What could possibly be wrong with loads of data on how language is used by native and (increasingly) non-native speakers? Well, nothing – if a bit of common sense/ a pinch of salt is added before it gets to the classroom. Sadly, that is too often not the case.
For example, someone who has been teaching IELTS or FCE for a couple of years could easily come up with better materials than those based on the official Cambridge corpora of typical student errors, only about 10% of which is relevant to my students and teachable. I’d also go with about 10% for how much of “spoken grammar” is of use to people who don’t live in English speaking countries or have a British or American boss.
Academic Vocabulary in Use is a much better model, but it still has about as much useless stuff for my students as the other, less scientifically put together, books in the In Use series. And although I’m as guilty of overusing it as anyone, the Academic Word List is an incredibly blunt tool that was responsible for most of the things that were wrong in a Garnet academic vocabulary book I reviewed.
On the plus side, corpora have given us evidence to back up spending less time on nearly pointless grammar points like Past Perfect and backshifting in Reported Speech, but I’m sure I’m not the only one who had already come to that conclusion from my own “personal data set”.
So, my conclusion is “So far, more bad than good”. A practical suggestion for how it could be done better coming up in my next post.