A very short entry today.
I’ve previously mentioned the work by Perkins et. al., as part of Project Zero at the Harvard Graduate School of Education (from quite a while ago) in which it was suggested that understanding is a performance. I’ve tried in vain to get a full-bodied theory from this work; it just doesn’t seem to be out there. If I had to guess, I’d think that the authors figured out that they didn’t have a defensible theory and abandoned the project. They have ceased academic publishing on the idea, but continue to sell books and PD materials to schools…
Anyway, here’s a quick thought experiment to dispel the notion that understanding can be reduced to performance. Imagine watching two people play chess. Can you tell something–it doesn’t need to be a complete picture–about what the players understand by watching their moves?
Well, first, you’ll need to understand something about chess yourself to make this judgment (careful: this is in danger of circularity!). So now you, using your chess knowledge and understanding, try to form a judgment about what the players understand about the position on the board. Are they beginners? Are they good coffee-house players? Masters?
But wait. What about context? Who are these people? How do you know that they aren’t actors playing through a script? You don’t. Unless you have access to their internal processes–their thinking–you are in a position of significant uncertainty. We don’t even need the fanciful possibility of acting or fraud. Perhaps the players have memorized some moves from master games, and can play them so long as the opponent doesn’t stray from the memorized lines. (The longer the game goes on, the less likely this becomes.)
This is a real problem in modern chess tournaments. From time to time accusations of cheating come forth–perhaps the player is getting electronic assistance during bathroom breaks. How do you decide if the player understands the game well enough to make those moves? The standard approach is laughably simple. You give the player some sample positions in test conditions and get him/her to explain what moves to make and why.
Natural language dialogue appears to be the most obvious and sensible approach to the question of student understanding. This is not a theory but is a starting point. More to follow.