Ever since Machine Translation (MT) was first mooted as a technology – and years before it ever became a viable tool in the translation workflow – different commentators have debated the effect of its introduction on the human translation community.
Still today, we see pieces on websites, blogs and in other media whose positions are diametrically opposed; just to pick two fairly recent publications at random, contrast Five Reasons Why Machine Translation Won’t Replace Human Translation (November 29th 2012) with “machine translation could replace human translators for written texts” (June 11th 2012).
Given the amount of MT and translation technology courses available (without too much effort, the other day I found over 120), it is a bit disappointing that despite all this effort to educate people – presumably at least a fair amount of whom have entered the translation industry – that the same ‘old chestnuts’ persist.
In a more recent piece still – Translation Industry at the Crossroads (January 24th 2013) – we see a section heading “MT developers seeking to replace linguists?” This article states that “MT developers are enthusiastically trying to replace interpreters and translators”. (Although soon thereafter we read that “that may not be their intention”, and later still that “IT developers do not set out to undermine translators and interpreters”.)
Interestingly, this article is written by a human translator. To the best of my knowledge – and I’ve been building MT engines of all types for 25 years now, and for most of that time educating young people as to the proper place of MT in the translation pipeline – I have never heard any MT professional utter such a phrase; we are all very happy to acknowledge the critical role of the human in the translation process (cf. Way & Hearne, 2011). This would appear to be a myth (call it ‘disinformation’ and/or ‘scaremongering’ if you like) put about by translators, presumably as a tactic to stir up negative sentiments towards MT in their colleagues.
Of course, this claim is incompatible with the other main comment from translators on MT, namely that the quality you get is hopeless. Of course, if MT was as bad as they say, there’d be no chance of translators being replaced; by the same token, if Machine Translation was so good that translators were going to be replaced, then there would be no need to criticise it, especially with the level of vehemence that abounds, a topic that we address in our next post.
The truth, of course, is that there is room for both. It is a fact that more and more use-cases are presenting themselves where there is no opportunity for human intervention in the process. This is a good thing for MT, and should not concern human translators one iota; only a fraction of what is translated actually is being translated, so there is – and always will be – room for us all.
It is also the case that many translators are using MT already. They have embraced it as a useful tool in their armoury. But that’s all it is: one tool among many. MT quality from respectable suppliers – like Lingo24 – is good enough to be useful to translators now, and I would encourage more to approach MT with an open mind. There will always be some who overhype what MT can do, but seek out those who can be trusted and you won’t look back. Rather than translation being at a crossroads, we’re all moving in the same direction – some more quickly than others – but there is no conflict between MT and human translation, and it would be better for all if commentators did not present the situation as such.
Andy Way and Mary Hearne. 2011. On the Role of Translations in State-of-the-Art Statistical
Machine Translation. Language and Linguistics Compass 5:227—248.