[From Bill Powers (980212.1545 MST)]
Tracy Harms (980212.13) --
Hi, Bill. As you see I've decided to jump back into your pond.
Hi again. Plenty of room in here.
In keeping with Gary Cziko's
line of theorizing we can expect that there is a significant randomness
involved in all *learning*. (We wouldn't want to exempt adults from
Lifelong learning is a nice idea, but it slows down a lot as we get older.
Presumably we find solutions to various problems that reduce the need for
reorganization. If we went through life having to reorganize just as fast
at the end as at the beginning, one would wonder just how useful a process
Unfortunately, this doesn't mean we end up with the _best_ solutions --
just good enough solutions to avoid most of the pain.
Don't miss my main point, though. If you look at the way people build
things, fix things, accomplish things, it becomes perfectly obvious that
the random component in behavior is extremely small. It's really
astonishing that anyone could miss all this precision and regularity, and
think that a correlation of 0.6 is good.