2011-09-13

Once Upon a Time

A raven, flitting around somewhere, notices a silver spot. Being curious, it pecks at it, and a seed falls out of a nearby hole. It pecks again, and another seed falls out. Now, whenever it pecks that silver spot, it expects a seed; whenever it wants a seed, it will peck the silver spot.

This is learning. Learning is entirely based on predictive pattern analysis: the ability to say, "in the past, every time X has occurred, Y has resulted," and use that information for future prediction. Advanced learning is being able to extrapolate that cause/effect into areas not identical but similar to the original situation: maybe the spot is gold, and maybe it's a piece of candy instead of a seed.

In simpler terms, the act of learning is an act of building stories - even ones far less interesting than "Jack and the Beanstalk." We build, in our heads, a narrative course of actions that we can use to inform ourselves as well as others in the future. Our brains are very good at building these stories: being good at stories - being good at learning - provided a heritable survival advantage, so our ancestors who were better building stories tended to survive more.

It's important, however, to recognize when our story-building goes awry: when we don't have enough data or experience to build a realistic cause/effect model, or when we highlight the wrong thing as being the cause or effect. This is also something we tend to be good at, unfortunately, because it's a side effect of constantly looking for stories.

The most critical part of any story we build is that we be open to modifying, expanding, or removing it based on experience in order to keep it useful: remember, if the story can't predict anything, it's not useful and may even be counter-productive as we waste energy in support of a pattern that doesn't exist. If the raven finds that it's not just silver spots, but square silver spots, that result in seeds, continually pecking a round silver spot and expecting a seed is a waste of energy; if it goes on long enough, the raven may starve.

Stories are extremely powerful, and while they can be useful, we need to constantly check and validate them to prevent them from becoming detrimental, both to ourselves and to society.