The paper aims to show, first, that O'Regan's and Noë's Sensorimotor Theory of Vision and Visual Experiences suffers from circularity, and that evidence from empirical research within perception psychology unequivocally invalidates their theory. Secondly, to show that the circularity in O'Regan's and Noë's theory of vision and in other general causal and functional theories of perception (i.e. Gibson's and Marr's theories of perception) is the inevitable consequence of mutually conflicting assumption of Cartesian dualism underlying these theories. The paper concludes by outlining the consequences of this conflict of assumptions for psychological theories of perception.