(I drafted this entry before I read about the Utah Court) I recently ‘attended’ a webinar about how translators can use corpora to investigate their target language. I’ve been fascinated by corpora since I first encountered the Collins Cobuild English dictionary when I was teaching English – I think it was in the 1990s. The dictionary was quite a milestone: it used a database of usage examples to show that what people say is not always the same as what language teachers say they say. Once I even tried to learn Python, after Mark Liberman said it was a good project for Christmas, but I did not get very far and suspect he has a larger brain than I have.
If you've got a free weekend or two, you could do a lot worse than to spend some time messing around with Python and NLTK — there's even an online book to guide you.
I’ve also been to a (bricks-and-mortar) seminar on corpora, but at that time I did not follow it up by preparing my own corpus. And this is where the ecpd webinar was so helpful, because it got me that far in half an hour the following evening, using free software (BootCat and AntConc). In a later post I will give a description of how it works ...Zum vollständigen Artikel