. "In other words, one should postulate only as many kinds of m" . . .