Complex class, day 1
We're studying three key questions: (1) how do interactions give rise to patterns of complex behavior; (2) how do we describe complex systems; and (3) how do complex systems arise through evolution. Our instructor, Yaneer Bar-Yam, is unafraid of building models in Excel spreadsheets on the fly in public. (He should really use RuntimeRevolution.) We had foxes eating sheep, droplets absorbing other droplets, lots of pixillating pointilism.
One simple point that I enjoyed today: all of statistics is based on bell curves and non-interacting elements. That's just not how the world works. Parts do interact. There was a good story from a participant about being told in a research project about the behavior of college students that he/she was not allowed to assume that the students could talk to another. Nope– that would invalidate all the assumptions under which the research grant had been obtained.
I also love fractals. I'm not sure where that will get me in life, but they're great. Same with neural networks and the strengthening/weakening of synapses.
There's a lot of modeling going on in this course. From my perspective, models are just fine as projects. But they don't really help at all with hard questions. Take policy debates. No one will ever be explicit about what their interests are or how much they care. So you'll never be able to say: here's a model of how things may work out given everyone's desires. All your assumptions are infinitely attackable.
And models are based on rules being applied to helpless colored squares, as far as I can tell. My question today was: why can't the elements make rules for themselves, based on what they're learning? No answer yet.