News

With Google Translate being a free and easily accessible tool to quickly translate a few words it is no surprise most of us are guilty of doing this. However although some will end up receiving the correct outcome, there is a higher probability of receiving a disjointed sentence in the target language.
Below is a breakdown of how machine translation works, and why it should never be used in professional translations.
What is Machine Translation?

Machine Translation, also known as Automated Translation, refers to when computer software is used to translate a piece of text, without human intervention.

How does it Work?

There are three types of Machine Translation systems:

1. 1. Rules-based system:
   This type of system uses a combination of language and grammar rules, as well as specialist dictionaries, to produce translations for specific industries and disciplines. This system is said to deliver consistent translations with accurate terminology.
2. .
3. 2. Statistical system:
   Within this type of system, there is no previous knowledge of language rules, so the system ‘learns’ to translate by analysing large amounts of data in both the source and target language. Additional data is required if the system needs to be used for specific industries or disciplines. A fluent-sounding translation can be expected, but it will be less consistent across multiple translation projects.
4. .
5. 3. Neural Machine Translation (NMT):
   This system learns to translate through multiple processing devices that have been modelled on the brain, known as a neural network. This approach has become increasingly popular amongst researchers and developers as NMT systems have higher performance rates compared to statistical approaches.

Why Human Translation over Machine translation?

Whilst artificial intelligence and technological innovations are on the rise, there are several aspects of translation that computer systems cannot understand and will be unable to detect, without the assistance of a human translator.

Culture
This is a key aspect of translation as different cultures around the world possess different lexical items, such as idioms and slang, that are unique to that specific culture, which need to be understood and recognised in order to be translated in an appropriate manner.

In the same way translators must update their knowledge and terminology databases, Machine Translation systems would require continuous updates and more sophisticated algorithms in order to deal with modern day vocabulary and cultural references.

Context
In many languages, there are words with dual meanings, which can cause significant problems for machine translation. Without human intervention, sentences could lack logical meaning, making the translation incomprehensible for the reader.

Style and Tone
Whether you are looking to translate a poem, a persuasive piece of text or an argumentative essay, style and tone can pose challenges for machine translation.

Whilst a Machine Translation system can translate a document, it cannot identify the purpose of the document. A Human Translator, on the other hand, would be able to identify the style and purpose of the text by looking closely at the language and format of the document, and replicate this to make their translation fit for purpose.
Glossaries and preference

Many of our clients provide glossaries for their translations which Prime Production always ask our translators to follow. In many instances as well clients have different preference on how things should be translated, especially in cases where there could be multiple translations for one sentence. A human translator will be able to develop their knowledge to be able to differentiate what translation should be used for which client which would not be possible through the likes of google translate.
Proofreading

A company will have a translation revised or proofread by another linguist to make sure the translation contains accurate information, and to ensure there are no grammatical errors or spelling mistakes. Whilst computer software can pick up on these types of mistakes, it is unlikely that it would be able to provide a professional editing service.

Machine Translation, whilst it is quick and often consistent, is intended to help Human Translation, rather than force them out of the field. Human translators could not produce a perfect translation without CAT tools and online research, in the same way that a Machine Translation would not be able to produce a perfect translation without human intervention. Linguists must continue to develop their knowledge in order to become better translators, so for now, machine translation will require further innovation that can provide regular updates in order to have a higher performance rate.

Here at Prime Production we strictly condone the use of Machine translations to ensure all texts are technically accurate and maintained at the same high standard we strive for at all times.

Prime Production Ltd is a one-stop language service provider based in the United Kingdom and Asia. Our Translating, Interpreting and Desktop Publishing services are available 24/7, 365 days a year for a diverse range of international clients. We have a track record of providing language services to tight deadlines at competitive rate, translating close to 2 million words per month, with 95% of our work being based on repeat business.

Should you need a document translating, do not hesitate to contact us at info@primeproductionltd.com to discuss your project, and to receive a free, no-obligation quote.