Computer Aided Language Learning - Principal Parts

[Jonathan Robie]

I could well be confused about your aims. But for what it's worth...

Or I could be confused about my aims ;->

I think we're trying to do more than one thing, with more than one kind of drill. And our exact approach is definitely still evolving. We're 1 month into thinking about this.

I think you need to decide whether this is a communicative type tool or an analytic one.
I can see great value to having components of a verb color coded. But that is a decidedly analytic way of learning verbs. Our language processors do away with the need for analysis because they are such absorbent sponges and because they intuitively recognize patterns. But giving loads of comprehensible input is necessary for us to apprehend the patterns. Drills, such as I think you are aiming at, could provide some of that comprehensible input and forced output.

I see value in all these kinds of drills. But to master the forms of the verbs, I think it's important to present examples in a systematic way so that the patterns become apparent. At least for those who do not have a master teacher present, and there aren't enough master teachers who use a communicative approach.

In the drills I'm talking about now, nothing is color coded.

I also think it's important to teach that there is an overarching pattern that is shared in the grand scheme, as Funk illustrates with this table:

patterns.png (50.04 KiB) Viewed 541 times

I'd suggest that you choose which purpose your drills master to serve, to aid in easy analysis or to provide comprehensible input. If input, then colors are a distraction to how you set up the drill and a distraction to the learner. If analysis, then aiming at artificial replication of genuine communication is a distraction.

I think we need both kinds of drills. But perhaps the drills that teach the forms for analysis can be simpler by not trying to mix in extra words, that makes them a lot easier to generate. I do think they should be mixed with exercises involving real sentences. If we draw them from the corpus, a human being needs to select them for readability, or we need a metric to do so.