ON THE NATURE OF INFANTILE AUTISM