Everett's theory assumes the completeness of the description by the wave function, the linearity of the dynamics for the wave function, and multiplicity. In this paper, I argue that these three assumptions of Everett's theory may lead to the violation of psychophysical supervenience.