Translated Target segments in the Source text


I have  source documents with some segments are already translated to target lang.  What happens to these translated segments if I submit this documents to MT?

Hopefully, the MT skips such segments and returns these segments with no changes/translation?

  • Hello Miyuki Komuro,

    Machine translation will try to perform segmentation first to split the content into sentences based on the segmentation rules used for the source language. Then it will perform translation on  sentences and if the words are not recognized in the source language it will leave them untranslated. However, if you have a lot of unknown words in your source sentences, you can expect very low translation quality and in some cases completly unexpected/unrelated  target content.   

    For the best MT quality, you need to hide already translated text from your source. Hope that helps.