Korva is a scroll of the internet’s best ideas.
Reward close, then raise the bar
Prompts
A prompt is a cue or stimulus that encourages the desired behavior. Prompting and shaping work together to change behavior in an iterative way.
How shaping works
When shaping begins, any approximation of the target response is reinforced. Once the approximation occurs regularly, the criterion gets stricter, so the actual behavior has to match the desired behavior more closely. The criteria keep narrowing until the specific target behavior is performed.
Where it's used
This tactic is often the preferred method of developing a habit gradually and of training animals to perform a specific behavior.
How did Skinner shape pigeons for war?
In World War II, Skinner led Project Pigeon, which trained pigeons to peck at an image of a target so that their pecks could steer a missile. The birds were shaped step by step toward pecking the target image. The National Defense Research Committee funded the work, and the military cancelled it in 1944 without using it in combat.
How do dolphin trainers use shaping?
Dolphin trainers reward successive approximations of a behavior, such as a jump a little higher each time, using a whistle to mark the exact moment of success. Karen Pryor helped bring the method into mainstream animal training in the 1960s while working with dolphins at Sea Life Park in Hawaii. She later adapted it for dogs as clicker training.
Why does the standard tighten only gradually?
Behavior varies from one attempt to the next, and shaping uses that variation: the reward picks out the attempts closest to the target. If the standard jumps too far, no attempt qualifies and the behavior can fade. Skinner described this selection of behavior by its consequences as parallel to natural selection in a 1981 article in the journal Science.
More from Shane
