I was talking yesterday with a young Swiss artist, Tobias Kaspar – who’s quite interested in translation in general and Robert Walser in particular – about how it is that in many parts of the world growing up bilingual doesn’t necessarily make one good at translating. In all those Swiss towns in which both French and German are spoken, though, I suspect that people are much better at translating than the international average, even if they’ve had no special training in it. This is because in this explicitly multilingual environment in which e.g. most supermarket products are labelled in three languages, people grow up accustomed to thinking about and in the interstices between languages where translation occurs.
And after years of translating and helping generations of students become more proficient at doing so themselves, I think that translation must reside in a part of the brain separate from (but linked to) the parts that govern writing, speaking and language-learning. When we work on translations, we strengthen the synapses that link all these lobes together. At least that’s how I picture it. I’d love to know more about the science behind this. While discussing all these things with Tobias, I mentioned one particular friend of mine who grew up bilingually in a context in which very little translation was required: everyone in the family was fluent in both languages – one of which was always spoken inside the home, the other outside it. As an adult now, she can speak and write elegantly in both, but isn’t so good at translating between them. And last night – just a few hours after this conversation – she came over while I was dreaming to make me dinner in my own kitchen. When the food was ready, she served it by ladling out spoonfuls onto the two front spiral heating coils on my stove. So this dream answered the question of what gets lost in translation: it’s the plates.
Translating strike me as being like rewriting a computer program in one language into another language that uses a different paradigm or has different functions/libraries – one has to focus on the application domain and then work out how to rework this into the chosen language (which may involve redesigning the process of how it works). Unfortunately, unless you’ve experience in this area (and understand the terminology) it probably means nothing to you as an analogy 😉 Funnily, lots of computer scientists have pursued the goal of being able to automate this (particularly for translating COBOL to other languages – there are billions of lines of COBOL out there used by big banks and insurance companies that are considered to old to maintain but to critical to ignore) – with the results generally considered to be ugly and unreadable. This also suggests that automated translation is unlikely to ever work well. Google translate will never work for programming languages because there is no archive of translated programs out there – for the same reasons Google translate doesn’t work well for ideas that haven’t been expressed in that way and translated in their archives.
The number of retranslated books also suggests that translation can’t be automated. That’s my rant for today 😉 A bit off-topic but I tried to start on-topic 😉