You can simplify the problem into straight behaviorism.… I’d have to look up which book I read this in (Don’t Shoot the Dog, maybe?), but there is a game you can teach dogs, dolphins, etc where you give them a box or something, and only reward them for novel behavior. So you reward them the first time they push it with their nose, but not any subsequent times. This seems to “teach” creativity, in that animals that play this game regularly get good at quickly coming up with unusual actions.
Note: I’m not saying the CORRECT thing to do is ignore all the substeps, conditions, pre-requisites, etc and go straight to “just reward the thing you want”. It was just a cute anecdote that seemed relevant.
While this might produce novel behavior, I doubt that it trains insightful behavior. The heuristics for doing something novel are likely not the same as the heuristics for actual integrating different ideas into insights.
You can simplify the problem into straight behaviorism.… I’d have to look up which book I read this in (Don’t Shoot the Dog, maybe?), but there is a game you can teach dogs, dolphins, etc where you give them a box or something, and only reward them for novel behavior. So you reward them the first time they push it with their nose, but not any subsequent times. This seems to “teach” creativity, in that animals that play this game regularly get good at quickly coming up with unusual actions.
Note: I’m not saying the CORRECT thing to do is ignore all the substeps, conditions, pre-requisites, etc and go straight to “just reward the thing you want”. It was just a cute anecdote that seemed relevant.
While this might produce novel behavior, I doubt that it trains insightful behavior. The heuristics for doing something novel are likely not the same as the heuristics for actual integrating different ideas into insights.