Reconstruction : post by François Hébert

After reading a number of recent conversations, I noticed that many myths surround translation memories. The main myth is we become enslaved both by our tools and translation agencies once we purchase the tools. I believe it is a good time to closely examine the topic.

When used correctly, translation memories offer much to translators, namely automatic formatting in various files like Excel, PowerPoint, and Word. Automatic formatting is a must when translating information technology (IT) products such as video games, websites, or complex machine interfaces.

Integrated terminology management makes consulting glossaries faster and easier. They can be classified by client, sector, or brand.

Automatic transcriptions of placeables such as proper names, mailing and e-mail addresses, and websites offer great advantages in speed and accuracy.

In addition, you can search for previously-used formulations that are neither complete sentences nor terminology units. In so doing, you create your own concordancer. As long as it contains your translations, it is more reliable than public concordancers like Linguee.

Most memories offer quality control to conduct checks that differ from proofreading software. For instance, memories check that target segments have the equivalent terminology in source segments. Memories also check that target segments have the same numerical values as source segments. In other words, these types of checks can only be carried out in bitext format.

Memories can quickly analyze large documents to spot the quantity of internal repetitions. A one million-word technical document with two thousand identical sentences is not the same as an equivalent text without repetition. It is almost impossible to carry out this type of analysis manually.

I have not gotten to the core of memories of late: recalling previous translations. I think that the aforementioned functions are considerable benefits.

Important productivity gains clearly come from reusing previous translations. This is where translators are to use professional judgement. Some sectors are well-suited to previous translations, while other sectors are not. Exact matches may sometimes have to be modified. Additional cognitive efforts are rewarded when productivity gains are justified. Professional translators can take advantage of memories when used intelligently.

But memories undoubtedly pose hazards. First and foremost, translators must relentlessly protect their memories. When many people share the same tool, it could quickly become polluted. Translations may be taken out of context, terminology may be inconsistent, or other errors may arise. In these cases, memories are useless, even detrimental.

Remember, we can never be enslaved by memories. As you know, I regret the way firms exploit translators when asked to do “patchwork.” Such work simply distorts texts and converts translators into technicians—or paralinguists, as Sylvie Lemieux suggests. The heart of the matter does not lie in using translation memories, but in the toxic environment a number of agencies create. The solution is simple: refuse to work in such an environment, and avoid working with agencies.

Given how agencies operate, quality of work is poor. This appears to sit well with some sectors of the global market. If we attempt to compete with this work model, we are asking for disaster. This is why it’s important to choose clients carefully and work with clients who care about quality of language in their projects, and believe that a high-quality translation is a worthwhile investment. It is up to us as professional translators to remember that we are not enslaved by our clients, in the same way we are not enslaved by our working tools.

François Hébert is a freelance translator and founding president of L'Odyssée, a training school for translators.

Translation and adaptation: Dwain Richardson