In the past, before technology made its way to translation, professional translators received the source text in a written document, and they would translate it entirely in another sheet. It sounds simple, until this human process suffers from minor errors. It required unnecessary effort from the translators and editors to spellcheck long documents, looking for things like numbering, or typos.
While the market is always looking for new ways to make any overwhelming task smoother, translation had its share with the advent of machine learning and software programs that are now-more than ever-stepping in to make the translation process less time consuming.
One of the most popular ways machine learning did that was by introducing computer assisted translation software, famously presented in computer programs like memoQ and Trados.
Definition of computer-assisted translation systems
Simply refers to the facilitation of the process of human translation by the aid of what computer software does best, storing, comparing, and retrieving of data.
Computer assisted, computer aided, or machine aided translations simply depend on a reservoir, or corpus of linguistic storage that is retrieved to aid the translator in his task.
CAT refers to a variety of tools that can be of aid to the linguist or translator. These tools include something as important as translation memory, grammar checking, and overall quality assurance tools.
The role of computer assisted translation systems
CAT systems are essentially made to make translation more efficient, less time-consuming, and cost effective. This is achieved by the integrated translation memory that helps maintain consistency in translating similar words or phrases.
That being said, CAT do not attempt any translations on their own, unlike machine translation, which we’ll dwell on shortly.
Stages of computer assisted translation:
According to an article published by IOP Science, computer assisted translation can be classified into four main steps, analysis, transfer, reorganization, and inspection.
The CAT system first analyzes the imported language file, dissolves it of all the non-textual elements and uses the segmentation tool to break down the file into independent stand-alone phrases.
While doing so, it consults the translation memory to look for direct or similar matches inside the document that it can auto-fill.
The translator proceeds to attempt his or her translation of the remaining segments in the document, all of which are also added to the translation memory database.
What is the difference between machine translation and computer-assisted translation?
Translation memory is not to be confused with Machine Translation (MT), both are two distinct things.
Moreover, the primary difference between machine translation and computer-assisted translation is that the former depends on the machine as the main provider of translation, after which a human linguist or translator might edit or proofread, this human process is known as Machine Translation Post Editing (MTPE).
However, computer assisted translation depends fully on the human translator to lead the way. It allows for edits and customization. This room for human intervention can make the target text more contextualized, consistent, and culturally specific.
Many of the free online tools available on the web now are considered machine translation.
History of computer-assisted translation
Attempts to create successful CAT go back to the post war decades of the 20th century where it was faced by difficulties. There were attempts in the 1970s and 1980s to develop translation memory ‘TM’ that were trained to store translated sentences in a database. In more recent decades, computers were able to offer more than just storing information.
CAT developed to encompass tools like terminology management, spellchecking, grammar checking, and even integrated machine translation.
During the late 1980s and 1990s, CAT tools were finally becoming more accessible either for freelancers or businesses. Programmes like Trados, Wordfast, and Déjà Vu as examples of CAT software gained popularity.
However, how far the CAT is efficient? To answer this question, let’s list both the advantages and the disadvantages of CAT tools.
Advantages of computer-assisted translation
1–CAT is cost effective for businesses. Clients will scan the documents for matches and repetitions, sometimes they might put them intentionally in order to offer translators a lower price.
2-Proven to ensure consistency of translation over time. This is because terminology management tools inside CAT ensure consistency of translation especially with several professionals working on the same project simultaneously.
3-Speed: It aids in meeting deadlines, especially with large technical documents.
4-CAT comes with quality assurance options like spell and grammar checking to ensure a polished and smooth to read target text.
5-Allows for several professionals to work on the same project, including project managers, proofreaders, and editors.
6-Ability to work on a huge number of projects or documents at the same time, given the significant processing capacity the CAT software is capable of.-
7-Automation of repetitive tasks.
Disadvantages of computer-assisted translation
It is important to keep in mind that double-checking the CAT translation can sometimes save the translator from massive errors. This is because CAT is prone to contextual errors. Like in the case of homographs (words that are spelt the same but have different meanings).
Given the CAT limitation to TM, it is not quite rewarding when it comes to translating literary texts that are loaded with double meanings, and metaphors, and rely heavily on cultural differences and localization.
Other disadvantages to keep in mind include:
-Need to always be up to date, undergoing additions, and improvements.
-Risking translators’ overreliance on technology and computer aid.
What is Translation Memory (TM)?
In CAT terminology you will find the term Translation Memory, also abbreviated as (TM) which refers to the database of translated segments or strings. Simply, translators don’t have to translate the same phrase, or word again when repeated, they already translated it once before, and so the TM steps in to take care of future similarities or matches.
What’s the difference between CAT and translation management systems (TMS)?
TMS combines the best features of CAT, as well as the speed and smoothness of machine translation. TMS achieves localization by using artificial intelligence to translate text, and then contextualize it with the help of professionals and native speakers.
Optimal use of CAT tools/fast trans
Lastly, there are best practices to ensure that you get the most out of CAT tools, below are some of them:
1-Upload the document in editable format, not a pdf format.
2-Make sure the source text itself matches the mechanism in which the TM works. This involves ensuring it is error free, and easy to read without ambiguity.
3-Knowing that the more a text is context specific, the less likely it will be used in future translations.
4-Double checking the CAT translation to avoid mistakes.
Conclusion
To conclude, CAT are not to be confused with machine translation because both of them work opposite to each other. CAT makes use of the simplest tasks the computers excel at doing, but it is tweaked in a way to aid in the translation process, without marginalizing the translator or making his role secondary. It remains one of the earlier attempts that sought to make the task of translation less distracting, easier, less time-consuming for linguists.
