Actually, they are not. Watching a translation engine at work, reconsidering translation choices within split seconds as more information becomes available, is impressive. If only I could think that fast! Why is machine translation viewed in a negative light by linguists?
To take the example of technical translation services, machines can think fast but they can’t think like humans, that’s why they make such bad translators. Human communication isn’t just governed by syntactic and semantic principles but also by empathy and intuition. Individual language users have quirks which a translation engine may not recognise. After all, it has only been programmed to process input by matching it against a database of words, collocations and standard expressions stored in its memory.
Professional technical translators don’t just consider the text in front of them but also the conditions of its production and reception. Why was it written, and why is it being translated? What kind of audience is it intended for? With technical translation services, this is often other industry professionals. Will readers of the translation need a little extra help in order to understand it? What is the author trying to say? Is s/he making that clear enough or is there a chance the meaning might be misunderstood? Is any of it humorous? If so, how can I make sure the joke doesn’t get lost in translation? (Machines don’t have a sense of humour. The only joke they know is the one about malfunctioning at the most inopportune moment.) Which aspect of the text is more important: single words or the overall sense? The idea is – usually, though by no means always – to aim for “dynamic” rather than “formal equivalence” by creating not a literal translation but a text which has a similar [...]