I recently received a request to participate in an interesting survey being conducted by Ana Frankenberg-García that highlights one of the dilemmas translators face when working with translation memories and CAT tools in general. Frankenberg García holds a doctoral degree in Applied Linguistics from the University of Edinburgh and currently divides her time between teaching at the Instituto Superior de Línguas e Administração in Lisbon, Portugal and professional translation. In 2001, she was awarded a prize by British Comparative Literature Association and British Centre for Literary Translation for her translation of Rubem Fonseca’s short story “Lúcia McCartney”. She has served as an external English-Portuguese translator for the European Parliament since 2006.
For readers who are unfamiliar with how computer-assisted translation works, the survey questions listed below are related to how a text is formatted before it is sent to a translator. If the source language author has constructed long, complicated “run-on” sentences, a translator often needs to divide them in order to maintain syntax and improve readability in the target language. CAT systems often break the text in an arbitrary fashion and divide it into neat little segments that are displayed in frames. Each frame of text must be “closed” before going on to another, making it very difficult for a translator to rework the word order or control the phrasing of extremely long sentences. As the translator does not have a coherent overview of the entire text, problems such as dangling modifiers and comma splices can often go undetected and imposing a coherent style on a text is much more difficult. Although current translation systems offer speed and vocabulary standardization, they encourage translators to make stiff, literal translations of each component sentence rather than accurately and fluidly translating its message.
I’m certainly not a Luddite. I’d never trade my computer for an old-fashioned Olivetti and I’m always eager to learn more about the technology that is transforming how we communicate. Although the majority of my work does not involve the systems described here, I do occasionally use CAT tools. I believe, however, that two distinct professional profiles are developing within the field of translation, one oriented toward speed and volume of production, and the other more oriented toward providing high-quality professional communications services.
The following constitutes the entire survey I received.
When translating segment by segment within a TM (Translation Memory) system, how often in your translations do you join source-text sentences together or split them apart? *
- Almost never
- All the time
Do you think you would split more source-text sentences apart or join more source-text sentences together if you were not translating segment by segment within a TM system? *
- Definitely yes
- Probably yes
- Not sure
- Probably no
- Definitely no
In your last translation job using a TM system, how many source-text sentences did you split or join? *
- 0 (I did not split or join any source-text sentences in my translation)
- Between 1% and 5%
- Between 6% and 10%
- Between 11% and 20%
- More than 20%
The results of this survey will be presented at the Corpus Linguistics 2011 conference in Birmingham, England, in July. It will be interesting to see the results.
Translators interested in participating in the survey can do so at https://spreadsheets0.google.com/viewform?hl=en&hl=en&formkey=dGVmU3F1SC12bmQ2dlVFeFJCUjBZaHc6MQ