[...] technologies can sometimes be useful to literary translators as well, says Bowker. “They can compile a list of an author’s favourite expressions, as well as sentences or words that were translated in earlier works. By cataloguing neologisms or lexical creations so that they can be compared with a larger sample, technology can help the translator decide whether a more creative approach is warranted.” Bowker concludes that these technologies should be regarded as tools, not as potential competitors: “Human translators will never run out of work.”
The magazine of the University of Ottawa

dictionaryFirst of all: What is a CAT tool? Yes, WikiPedia has it: http://en.wikipedia.org/wiki/Computer-assisted_translation It’s a software tool that helps translators doing their work. A lot of people call it “Translation Memory System” while I would rather call it “Translation Environment” (TE). Anyway, it’s all the same.
The first book I translated using a TE was the children’s book Mozart in the Future written by my wife in Brazilian Portuguese which I translated into German. Now I am translating my second book written by another Brazilian author. My wife will have to wait. She is already waiting in line with her second children’s book The Legend of the Black Lake. When I find the time I will translate that one, too, using a TE, of course!

Why do I use a CAT tool at all?

The COO of a company that produces and sells a TE asked me this question. Well, here is the answer: You don’t have to, actually. If you have always walked to the bakery which is less than half a mile away, why should you use a car instead? If you always talked to your neighbor personally, why should you call him using a cell phone?
Okay, this is not a philosophical question. I would say that “older” literary translators probably would feel more comfortable without a TE. Computer-savvy and rather “young” translators would and should use a TE. (What about the old and computer-savvy translators?) Here is why.
First of all some basics. In order to be able to translate properly and to be able to reuse already translated “parts” of the text, the TE splits up the whole document into segments. Most of the time these segments are sentences which start with a capital letter and end with a full stop. Consider this sentence: “He asked Mr. Green to paint the wall.” Now this should be one segment even though there is a period right after “Mr”. All TEs have segmentation rules which handle these exceptions.
In my case there was one main reason why I used a TE: In Brazilian Portuguese there are a lot of commas and semicolons. The German sentences have to be as short as possible so a German child can understand the text more easily. So I changed the segmentation rules dividing segments not only by full stops but also by commas and semicolons. Since not all commas and semicolons relate to full stops in the German target sentence, I used the “Join segments” feature of my TE to correct this while translating. This is some extra work, however, the benefits are that I get a lot of short segments which relate directly to more and better matches in the translation memory (TM). A TM is a database that holds all the translated segments. So similar or identical segments do not have to be translated from scratch.
I don’t use the matches in the same way I would use them for technical translations because I want the text to be rich rather than to have repeating or similar phrases. However, the matches give me hints on how I translated a specific sentence before. So I can decide if I want to use it or not.
The book I am translating at the moment is a non-fiction book about a ballet dancer who suffers from a Down syndrome: www.bailarinaespecial.com.br The book is called Eficiência na Deficiência. The dancer’s father is the author. The text is a lot more challenging and the short-sentence-rule doesn’t apply to the target text. Nevertheless, the semicolon-rule applies. So even with this book changing the segmentation rules helped a lot.

What about terminology?

A terminology database or term base is a kind of on-line dictionary which is part of the TE. Yes, also with literary translations the use of a term base can be helpful. You cannot remember all the words you looked up in dictionaries or in reference materials. Again, term base entries only give you a hint. It’s not about consistency or even quality assurance, it’s about helping you to decide if the same term or a synonym should be used.
With the first translation I didn’t use the term base at all. Although I started out to use it, I soon realized that there was not a single term I had to look up. My wife and I are married for 13 years or so (I can’t remember) and we have spoken to each other in Portuguese ever since I can remember. (This is what I do remember.) My wife also speaks Portuguese with our children. Her book Mozart in the Future is written for children, so the language used is quite simple. In the end I couldn’t find a benefit in using the term base.
With the second book it’s different because I haven’t been married to the author. There are some difficult terms which I don’t use on a daily basis so it really helps to store them in a term base for reference. Again, for reference and not only for reuse.

Is machine translation helpful?

I don’t want to discuss machine translation per se. It is definitely not a means for translating literary texts. But it can be helpful. Let me explain how.
When I look up vocabulary I use several resources. There are not a lot of on-line dictionaries for Portuguese-German and most of them are pretty weak. Since I know the meaning of most of the words anyway I use the on-line dictionaries to give me a list of possible translations which help me to decide which term to use. It is easier to use a tool than to let my mind come up with a list. Knowing the meaning of a source word and coming up with a translation from your mind are two different things. Since Spanish is very similar to Portuguese and I also speak Spanish very well, I use LEO quite frequently. If there is a similar word in Spanish I look it up and the German translation might give me an idea. If I am not sure I use an on-line German dictionary from the Universtity of Leipzip. I use the list of synonyms to check if the term I have chosen is fine or if there might be a more appropriate term. I also use the Google translator, especially the list of terms that appear on the lower right.
Looking up terms at different sites is somehow time consuming. So I thought that it might be helpful to have a proposal within the TE.
Steps:

  1. Open http://translate.google.com/toolkit, create a private TM, load up your document and have it MT-translated.
  2. Download the translated document and import both the original and the translated document into the TE using the alignment tool.
  3. Write the aligned matches into a separate TM
  4. Set the TM read-only and add it to the translation project as the secondary TM

I think I don’t use a single match of those that were machine translated, but they often give me a hint on how to (or how not to) translate. Quite often the terminology used by Google helps even though I have to use a verb instead of a noun or vice versa.
The use of MT is not always positive. In most cases it’s better to think first before you even read the machine translated text. I wouldn’t explicitly recommend to use MT for literary translations, however, I would recommend to give it a try having in mind the above mentioned thoughts and proceedings.

Conclusion

Forget everything I have written about literary translations using a TE. If you read this articles in this group, that means that you probably already use a TE. And if you would be faced with a literary translation project you would probably use a TE anyway. Everything I have written about the use of the tool itself, the use of terminology and the use of TM and MT really depends on your personal likes and dislikes.
The major drawback of using a CAT tool is that it may influence your creativity. And creativity is important when translating a literary text.