But if this is so, the contractees will certainly not decide upon a strategysuch as "no advantage to me is acceptable unless it is to the advantage of thoseworse off." For, while this may be reasonable in a situation of uncertainty, itis not reasonable in a situation with information. In the latter case, a rationalegoist will adopt a new strategy that will maximize his interests alo

lity. An egoist will not bind himself toa policy which does not promise to maximize his future expectations; so it isodd for Rawls to suggest that "they understand further that the principles pro-posed and acknowledged on this occasion are to be binding on future occasions.... They will be bound by it in future cases the circumstances of which areunknown.

The 'difference' principle is not sufficient either, since it gives no specifi-cation of the size of the inequality allowed in comparison with the amount ofthe advantage provided. Any inequality, no matter how great, would be justi-fied by any advantage, no matter how slight, to the badly off. Rawls does notcall this perfectly just, but he considers it 'just' all the sam

"Given the circumstances of the original position, it isrational for a man to choose as if he were designing a society in which hisenemy is to assign him his place. Thus, in particular, given the complete lackof knowledge . . . it is rational to be conservative and so to choose in accor-dance with an analogue of the maximin principle.... Moreover, it seems clearhow the principle of utility can be interpreted: it is the analogue of the La-placean principle for choice uncert

The second, or 'difference' principle is much too strong. It specifies thatinequalities are arbitrary unless they will work out to the advantage of therepresentative man. But this Pareto-inclusive rule prevents any redistributionwhich lowers the wealth of the rich from being considered just.