At the moment, I'm researching learner errors using the
Cambridge Learner Corpus, to feed into a couple of new books. It's work that I always enjoy, but inevitably, like everything, there are the good bits and the not-so-good bits. Anyone who's undertaken any kind of language research (and probably any other kind of research too) will know there's the stuff that jumps out immediately and is nice and clear-cut, then there are all the annoying bits and pieces which are much messier and much more difficult to analyse and that it's very tempting just to ignore!
When it comes to learner errors, there are certain language areas that are pleasingly productive, producing lovely patterns on the screen that then translate into nice simple notes to pass on to authors to incorporate in their writing. For example, collocations - I can bring up screens of do a mistake instead of make a mistake, make a photo instead of take a photo and pass time instead of spend time. They're all nice clear-cut errors which are simple to identify and easy to warn students about. What I dread are the trickier areas of grammar; tenses, passives etc. where errors tend to be messy, compound affairs, where it's often difficult to see exactly what the student intended to say, let alone figure out quite where they've gone wrong.
One of the good things about working for myself though is that I can organise work how I want. So presented with a list of areas to research, I can pick and choose what I do when. Yesterday morning, for example, I started off with the awkward area of modals in the morning (I swear students just use modal verbs randomly!), but then gave myself the afternoon working away merrily on gerunds and infinitives - coming up with a really juicy list of incorrect verb patterns (interested to do instead of interested in doing, decide do instead of decide to do, etc). Next on my list this morning are used to and would, both of which I know are going to be messy, which is perhaps why I'm writing this instead!
Labels: Cambridge learner corpus, learner errors