Let us perfrom such an “experiment”. There is an urn containing some balls such that each ball is labelled by a natural number and every natural number is used in labelling exactly one ball. Now we draw one ball from the urn. The sample space S={1,2,3…} By Probability Axiom, we have P(S)=1 Then it seems to me that the probability of drawing out some balls must be different from that of drawing out some other balls. For if the probability is the same for all balls, it can neither be positive nor zero; if it is positive, we can use another Probability Axiom to argue that P(S) is unbounded; if it is zero, we can argue that P(S)=0.