Most people who have ever learnt a language will be familiar with some of the potential difficulties when translating between languages. For example, there are numerous false friends, which are not what they appear.
Take the word 'handy'. In English, it's an everyday adjective meaning 'convenient' or 'useful'. In German it gains a capital letter and becomes a noun, which in England we would translate as a 'mobile phone'. Americans would say 'cellphone' instead. We can already see two of the main issues affecting translations:
1. it's easy to jump to conclusions about what words mean, and
2. different users of the language use different vocabulary.
Computer translation tools, such as the ubiquitous Google Translate or the translation tool in MS Office have always found it hard to deal with these types of difficulties.
For example, having spent much of my working life in Berkshire, the name Reading has caused a few issues, because it looks like it has something to do with books. Just along the M4, Bath has a similar problem. Language teachers in this area can easily spot Yr 7 homework, where the pupil has tried to use the computer to do the work, when they hand in 'J'habite à lecture' or 'Ich wohne in Lesen'. Are they trying to say that they live in a library? Computers are getting better at identifying this particular issue, thanks to the capital letters, which the software has learnt to recognize as a name. That's fine for English, but in German every noun starts with a capital. So 'Essen' might refer to the German city or it could also mean 'eating' or 'food'. A human brain can use the context to work out which meaning is correct, but computers struggle to find the right meaning for phrases like 'Essen ist super'. Fortunately, Baden-Baden tends to be recognized as the spa town, so we don't end up with Bath-Bath.
In a similar vein, a 'Hamburger' in German may be the food or it could be someone from Hamburg. The same applies to Frankfurter, Wiener (from Vienna) and Berliner, as JFK famously found out.
Another potential issue wth the name of Reading when translating for different audiences is that there is more than one Reading in the world. Even putting Reading, Berks may not help identify the English town in Berkshire, as one of the places named Reading in the US is in Berks County in Pennsylvania. As much of the internet tends to default to the US, this could easily cause confusion and it takes a human brain to check that the right location has been found.
In my next blog I will look more at how translations need the correct localisation for their audience.
Comments