no-machine-translation 318 × 292Machine translation vs human translation. It’s like something from a sci-fi film – the battle between humans and machines; artificial intelligence turning on its human creators and, at best, rendering us obsolete, at worst, turning us into biofuel.

It’s a debate which is always rumbling along in the underbelly of the translation industry, and periodically raises its controversial head, with the Machine Translation Enthusiasts/Companies doing their best to convince unwitting businesses that it’s The Way Forward, and the Sensible People just smiling a bit condescendingly and getting on with persuading companies to source professional linguists for their translation work (which is basically what I’m attempting to do now).

The thing is, no matter how sophisticated the tools become, they’re always going to lack the human aspects of language. They’ll never grasp that words carry with them a multitude of cultural connotations and historical significance which can never be programmed into a machine.

Additionally, machines won’t ever master the art of word play, or understand stylistic choices. Neither will they query something they don’t understand if the context doesn’t make it clear. They can’t do extensive research or ask colleagues or clients to clarify an acronym. Yes, they can use sophisticated algorithms to cross reference databases and dictionaries and web pages, but that’s still no match for a person, who won’t be restricted by programming limitations, and can think around the problem.

Moreover, languages change constantly, subtly, with each year that passes. People make up new words and hijack old ones to take on new meanings. I’m sure everyone reading this can think of at least one or two words that are used between friends or family members, which probably aren’t examples of standard usage. I forget sometimes that ‘autumn’ is not synonymous with ‘awesome’, except when I’m talking to my favourite Canadian; likewise ‘How’s cows?’, a greeting I sometimes use with one of my old school friends, is not a standard way of asking how someone is, not even between farmers. A machine would produce gibberish if it came across nonsense like that; a human would at least be able to ask me what on earth I was talking about and find a way to convey the meaning.

This doesn’t mean I’m against machines helping out – they can be really useful if they work alongside human translators to store their work, enabling them to utilise it in later jobs, saving time and increasing productivity. But even repetitious documents and manuals, the areas where machine translation is supposed to have the greatest advantages, need that human understanding behind them – people will be reading them and using them after all, not robots (unless they do actually win in the end).

Therefore, no matter what ambitions they might have of overthrowing their meatbag creators and taking over the translation industry, I just can’t see machines ever being able to do it, because language is tied up with so many other aspects of humanity that they’re going to have a really, really hard time truly understanding it, let alone producing it convincingly.

So, at the risk of being targeted by robots that look like Arnold Schwarzenegger, I proudly declare that Lingua Translations will always be a leader in the rebellion against the machines. We only use 100% organic human translators for each project, who have extensive training in their subject areas, and therefore guarantee translations of the highest possible standard which will be fully in line with the expectations of your equally human target audience.