Almost no other issue divides opinions in the translation industry as fiercely as machine translation. Will it become prevalent – and put human translators out of work? But how can it, when the translation quality is “so terrible”? At the end of the day, maybe it’s not a matter of “human or machine,” but of “human and machine.” Is all the fuss over nothing?
What are the supporters of each “system” actually fighting about?
For starters, the advantages of a machine translation over one by a human translator are very attractive for customers:
- It’s cheap: Machine translation costs much less than employing a human translator. If publicly accessible translation machines like Google Translate are used, they don’t cost anything at all.
- It’s available immediately: If there’s no time to wait for someone to translate the text, machine translation is hard to beat.
- It gets a lot done quickly: Translating huge amounts of data in a short amount of time takes multiple translators – or a machine.
But the limits of machine translation are just as evident as its advantages:
- A machine doesn’t translate idiomatically: As a result, machine translations generally don’t flow well and have a strange syntax.
- A machine can’t read between the lines: Humor, irony or metaphors are lost on computers. They only translate the words in their memories.
- A machine can only do what it has learned: Whether rule-based or statistical – a machine can only produce content that was fed into it at some point in time.
While these limits of machine translation are deal-breakers for literature, marketing or technical documentation, the role they play in less permanent environments with lower quality requirements, such as discussion sites or help forums, should not be underestimated.
In other words, the question “machine or human translation?” doesn’t generally come up. Instead, the question should be: “Machine translation without postediting or no translation at all?” And for translation projects with higher quality requirements, larger budgets and longer deadlines: “Human translation or machine translation plus postediting?“
The fact is, machine translation always needs some postediting. And why shouldn’t it? If a translation can be produced at virtually no cost, you’ve already saved a significant part of the total budget. Having saved that much, it’s always a good idea to invest in a more or less in-depth revision of the translation.
Which brings us back to the question of whether computer translation programs threaten translators’ jobs. The answer is no, quite the contrary! At the same time, the future of translators is no longer just translation, but rather postediting machine translations. This opens up a whole new line of work that calls for new abilities and qualifications. Translators who stay on the ball won’t have to worry about their jobs in the future. After all, the amount of text that needs to be localized in the globalized world is so large that, even if computer translation programs are used on a large scale, there will be plenty of work for humans and machines.
Comparing human translations to machine translations is like comparing apples to oranges. There are reasons to use each solution, and if they’re used appropriately, they almost never come into conflict with each other – even if new findings in computer linguistics can further improve the quality of machine translations. The day when a computer can replace a human translator is still very far off. It will probably never come. On the other hand, the day when today’s translators can be a useful supplement to computer translation programs is already here. This is an opportunity no one in the translation industry should miss.
ABOUT THE AUTHOR
As Quality Manager at text&form, Katrin Marheinecke focuses on enhancing the quality of translations and continuously improving translation processes. For her, observing the latest trends in the translation landscape is both business and pleasure.
Read more from Katrin: 5 Reasons why you should outsource translation projects