Proper Process for Translation Work

Different methodologies should be used depending on what material is at hand when performing translation work. When working with shorter and more repetitive texts, we have a somewhat different process compared to when we work with longer texts that are more original. This holds true also for product descriptions. Some owners of web shops prefer to have shorter texts when writing their marketing material and product descriptions.

Examples of this might be web shops which sell technical units, where we often have many different types of units being sold, but with relatively little variation in description. Compare this with for example an iPhone or a car, which are still technical units. However in this case we are more able to write marketing material and product descriptions, and adapt our translation work to touch upon emotions in the reader.

How Technology Comes into Play in Translation Work

In the first case with the more repetitive descriptions it is hence useful to utilize software in order to aid the human translator in the task. If the translator has two product descriptions where the phrase “Technical Description” is present, then the software will remember that this is the proper translation when the translator has done the first product description, and transfer it automatically to the next one. By doing so we avoid duplication of work, the client gets the material faster and since the translator needs to put less time on each document the cost goes down (scale economy). Everybody wins.

When we perform our translation work it is however important to note the difference between translation software and machine translation; here lies a crucial distinction. Machine translation is an automatic translation done using a website or freestanding software, where the translator does not oversee the process; i.e. the translator does not perform the actual translation by using his own sound judgment in regards to what the translation “on the other side” ought to be. This opens up for all kinds of errors, such as misrepresentation of meaning, errors in grammar and structure and straight out nearly incomprehensible texts in some occasions.

This is because the software can not notice the fine differences between different types of expressions. The software is not aware of the outlined purpose of the text, which is a result of the discussion between the translator and the client. In order to create a software which captures all the necessary aspects of translation, we then need to build an artificial intelligence for translation work which is quite close or identical to the human mind. But this is a philosophical discussion that we will leave be for now.

If the translator however uses a software as aid in the translation work, then we do not have a machine that is translating the actual text, but simply a process that is being repeated in order to avoid duplication of work. In other words the software simply does an automatic repetition of phrases which have already been translated. Using this kind of translation software for relatively repetitive texts is also preferable since we then diminish the risk of errors, which are often the result of monotone work with phrases which are exactly the same. For each phrase in the previous texts, that has an equivalent in the coming ones, we also exponentially lower the overall cost of the project since the automation of the repetitive translations stacks up.

Now if we instead have longer and more elaborate texts, such as for example an iPhone or a car above, then it is paramount that we limit or all together avoid using translation software. In this case we have more qualitative texts, as opposed to quantitative. The process therefore becomes more similar to translating a novel and psychological factors related to marketing come into play, as well as SEO (search engine optimization).