Abstract: Based on intellectualism, many mainstream researches anticipate understanding other minds as “one mind thinking about the other minds”. In recent years, however, this view has been challenged by the “direct social perception” (DSP). According to DSP, problems of accessing to other minds should not be regarded as “mind thinking about minds”, instead, the explanandum in intersubjectivity must include “embodied mind perceiving embodied minds”. As a substantive theory of DSP, embodied simulation (ES) emphasizes on reusing the core concept of mental simulation and posits a motion simulation as an automatic, unconscious, pre-reflective mechanism to go straight to the other minds, which is generated by the activities of mirror neurons. I propose five problems that should be fulfilled in order to support ES, and brake on direct access to other minds: (1) the similarity of brain-body system is not necessary for social interaction; (2) the priming of ES is affected by plenty of stimulus rather than the lack of stimulus; (3) activation of own motor program is not necessary for action understanding; (4) the accounts for classic experiments of mirror neuron is compatible with mentalizing accounts; (5) mirror neurons originate from sensorimotor associative learning rather than gene-based natural selection or adaptation.

Key words: other minds problem丨direct social perception, mirror neurons丨embodied simulation丨mentalizing