Previous: wanting and likingNext: learning from failureContents

Society of Mind

9.2 gerrymandering

We all know how accomplishment can bring satisfaction, and we tend to assume a direct connection between them. In very simple animals, where satisfaction means no more than meeting simple, basic needs, satisfaction and accomplishment must indeed be virtually the same. But in a complex human brain, a great many layers of agencies are interposed between the ones that deal with body needs and those that represent or recognize our intellectual accomplishments. Then what is the significance, in these more complicated systems, of those pleasant feelings of accomplishment and disagreeable sensations of defeat? They must be involved with how our higher-level agencies make summaries.

Suppose that you once had to send a present to a friend. You had to choose a gift and find a box in which to wrap it. Soon, each such job turned into several smaller ones — like finding strings and tying them. The only way to solve hard problems is by breaking them into smaller ones and then, when those are too difficult, dividing them in turn. So hard problems always lead to branching trees of subgoals and subproblems. To decide where resources should be applied, our problem-solving agents need simple summaries of how things are going. Let's suppose each agent's summary is based on other summaries it gets from the agents it supervises. Here is a pathological example of what could happen if every such summary were based on a simple majority decision:

When all is done, if someone asked if you enjoyed the whole experience, you might say that it was fun or terrible. But no such summary can say very much of what your agencies actually learned. Your knot-tying processes learned which actions worked and failed, your paper-folding and gift-selecting processes had other failures and accomplishments; but your overall assessment of the experience cannot reflect all those details. If the entire episode left you unhappy, you might be less inclined to give presents in the future, but that should not have much effect on what you learned about folding paper and tying string. No single sense of good or bad can reflect much of what went on inside all your agencies; too much information must be concealed. Then why does it seem so satisfactory for us to classify our feelings into positive and negative and conclude that on the whole the net effect was bad or good? True, sometimes feelings are more mixed and everything seems bittersweet, but, as we'll see, there are many reasons why we have to oversimplify.