The pros and cons of machine translation

Machine translation, meaning translation that is done by computer systems (with or without human intervention) started being developed in the 1950s. To the extent that machine translation (MT) has improved, its use has increased. However, it wasn’t until free machine translation tools became available on the Internet (such as Google Translate) that the use of this method of translation grew considerably among the general public.

This is understandable: if you need to translate a short text in order to get an idea of the overall meaning, then not hiring a professional makes sense. The objective and expected quality of that translation have nothing to do with accuracy, perfection or professionalism, given that it’s probably not even targeted at the public. In such cases, using a machine translation tool will meet your needs:

  • You get an immediate translation.
  • It’s free (paid translation tools are also available, which offer better quality).
  • It’s simple to get (you just copy and paste a text or write it, and tools such as Google Translate or DeepL will display the result in another language).

Plus, machine translation tools continue to develop as the technology advances. Right now in the market, there are three main types of machine translation technologies: rule-based machine translation (which uses linguistic rules and bilingual dictionaries), statistical machine translation (which uses monolingual and bilingual corpora) and neural machine translation (which forms associations by simulating the connections between ideas made by neurons).

The most sophisticated translation tools can even be trained so that they offer increasingly better quality (although involving the consequent costs in time and money).

But if they offer such advantages, why hasn’t the profession of a translator disappeared?

Considering all this, you might wonder why professional translators continue to be essential.

We’ll use an example to show why. Let’s say you need to translate an email into Spanish for a possible partner or client. Since it’s handy, you use a machine translation system. Once you have the target text (meaning the text translated into the target language), do you send the message directly without reviewing it first?

Probably not. You’ll most likely have to re-read the final text and make changes or corrections. There will be poorly translated words because they’re not right for the context, or the register might be too informal to send the message to a client. And to make such corrections, you need to have good knowledge of the target language.

In other words, the human factor is still key to obtaining a quality translation.

No matter how well developed a machine translation tool might be, there are certain aspects that only a human can recognise:

  • the tone of a text
  • certain plays on words, irony or double meanings
  • recognition of the (emotional) impact that one word or another has on the reader
  • the objective of a text

And professional translators specialise in the area of a text to be translated, meaning that they know the rules or standards and customs in each language.

Not to mention the cultural factor: for example, there’s no way for a machine translator to know if a direct question sounds too aggressive in the target language and if asking an indirect question in that culture might be more appropriate.

A professional translator can adapt to the register of the source text and follow the author’s same style (if it’s suitable).

Aside from that, when working with a professional translator, there’s always an open communication channel for expressing needs, requirements (specific vocabulary of a company, for example) or desires.

What conclusion can we draw from all this?

As we can see, machine translation is a very useful tool: in some cases, it can be used to obtain a first draft of a translation (especially if you’re not looking for business level quality) or even to supplement the work of a professional translator.

But it never works alone; human intervention is always required in the process. And if it comes from a translation professional, even better.

At Siens, we trust in the experience, work and excellent knowledge of people, and we view machine translation for what it is: a tool that can sometimes be efficient in the hands of a professional translator.