Could you use a corpus?

Using language corpora in translations

All translators have specialist areas, but most of us often work at the edges of them and sometimes outside them. And when we do, we are much less certain of our choices of words and expressions because it’s much more difficult to be sure that experts in the field wouldn’t say what needs to be said in a slightly different way.

Translators, too, often live for long periods of time away from their home country, where language is moving on and slowly leaving them behind. How can we be sure that an expression that was common currency 20 years ago hasn’t become outmoded and superseded by others. We can, of course, listen to the radio and make regular visits home, but that only gives us a subjective view of the way language changes.

These are just two of the reasons I was so excited last week to go to an all-day training session on language corpora organised by Mediterranean Editors and Translators (MET) in Barcelona. And, when I came out, the fireworks of new ideas for using these computerised collections of texts were going off inside my head.

So what can be done with target-language corpora (I’m not talking about parallel ones in two languages here) that makes them so potentially useful to translators? Well, they can be searched and analysed quickly using computer-based or online tools to answer some of the questions we face every day. Do we really say that or is it a translationese copy of the source language structure? Do experts put it this way or that way? What verb usually goes with that noun in my context? What preposition is used with that verb by people writing in specialist journals? Has that expression gone out of fashion nowadays? With a little thought, the possibilities expand even further.

Not only that, it turns out that other tools are available so we can make our own corpora very easily, using just a few key words, from texts available on the Internet. I tried out the tool we were shown and quickly made one containing three-quarters of a million words, all from texts that should be relevant to a big town planning translation I’m currently working on. I shall try it out and, if it proves as useful as I think it will, buy the tool.

What impresses me most is how practical it is to use this technology. You don’t need to be an obsessive or a techno freak to be able to do it, as building a corpus takes just a few minutes and using it is almost instantaneous. Any translator who wants to can have a useful tool available almost guaranteed to improve his or her work  in just the time it’s taken me to write this blog.

I can honestly say that not since I bought my first CAT tool, Déjà Vu, ten years ago (now superseded on my computer by MemoQ) have I seen anything with such potential to take my translation to another level in terms of sounding natural even when I’m not an expert on the subject matter. I shall keep you posted.

You can find out more about MET by going here. An article on the use of corpora co-written by one of the MET workshop leaders can be found here. And here is a link that should allow you to download details of WebBootCat, the tool for creating your own corpora.

on Twitter on Facebook