaning of the homonym “port.” Perhaps the software could be reprogrammed to identify the proper context, but suppose it was to translate a manual for a marine application that uses two different meanings of “port”? What if it advised seafarers not to put the USB-stick into a free port on the port side while in port drinking port? This may be nonsense, but it is easily understood. Even a child would see the humour in it. And we have yet to consider truly complex matter, such as poetry, philosophy or your next statement of account.
Until further notice, your best bet for sustainable translations remains the proven recipe of competence, close cooperation, and long-term business relations – complemented by quality assurance standards and all the clever things a computer will do under the care of a bright mind.
This is not to say that a translation can actually be “sustainable.” Yet obviously, the idea underlying machine translations is sustainability in the general sense of “sustaining the status quo.” Any customer will expect a translator to keep delivering translations of unchanging quality, consistent in terminology, register and style. Here, the challenge of sustainability narrows down to quality assurance, and it is tempting to think that a computer-assisted solution would deliver the necessary consistency in demonstrable ways. With mounting processor power, you might even teach the computer’s unfailing memory that “spirit” is not the same as “spirits.” Yet while perfect machine translations may be just around the corner, chances are that they will never take the lead in professional translating, but will remain a sideshow playing to search engines and tourist-guiding gadgets.
Does this mean that translating software is dead in the water? It does not. On the contrary, it is the up and coming thing. Not the kind of thing that would have a machine in the corner do the whole job and substitute whiskey for spirit. Rather, the sustainability issue is being addressed through a combination of computer, databases and – yes – brain. True to the holistic concept of sustainability, professional translation software integrates electronic assets, personal expertise, and client style guides in a unified approach to quality assurance. The computer’s part is to ana- lyse the source text, to identify identical sentences previously translated, and to deduct these matches from the quote. Aside from ensuring the consistency of your translation, it is a good way to reward returning customers with cash savings. Other value-added aspects include glossary, concordance, and import/export features as well as resource sharing and remote desktops. As this trend and the ongoing software development mutually boost each other, translations have become better, less costly, and more sustainable. So, where does the brain come in? For one thing, it handles the actual translating, because that is not what this type of software is designed to do. Also, it would recognise a Biblical quote, and research the existing translation in the target language rather than translate it. And finally, the mind, being the very source of human discourse, is less vulnerable to its stylistic booby traps. It understands allegory, irony, and metameanings, or at least more so than the computer.
The other day, a well-meaning client sent a “pretranslated” source text he had run through an automated translation program to make my life easier. It did not. In fact, we could both have gotten into trouble if I had let it slip, because the computer-translated manual advised buyers of a memory stick to put the device “into a free harbour.” The program had mindlessly chosen the wrong meaning of the homonym “port.” Perhaps the software could be reprogrammed to identify the proper context, but suppose it was to translate a manual for a marine application that uses two different meanings of “port”? What if it advised seafarers not to put the USB-stick into a free port on the port side while in port drinking port? This may be nonsense, but it is easily understood. Even a child would see the humour in it. And we have yet to consider truly complex matter, such as poetry, philosophy or your next statement of account.
Until further notice, your best bet for sustainable translations remains the proven recipe of competence, close cooperation, and long-term business relations – complemented by quality assurance standards and all the clever things a computer will do under the care of a bright mind.
RSS-Feed
