JOINT ATTENTION IN THE MIMETIC CONTEXT — WHAT IS A "MIMETIC SAME"?

Mimesis should be meaningful imitation that reproduces not only superficial behaviors but also the behavioral intentions underlying the behaviors. We propose mimetic learning based on a self-evaluation function (MLSE), which enables the imitator robot to assess the model robot's behavioral intention by referring to the rate of change of its self-evaluation function. Here, the success of the imitation depends on a coevolving mechanism that consists of two learning contexts: identical and situational contexts. Experimental results show that MLSE enables the imitator robot to reproduce behavioral patterns by taking into account the model robot's behavioral intention. Additionally, the imitator robot succeeded in reproducing the model robot's intention even if there was a slight difference between the model and the imitator robot's body size. In this chapter, this phenomenon is regarded as a special case of joint attention in the mimetic context. As a result, this chapter aims to answer the question "What is a mimetic same?"