I want to discuss some technology-related challenges in the language services industry, particularly the impact of translation memory software and machine translation (collectively known as CAT tools) on Chinese translation services
Keeping up with changes in technology is an ongoing challenge for translators, and certainly one that I struggle with. I am a classic “late adopter” and I often avoid using a new technology until no other option is available when I provide Chinese translation services to my clients, who often tend to pay a premium price to have good quality work done.
I have certainly never been a fan of TM software – I have found that generally it is expensive (up to 1000 Euros!), slow to run, and prone to faults and unexpected problems. Though useful for files in unusual file formats, it still struggles to properly recognize and format text in PDF files, by far the most common type of file in medical and technical translation. The software seems to have been developed by people with European languages in mind, and doesn’t handle Asian languages as well. The use of tags slows down the translation process, and the use of segments takes away context, divorcing the translation process from the end target text. Even worse, clients make you reuse translations made by other translators which are often unsatisfactory, and pay you less because of “repetitions” and “matches”. I feel that the main beneficiaries of TM software are translation agencies (greater standardization) and the end clients (lower prices). Any advantages in productivity for translators tends to be negated by a decline in quality and a lower overall price per word.
With all the advertising done by software providers and the vigorous promotion of software by many translation agencies, it would be easy to get the idea that in this day and age all translators are using translation memory software all of the time. In fact this is not true at all. Neither I nor any of my close colleagues in Chinese to English translation habitually use translation memory software for most jobs. On the other hand, we do realize that some jobs do benefit from the use of TM software (usually jobs involving large files with repetitive content), and there are many more jobs in which use of TM software is required by the client, and if you do not have the software, you will not get the job. Therefore it is very important for a full-time translator to own at least one software package and know how to use it properly. The better and faster you can use it, the greater the productivity gains.
A trend for Chinese translation services in the use of TM software is for agencies to utilize their own proprietary online software, usually integrated into an online job management platform. My experience with this mode of translation have generally been very negative – this kind of software has all the disadvantages of regular TM software but is usually even slower and more buggy, and often lacks useful features offered by the more established software packages. Once I even used a program without a find/replace function, which is probably second only to copy/paste for saving time. I have responded to this challenge for the moment by charging higher rates for these jobs and/or avoiding these agencies wherever possible, however in future this may become more difficult.
Fortunately, most of the technology that really does boost translation productivity is inexpensive or cheap, and doesn’t require frequent updating. Microsoft Office (at least 2007) is essential, and a good optical character recognition (OCR) program (I use Abbyy FineReader) makes translating unformattable PDF files much faster, and also makes quoting easier. Google search and Google translate are the most useful online tools for translation, and are happily completely free. There are a number of free online Chinese-English dictionaries. These are the technologies I use every day in my work, and it is difficult to imagine translating without them.
Machine translation, and Google translate in particular, is the cause of much debate and consternation in the language services industry. Some agencies ban the use of Google translate and similar tools on the grounds of security. Many translators hate machine translation and see it either as low quality garbage or as a threat to their jobs. Nevertheless, there is no doubt that machine translation has become much better over the last 10 years, and despite this there are more translation jobs and more translators than ever before, so any threat to employment has yet to show itself. I personally love Google translate, and I can’t believe that it is provided for free. In productivity gains, it is worth hundreds, perhaps thousands of dollars a year to me. In my work, a machine translation is an essential first step for most technical translation jobs. It serves as a template to work over that reduces time spent typing in words and looking up dictionaries. For common simple sentences and lists of terms, a good machine translation may not require any altering at all. Because Google translation is based on a massive corpus of existing texts, every now and then it will produce authentic subject-specific phrases and sentences that are actually better than what I can do unaided. Although Google translate destroys formatting of a file, there are online tools that can preserve formatting, which is very useful for files that contain tables. Of course, no machine translation should be trusted implicitly, and research is always necessary to double check unfamiliar terms or phrases.
Clients are slowly waking up to the increased quality of machine translation, and this has spawned another trend in the industry, the so-called “post-editing”. In post-editing, the translator merely ‘edits’ a machine translation instead of translating from scratch, and because of this, charges a lower price. Because regular translation is now increasingly produced from a machine translation template anyway, post-editing could very well be seen as an attempt to pay translators less for producing the same product. However, it is my experience that in practice most clients who commission a post-editing job are expecting a lower quality result than regular translation, and usually what I do is just send in something of a rough draft quality, and that seems to be acceptable. At the moment, post-editing is usually commissioned when there is a lot of content to process and it would be too expensive or time-consuming to order a regular human translation. It is still fairly rare, and a full-time translator can easily get by without doing any post-editing for now, but it seems likely that in the future, more and more translation jobs will morph into post-editing jobs.
The common theme here is that technological progress is constantly making translations faster and cheaper for clients, and easier to manage for agencies, but does not necessarily benefit the translator in terms of improved quality or more income earned per time spent on the job. The challenge for translators lies in finding ways to make the technology work for us too.
At Capital Linguists, we employ translators of the highest caliber in Chinese translation services, and manually produce translations that are faithful and authentic to the original, while capturing all the nuances intended by the original speaker. If you are interested in accessing high-quality translations, please do not hesitate to contact us at info@capitallinguists.com
