In The Hitchhiker’s Guide to the Galaxy, Douglas Adams introduced the babel fish which, when inserted into your ear, allows you to understand any language. Ironically, Adams’s trope made its debut during a decade-long lull in the real-world search for universal translation capability.
‘Machine Translation’ had been identified as a research theme by Warren Weaver back in 1949. Interest grew rapidly in the 1950s, but funding was cut back in the mid-1960s after it became clear that little had been achieved.
More recently, web-based translation has made a break-through conceivable. Online tools use statistical methods, searching for patterns across hundreds of millions of texts that exist already in parallel versions - for instance EU documents.
It’s a far cry from a Chomsky-inspired vision of universal grammar, and is prone to egregious errors, not least because in many cases translation involves two hops - from language A to English, then from English to language B.
Work on the next generation of online translators will bring us closer to a world of mutual instant intelligibility.
It doesn’t end with texts - tools already exist that attempt real-time translations of the spoken word. Combined with wearable devices, we can imagine a world in which perfectly sensible conversations are possible with someone speaking another language.
On this basis, it could be argued that MFL is the curriculum area most likely to be transformed by digital technology in the next decade. Improved online translation could conceivably be a game-changer, the question becoming, not how foreign languages should be learned, but whether it’s necessary to learn them at all.
With automatic translation of text and speech making it straightforward to navigate in a foreign-language environment without prior exposure to the language, what reason will there be to learn a language in the formal sense?
There are several enduring reasons to learn languages. The next generation of digital dragomen, even if they had access to the full lexicon and to every syntactic rule, will not solve the ambiguities of context, tone and irony that spice up speech. Translation and interpretation are very different things.
Some teachers baulk at the easy availability of online translation, but others encourage their students to use them as a heuristic device, comparing them with their own efforts.
Indeed, automated translation works best in a feedback loop, being checked and improved on by people who have some background in the target language.
Operational competence - being able to understand and be understood in transactional contexts - still requires a critical engagement. But academic MFL amounts to more than this; it is about making connections with different language groups, and exploring cultural as well as purely linguistic dimensions.
Academic language study requires discursive engagement. Translation software will not be able to turn a mediocre piece of writing into a good one. It will not turn a poorly structured argument into a better one.
Meanwhile, improvements in translation technology might help remove ‘barriers to entry’ for students, drawing them in to engage more deeply with foreign languages, cultures and peoples.
Dr Kevin Stannard is the director of innovation and learning at the Girls’ Day School Trust. He tweets as @KevinStannard1
Want to keep up with the latest education news and opinion? Follow TES on Twitter and like TES on Facebook
