BlogBosses BlogHow reliable is machine translation at the moment and what are its limitations?

How reliable is machine translation at the moment and what are its limitations?

How reliable is machine translation at the moment and what are its limitations?

The most popular machine translation software at the moment is Google Translate, although other programs are available. The software has improved remarkably over the last few years, with regular “breakthroughs” being announced by its developers. But what are the limitations and how reliable is it in real use?

Advantages

  • It’s free! There’s no cost for most users and a reasonable amount of text can be processed.
  • It’s instant and it is available online: There’s no tedious waiting around and you can get an answer using a range of devices.
  • It’s versatile: You can use it to translate a small snippet of text, a page, an email or a whole website.
  • It covers many languages: There are more being added all the time.

Disadvantages

  • There are limits: The service does not cover everything.
  • Your language might not be covered: Many languages are offered, but what do you do if yours is not one of them?
  • Its translations are not natural: Google gives a mechanical, direct translation of the words you enter. The language barriers sometimes remain in place with confusing translations that can be ambiguous to downright unintelligible.
  • Context is not taken into account: This is the biggest issue with machine translation, in particular Google Translate - the context of your text. Although some machine translation systems can “learn” specific context and use bespoke glossaries, context is always an issue for this type of system.

In a language like English, a single word can have hundreds of different meanings, depending on the context. In fact, the Oxford English Dictionary’s lexicographer for the letter “R,” Peter Gilliver, claims that the verb-form alone of “run” has no less than 645 distinct meanings. Can a machine ever learn each of these meanings for every word in not just one language, but two? This isn’t an easy question to answer.

Not only that, but word-for-word translation is impossible.  When humans use context to figure out meaning, we think not just of single words, but how those words interact with the ones around them. Those combinations are constantly changing and multiplying, making it impossible for machine translation to keep up.

It’s also important to remember, that the software also won’t know who your target audience is, what their ages, education or culture are. It won’t know if it’s dealing with a press release or a training manual, a speech or a technical report.

There is still no way machine translation could successfully translate a website for example, taking into account SEO keywords or brand names.

Google is cutting edge in many ways, with unique algorithms offering some of the best automated translation available today. There is no substitute however for professional human translation, which looks at the text for translation from many angles, considering audience, intention, context, style guidelines, cultural sensitivities and more.

By all means use Google for getting the gist of something, informal communication or finding out what language something is written in, but it should never be used for anything customer-facing or anything official.

If you’re at all unsure about whether you could use machine translation, get in touch now on 0844 856 1086 or email enquiries@midlandtechnical.co.uk. We offer professional human translations of all types of documents – a cost-effective solution that won’t risk your reputation.

  • Recommend:
  • Share: facebook
  • twitter

0 Comment(s)

Be the first to comment on this blog!

Leave a Comment


Simple BBcode supported, more info here.

Homepage Link / Website: (Optional):

Please enter the text shown on the image. (Proof that you're not a robot)
Please type in the value of this image
Refresh Image


(Please note that your comment may first be submitted for approval before being added to this page)