People Teach with Rewards and Punishments as Communication not Reinforcements

Carrots and sticks motivate behavior, and people can teach new behaviors to other organisms, such as children or non-human animals, by tapping into their reward learning mechanisms. But how people teach with reward and punishment depends on their expectations about the learner. We examine how people teach using reward and punishment by contrasting two hypotheses. The first is evaluative feedback as reinforcement, where rewards and punishments are used to shape learner behavior through reinforcement learning mechanisms. The second is evaluative feedback as communication, where rewards and punishments are used to signal target behavior to a learning agent reasoning about a teacher’s pedagogical goals. We present formalizations of learning from these two teaching strategies based on computational frameworks for reinforcement learning. Our analysis based on these models motivates a simple interactive teaching paradigm that distinguishes between the two teaching hypotheses. Across three sets of experiments, we find that people are strongly biased to use evaluative feedback communicatively rather than as reinforcement.