10 Reasons why Human Translation beats Machine Translation
Technology is very clever. It can do many jobs for us and go a long way towards making our lives easier. One of the jobs it can do for us is translating between languages. Great! Stick your content into Google Translate, pick your languages and you’re good to go – or are you? There are many reasons why you might want to use automated translation: it’s quicker, often cheaper, and you can translate large quantities of text in seconds. But there are also many reasons why expert, qualified human translators might do the job better than machines.
What is machine translation?
Before we start, what is machine translation and what is human translation? As you might have guessed, human translation is the process of relying on a person’s intelligence, linguistic expertise and professional experience to translate a text. Machine translation, therefore, is when an application or piece of technology does the translation work. You might also have heard it referred to as automated, automatic, or instant translation, but they all mean roughly the same thing. Google Translate is the most famous example of machine translation, although there are several types, including some that can be edited and trained to get better results. Let’s also establish what isn’t machine translation – there are many technology resources available to language professionals to make the translation process faster, more accurate and more efficient: CAT tools, translation memories and project management software, to name a few.
Here we go – machine translation vs human translation, who wins?
10 Reasons why Human Translation beats Machine Translation
1. Accuracy
2. Language is constantly evolving
The way we use language never stands still for very long. Words fall into or out of use. They gain second, third meanings that are completely different to their dictionary definitions. Take as an example the world of social media, and the dizzying speed with which the meaning of words like ‘cloud’, ‘catfish’ and ‘viral’ can change. A human translator immersed in the culture of their source and target languages will already be aware of these new meanings. Machine translation engines will need a bit of fine-tuning to catch up.
3. Localization and cultural differences
We’ve established that machine translation can’t understand cultural nuances in the same way as expert human translators. You might be thinking, “But my text doesn’t contain any cultural references, I’ll be fine!” However, things like the level of formality, any idiomatic expressions, and aspects of the language specific to where your audience is located (UK or US English?) will need to be understood and managed by your translator. You want your audience to feel that the content was tailored to them. Would you trust a machine to achieve this for you?
4. The issue of tone
The tone can really make or break a text and how it is received. Whether your text is consumer-facing, creative, persuasive, or intended to leave a particular impression, the translator must get the tone right. Will a machine know to translate agité as excited rather than agitated or nervous? Will they know that your Linsen refers to lentils, rather than camera lens? A human translator will use the context to decide which word to use. A machine translation engine will try its best, and in some cases will be able to register the context and the nuance to be able to choose the appropriate translation, but not every time.
5. Creative texts don’t suit machine translation
It’s also important to note that some texts suit automated translation better than others. It might be a good idea to use machine translation if you only need to understand the general gist, or if your text uses very standardised language. Simple or formulaic texts would work well with machine translation. However, it would not be ideal to use machine translation for your complex or creative texts. A machine will not be able to create prose as reader-friendly as an expert translator would and could cause some embarrassing problems if your source text contains any sort of hint, metaphor or euphemism.
6. Marketing texts don’t suit machine translation
If your text is written for marketing purposes, or is consumer-facing, you should avoid machine translation. Due to the cultural differences I mentioned earlier, there often needs to be quite a lot of editing and altering for a marketing translation to be successful. A translator who is knowledgeable about SEO, your marketing goals, and the brand voice you worked so hard to create can help your brand shine in your target market. Google Translate probably won’t be much help with this.
7. Collaborate with real humans
As good as it is to be able to enter a block of text into one box and have its translation quickly appear in another, there’s also a lot to be said for working with real people – real language professionals. Whether you choose to work with a freelance translator or a translation agency, your chosen professional will be able to provide a helping hand throughout the process. A machine translation will give you the same output each time, but a human translator will be able to make changes and revisions based on your instructions and feedback. Collaborating with a real person means better communication and working with someone who really cares. They will be happy to check over your source text, provide advice on your target market, and make sure that you are completely satisfied with the service provided, every time.
8. Confidentiality and data security issues
When you use machine translation tools, the content you have input is likely to be subject to terms of use agreements. This means that the provider, including Google Translate, is usually allowed to store and process your text. It’s not necessarily malicious – the algorithms use it to learn and improve the machine’s translation skills. However, this data could potentially be hacked, and inadvertently exposed. Therefore, if your content should remain confidential, you can be sure that a human translator will be more discreet. At Planet Languages, our information security management system has been independently verified and certified to the internationally recognised ISO 27001 information security standard, so you can be sure that your text is in good hands.
9. Documents for legal or official purposes
Certified translations are often required when the translated document has a legal or official purpose. A certified translation service will provide you with a stamp of approval, declaring that the document is a true and accurate translation of the original, completed by a suitably qualified professional. Machine translation tools can’t certify translations.
10. High-resource vs low-resource languages
While the type of text might dictate whether a human or a machine should do your translation, it is also important to consider the languages you’re working with. If you’re working with a widely spoken language, then the quality of machine translation is more likely to be better. This also means that a less-spoken language will have fewer resources available to draw reference content from, therefore the quality of machine translation is likely to be poorer. If you’re unsure whether your languages are high or low resource, contact your chosen translation partner, who will be able to advise you on the best course of action.
What is human-aided machine translation?
The decision between human translation and machine translation isn’t quite as black-and-white as you might think. There’s also the possibility of human-aided machine translation, also referred to as Machine Translation Post-Editing. This is where an expert will look over the translation done by a machine and correct any errors. This can be a really good way to save time, and therefore money. However, if the text becomes particularly mangled by machine translation, this can sometimes take longer than the regular human translation process. Here at Planet Languages, we’re certified to the ISO 18587 – post-editing of machine translation output standard – and we take a responsible approach to integrating machine translation into workflows.
Conclusion
Technological advances in different walks of life often create uncertainty. To make the most of recent advances, you need an expert human in the loop to achieve the best results. Many language professionals, freelancers and agencies alike, have welcomed these developments. Machine translation and language translation software can make our lives much easier – we must just use them wisely. It’s important to know when you might be better off sticking to good, old-fashioned human translation.
Contact us now to find out how we use technology to get you consistently high-quality translations, while maximising consistency and reducing your costs. Why choose when you can have the best of both worlds?
About the author
Emily Barnett graduated from Lancaster University with a 2:1 in her Integrated Masters in French & German Studies, where she also received the Swiss Ambassador Prize for her performance in German. She went on to graduate from the University of Manchester with an MA in Translation & Interpreting Studies. She is passionate about content writing and languages, and currently learning Dutch in her spare time!