[From: Bruce Nevin (Thu 931014 10:58:49 EDT)]
Avery Andrews 931014.1314 --
It seems to me that what is dysfunctional in mood disorders is that
reorganization is prevented from reaching a resolution.
It may be that the cycling from manic to depressive and back is a
metaconflict, as it were, a relation between program conflict and system
conflict, and that each in some way prevents the other from being
resolved through reorganization.
My architect friend enjoyed at least some aspects of it, or was
fascinated by it. After a period of stability he would stop taking
lithium, claiming that it deadened his creativity, but my sense was that
he got some kind of rush from the indulgent destructiveness involved. He
could get away with doing very hurtful things because, after all, he was
crazy. After he did some possibly homicidal things driving his car,
involving his wife and stepson as they walked along the roadside, and the
police dragged him in impotent red-faced rage from court in a hammerlock
and put him in Bridgewater (not a pleasant place) instead of MacLean
(where he had burned up most of his trust money from his father for prior
treatment) he seems to have found it less fun to be crazy, and so far as
I know has been less adventurous. So in his case I felt that he was
controlling some perception of sticking it to old Dad, who he described
was a self-centered bastard, and to do this he appears to me actively to
have courted his mania. I don't know about depression with him.
In other cases, with which I am not so familiar, it smells to me
something like Bateson's double bind. An image from an R. Cobb cartoon:
dejected fellow sitting on the pavement near the end of an alley posted
with signs that say "one way" (pointing inward to the end of the alley)
and "no loitering", and "trespassers will be prosecuted" (or some such
combination). So that's another possibility: depression in the double
bind, mania in the attempt to reorganize out of it.
Avery Andrews 931014.1530 --
Your modified GATHER agent would have to be able to perceive objects at a
distance very much greater than its avoidance reference. For the
analog to vision, this is the distance to the horizon, which varies but
can be in principle infinite.
There is something like this in its present perception of where the goal
is. It knows that regardless of intervening objects? I don't have the
program (not owning a mouse). I think that in the agent's perceptual
universe objects other than the goal don't even exist so far as the agent
is concerned until they and the agent are as close as the agent's
avoidance distance. They can't be perceived as intervening between the
agent and the goal (so far as the agent is concerned) until they can be
perceived. So the new controlled perception is a relation between the
long-range perception of the goal goal (somewhat mysterious to me) and
long-range perception of objects other than the goal.
Bruce
bn@bbn.com