Rewarded actions get stamped in and repeated while punished ones fade
Evidence
What Thorndike observes—while eating those apples and smoking those cigarettes—is that the animals, who have no idea how the box actually works, initially behave almost purely randomly: biting things, nudging things. Once one of these random actions gets them out of the box, they very quickly learn to repeat that action, and become more and more capable of quickly escaping that same box in the future. “The one impulse, out of many accidental ones, which leads to pleasure,” he observes, “becomes strengthened and stamped in thereby.” … When the result of an action is “satisfying,” we tend to do it more. When on the other hand the outcome is “annoying,” we’ll do it less. The more clear the connection between action and outcome, the stronger the resulting change. Thorndike calls this idea, perhaps the most famous and durable of his career, “the law of effect.” — Brian Christian, The Alignment Problem, 538
Categorization
- corroborated: 2 independent passages support this claim — Brian Christian, The Alignment Problem, 538
- XREF: connects to: Behavior is strengthened by satisfying outcomes and weakened by annoying ones — Brian Christian, The Alignment Problem, 539; Sound incentives require losing points when progress is undone — Brian Christian, The Alignment Problem, 765; Complex behavior emerges from rewarding successive small approximations — Brian Christian, The Alignment Problem, 660 (+5 more)
Related
- Behavior is strengthened by satisfying outcomes and weakened by annoying ones (similarity 0.86)
- Dopamine encodes errors in expectation, not reward or surprise itself (similarity 0.79)
Broader Connections
- Gambling addiction may stem from intrinsic surprise rewards overriding negative expected value.
#complicates(0.70) — Brian Christian, The Alignment Problem, 909 - Executing discipline with minimum necessary force shapes behavior effectively.
#responds-to(0.55) — Jordan B. Peterson, 12 Rules for Life An Antidote to Chaos, 459