Monday, September 12, 2016

Explaining ‘Post-Editing’




As much as translators sometimes wish it wasn’t true, there’s been enormous technological progress made in the field of translation. Before 1970, machine translation was almost unheard of, but what initially began as an experiment has today become a reality in the world of translation.

Translators are quietly confident that machine translations won’t replace human translations, certainly not in the short term anyway, but we can’t pretend that this technology doesn’t exist. It does exist, and more and more translators, agencies, and clients are using machine translations.

And this is where the term ‘post-editing’ has come from. Post-editing can be defined as the process of editing an automated translation. Today translators can use Google Translate and work with a bilingual document using Trados: a translation memory can be created, and this will be of great assistance for future reference and future translations. As the user runs the commands, the automatic translation will be displayed segment by segment. Enter the editor! The editor’s job is vitally important when using machine translations because the translated text will require human interpretation.

The Post-Editing Process

What should we expect from the post-editing process? Well, from a completed translation document’s point of view, the task of post-editing must be approached as one would approach a regular editing task. The editor will take the automatically translated text and perfect and polish it the same way they would a translation completed by a human translator. Just because a translation has been completed using AT tools, it doesn’t mean that the editor can lower the quality of their work. In fact, it’s quite the contrary! With automatic translations, it’s probably even more important that the editor pays very close attention to the translated text because we must realize that a machine has no analysis capabilities.

When Should Post-Editing Be Used?

Automatic translation is probably best suited for texts that are scientific or technical in nature; texts with a limited vocabulary. And when we say ‘limited,' we’re referring to the number of meanings that a specific word might have. It’s certainly true that both the scientific and technical fields are becoming more complex; however, the greater the technical complexity in the document, the more specific the word translation must be, so we must assume that it’s more likely that the machine will select the right word.

Texts Not Recommended for Machine Translation

The least-recommended texts for machine translation are probably literary works, such as poems, novels, and so on, because, in literature, the author's intention is an essential element; and also because these type of texts require the translator to have great interpretive qualities. In these instances, the post-editing process would become so complicated and so large that the whole document would probably have to be re-translated.


So, for translators to stay current and up-to-date, and to ensure they continue adapting to the reality of today’s translation industry, it’s important that they accept (and even welcome) the fact that machine translation is a trend that’s growing, particularly in the area of industrial relations.

No comments:

Post a Comment