Jan 17, 2008 22:11
Controlling causes leads to statistical probabilities of outcomes. The more you can control, the higher probability of your outcome or effect. The problem is in trying to manage all the causes in human interaction. To be an effective manager of causes, the fewer causes one has to manage the greater one's ability is to manage them. When trying to manage so many causes, especially in a situation where there are more causes than usual, more going on than usual, we tend to not get what we want because when we try to control *all* of the causes, there is an even *smaller* statistical probability that we will get the effect we want, we try to devote so much attention to so many causes, and we become so much less efficient. Not only less efficient, but we surely miss some causes, and screw up others, and in the process we even end up producing other undesirable causes without knowing it---the effect has such a high probability of coming up far from what we expected, and out of our control. The only constant is chaos. We can control (1) chaos, and (2) not chaos, with a continuum in between. Usually. I think. There are different kinds of causes, too, with different durations.
When things get too complicated in human interaction, a rule should be to stop everything, time-out, and actively search and strive for simplicity.
effect,
simplicity,
psychology,
control,
interaction,
minimalism,
cause