Sitting her comfortably in Virginia, I’m quite happy to bathe in an English-only sea of words. So much of the content on the Internet is in English these days that it’s easier and easier to ignore everything interesting taking place on online islands where English either isn’t spoken, or just isn’t used on websites.
As a European historian, I know how bad this is, because it’s hard enough for me to keep up with the literature in my field in English, much less what people are writing in German, Czech, and Slovak, not to mention Hungarian (which I can’t read), or French (which I can sort of read). Some of the best works in my field have never seen the light of day in English and probably never will. This means I have to force myself to be attentive to what’s happening in those other languages and, if we’re being honest, I’ve not been as attentive as I ought to be.
Sadly, I’m sure I’m not alone. How many of us really have the time to keep up on our reading as it is? And unless we are especially diligent, it’s almost always the work published in languages other than English that slips further down on the reading list.
All of this is by way of explaining why it is that as we begin to ramp up Global Perspectives on Digital History, we’re so focused on figuring out how best to publish our “journal” in multiple languages simultaneously. Unfortunately, the auto-translation algorithms aren’t where we need them to be today, so we’re going to have to do a fair amount of working around the short comings in the software.
For instance, I recently ran the introduction to my forthcoming book on teaching history in the digital age through the Google Translate algorithm to covert it from English to German. The resulting German translation was workmanlike and mostly correct in the purest technical sense of the word “correct.” However, virtually all the nuance in what I wrote disappeared–and I’m not a sophisticated enough German speaker to notice everything that was wrong. But even with my more limited skills, I could tell that the resulting translation was a bit like what one could expect a second year student to produce for a class assignment. Almost all the words were correct, but the sense of the sentences was often all wrong.
On the good side, Google Translate produced this translation of a 17 page text in less than five minutes. Now we have to figure out whether it’s easier to begin with such a rough and ready translation and then correct it, or to just start from scratch. My hope is that the former will work, if only because I can imagine the mechanisms under which we might work with such an approach. If we have to start from scratch on each item, the level of human involvement could be prohibitive.
Over the next month or so, we’re going to experiment with some more sophisticated translation software. Once I have samples to show, I’ll report back.