Anyone who can speak another language or has had to translate from one language into another will have come across the question: ‘Why can’t we just use Google Translate instead?’ – This is a question that frustrates a lot of translators as well as language learners.
Today we’re looking at the role of machine translation tools like Google Translate and other computer-assisted translation software within the industry. We’ll be asking the question: Will we ever be able to rely fully on machine translation and no longer need human translators?
In an increasingly electronic world, it’s no wonder that technology has found its way into the translation industry, and it has most certainly not been a bad thing either! There are two main types of translation technology: machine translation (MT) and computer-assisted translation (CAT). MT relies on the software alone to do the job, whereas CAT does what it says on the tin – it assists translators. Below are some of the ways that CAT software can help speed the process of translation along:
- It can be used to create translation memories which are databases that store
phrases, sentences, headings, titles, etc. that have been translated before
- It can be used to create glossaries of various kinds, such as a glossary
of specific terms within a certain subject. This means that consistency is
maintained throughout even the longest translation
- It can count words
- It can split texts into segments
- It can analyse these segments to see how much work the translation memory has done and how much is the work of the translator
- It can align sections of the original with the translation, so that the
translator can compare the texts easily
Basically using CAT tools means that more work gets translated within a shorter time!
But what about MT where the software itself does the bulk of the work? Here,
the translator’s role is more like that of an editor, rewriting and amending
parts of the text that have not been successfully rendered into the target language
by the machine. In some ways, this is just as much work (if not more!) for the
translator as if they were to have written the entire translation without any
technological assistance. The language produced by machine translation is often
unnatural and does not read fluently.
We only need to look at a classic example of Google Translate gone wrong to see that machines aren’t always the most reliable of translators!
An article in the Guardian from last year described how a culinary festival in Spain became a rather different kind of festival thanks to Google Translate. In As Pontes (north-west Spain) where Galician is spoken, a festival was being held to celebrate grelo, the leafy green vegetable that is grown in this region. When the organisers used Google Translate to advertise the event to Castilian Spanish speakers, it became something quite different: a clitoris festival…
And take a look at Google’s English translation of the start of a German news report:
Exhaustion scandal – Examination report defused?
Too much indulgence with the car industry? The clearing-up of the diesel co-operatives
is intended to have agreements between the Kraftfahrtsbundesamt and Autobauern.
Critical passages should then have disappeared from the examination report.
As we can see, sometimes machines get it wrong. They may not have certain words
in their translation databases and they certainly don’t have the same cultural
understanding as humans.
So while there will definitely be technological improvements in the future, it seems that the role of the translator is safe for now.