The Philosophical Quarterly vol. 23 (1973): 241-252.

CRITICAL STUDY

RAWLS' THEORY OF JUSTICE -- II1

By R. M. Hare

A Theory of Justice. By John Rawls. (Cambridge, Mass. : Harvard U.P., 1971. Oxford : Clarendon Press, 1972. Pp. xv + 607. Price £5.00, U.S. paperback $3.95.)

(4) In this second part of my review of Rawls' book I shall be examining his normative views about justice, and in particular the question of whether he succeeds in establishing them by argument. Since my own argument will be for the most part destructive (though I shall give some hints as to how I think the job might be done better), I must start by making it clear that I am not criticizing the project of bringing philosophical argument to bear on practical questions, but only Rawls' attempted execution of it. It is my hope, as it is his, that philosophers can so clarify these matters that we shall be able to argue more cogently than we do at present about what is just or unjust. I differ from him in thinking that unless the philosophers who attempt this base themselves on a thorough understanding of the concepts used and their logical properties, which is their proper and peculiar philosophical contribution, they run the risk of doing no more for the topic of justice than journalists and politicians.

I argued at the end of Part I that the simplest form of the rational contractor theory would have the same normative consequences as a certain version of the ideal observer theory and as my own theory, and that these consequences would be of a utilitarian sort. I suggested that that was why Rawls himself does not adopt this simple version. It remains, therefore, to ask whether his more complex version (if we can ever determine exactly what it is) can bring him to the non-utilitarian conclusions that he so earnestly desires.

The crucial questions are :

  1. Who are to be included among the parties in the original position? (Let us call these, as before, POPs, and the "people in ordinary life" POLs.)
  2. What are they to be allowed to know? This I have already discussed in Part I, but I shall have to return to it.
  3. How, given these restrictions on membership and knowledge, will the POPs set about making their choice? Under this heading come questions about their use of the "principle of insufficient reason", and of their motivation (especially their aversion to risk).
We have to ask whether the answers to these three questions suffice to determine, without appeal to Rawls' own preconceptions or intuitions, what principles of justice the POPs would choose.

The principles which Rawls says they would choose are summed up by him in their "final statement" on 302/15,2 thus :

First Principle. Each person is to have an equal right to the most extensive total system of equal basic liberties compatible with a similar system of liberty for all.

Second Principle. Social and economic inequalities are to be arranged so that they are both : (a) to the greatest benefit of the least advantaged, consistent with the just savings principle, and (b) attached to offices and positions open to all under conditions of fair equality of opportunity.

He also has what he calls "priority rules" determining the relative priority that we are to give to these principles and their parts, and to the principles of justice over other principles. I shall not have space to deal with these. The most important of them is that which gives priority to liberty over the other principles. All the principles of justice are said to rest on what he calls the same "General Conception", which runs thus:
All social primary goods -- liberty and opportunity, income and wealth, and the bases of self-respect -- are to be distributed equally unless an unequal distribution of any or all of these goods is to the advantage of the least favoured (303/7).
Thus Rawls founds his theory of justice (rightly, as I think) on an account (though I think it is the wrong account) of distributive justice. This in turn is founded on a view about procedural justice in the selection of principles.

(i) The Membership of the POP Committee

It might be thought that the simplest and safest membership-rule for the assembly of POPs would be to include everybody who might be affected by their choice of principles -- i.e., "everyone who could live at some time" (139/4). This simple rule Rawls rejects, on the ground that "to conceive of the original position [thus] is to stretch fantasy too far; the conception would cease to be a natural guide to intuition" (139/5). In a book which relies so much on fantasy, this is a surprisingly weak ground. The membership is restricted in several ways :

(a) Animals are excluded, thus neatly removing them from the direct protection of principles of justice.
The discussion of this topic is full of expressions like "presumably" (505/6, 506/19), "the natural answer seems to be" (505/12) and "seems necessary to match our considered judgments" (509/25). Rawls himself says, very frankly, "Now of course none of this is literally argument" (509/32); and it would certainly not convince a determined vegetarian. It is of course difficult to include animals among the POPs (they could not make speeches in the assembly); but that is an unfortunate effect of the dramatic scenario -- which might perhaps be overcome if we could suppose that the POPs do not know whether they may not be, or be going to become, animals.
(b) By adopting the "present time of entry interpretation" (140/11 -- an opaque phrase which I have nowhere found explained) he seems to exclude all but the members of just one generation in the world's history.
He thus lays up for himself, as he believes, troubles about justice between generations. These troubles are quite gratuitous, and could easily have been solved by allowing all generations to join the assembly. Rawls himself attempts to get over the difficulty by an obscure and contrived assumption that the POPs are to be "thought of as representing continuing lines of claims, as being so to speak deputies for a kind of everlasting moral agent or institution" (128/25). "Everlasting" is toned down, two lines further on, to "over at least two generations". It might be thought that there was an important difference between two generations and eternity, and that, for securing impartiality between all generations, only eternity would do. But in fact it is strictly unnecessary to make any such assumption to secure impartiality if, as Rawls also rightly stipulates, the POPs, though they all belong to one generation, are not to know which this is (137/14, 288/9). For in that case they will not favour a particular generation, any more than they will favour particular individuals, not knowing who they themselves are or when they are born (cf. 137/23).

It is not the restriction of membership to one generation, by itself, that gets Rawls into trouble at this point, but rather this, combined with another feature of his theory to which we must now attend. He writes as if the POPs were not prescribing universally (or, as he would put it, "generally") in choosing their principles of justice, but only prescribing for their own behaviour (and possibly also for that of subsequent generations -- 13/4). From this it follows that (in default of the ad hoc restriction which he imposes) they can happily say "Let our generation, whichever it is, consume all the world's resources and leave none for succeeding generations". If, on the contrary, they were prescribing universally for all men at whatever time, and did not know at what time they were to be in the world, they could not happily universalize this prescription; for they would then be prescribing equally for their own predecessors. Thus Rawls has (characteristically, and as a consequence of his contempt for such logical tools) failed to avail himself of one of the "formal constraints of the concept of right" to which he himself has earlier drawn attention (131/14). If the POPs do not know to what generation they all belong, and are prescribing universally for the conduct of all generations, they will have (if they are rational) to adopt principles of justice which maintain impartiality between the interests of all generations. We can say that they are either prescribing for the past as well as the present and future, or choosing the principles by which they want society to be governed in the future, and hope that it has been governed in the past. I have heard rumours that Rawls himself is now attracted by this manoeuvre. Alternatively, we might take a hint from Kant via Mr. Richards,3 and suppose that, when they make their choice, they are in a noumenal atemporal green room, and do not know at what point they are to come upon the stage of time as POLs. Any difficulties which attend these modifications to the scene arise from the creakings of the stage-machinery and not from the logic of the argument, which could be set out in universal-prescriptive terms without any such machinery. That the POPs cannot affect the past (292/5) is strictly irrelevant; Rawls thinks it relevant only because he takes his machinery too seriously.

(c) The principles of justice are to be chosen from the "perspective" or "point of view" of "representative men in all relevant social positions" (96/1, cf. 64/9).
This is not strictly a membership-restriction, but it has a similar effect. Those who are on the committee are compelled to choose from the point of view of representatives of these rather gross classes (in defining which, for example, we are not allowed to differentiate between large and small farmers (96/4); this perhaps illustrates how prone Rawls is to iron out material differences between cases in the interests of "a coherent and manageable theory" (96/7) ).

Does any of these membership-restrictions enable Rawls to avoid utilitarianism? It seems not; what they can do is, not to establish or refute any particular principle of distributive justice, but only to confine or limit the class of those protected by it. They will do this if the POPs know that the membership is so limited (as, perhaps, in the case of the disfranchisement of animals); if they do not know, then the membership-restriction can make no difference at all, since each POP will have to envisage the possibility of his being any person out of the set a, b . . . n, even though the person who is going to be, say, b, or m, is not actually present at the meeting. It does not matter if some person is not present, provided that nobody knows that he is not; and for that reason a committee is strictly unnecessary; one POP would do, provided that he did not know which POL he was going to be (cf. 139/23).

However, there is one membership-restriction stipulated by Rawls that does seem to make a big difference :

(d) Only people who actually do or will exist are allowed on the committee.
In a passage which I quoted earlier, it has been implicitly laid down, on rather slender grounds, that (merely) possible people, as opposed to actual people, are to be blackballed (139/4); later it is explicitly stated that the POPs "know that they already hold a place in some particular society" (166/12) -- though of course they do not know what place in which society. This means that in Rawls' system the interests of possible people are simply not going to be taken account of. This would seem to be crucial for questions about population policy and abortion, for example. The person that the foetus would turn into if not aborted, and the people who will be born if contraception is not practised, get no say if they are not actual people -- i.e., if it is actually the case that abortion and contraception are practised. This would seem to have the curious consequence that, simply by performing an abortion, I can make sure that my act does not contravene Rawlsian justice, because I shall thereby have disfranchised the abortee. Since POPs have to have reached "the age of reason" (146/15), it looks as if the same would apply to infanticide (since merely possible adults are excluded). But 248/37 ff. and 509/19 ff. seem to support a different and more orthodox view (if we can swallow the assumption that the veil of ignorance could conceal from one of the contracting parties the fact that he is a babe in arms).

It is only this membership-restriction that enables Rawls to pronounce so easily that the average utility principle, which bids us maximize average utility, is superior to the classical utility principle, which bids us maximize total utility (166/3). To understand this, consider a possible person P whose birth would have lowered average utility but raised total utility, because his own happiness would have been less than the previous average, but more than the combined losses suffered by the others owing to his arrival. If a POP might, for all he knows, be P, he will find the classical principle more attractive; but if he knows that he cannot be P, he will prefer the average principle. This is because the classical principle would require population policies which allowed P to be born, whereas the average principle would require policies which debarred him from existence. By excluding P from the committee, and allowing this to be known, Rawls makes sure that it will disregard P's interests, and thus brings it about that, from the POPs' point of view, the average utility principle is a stronger candidate against which to pit his own principles of justice than the classical utility principle. If the exclusion is unjust, he is perhaps choosing the weaker and not the stronger opponent; but I do not think that in fact this gives him much advantage. In any case, it would not seem that this membership-restriction helps him much in his fight with utilitarianism in general.

We may note here an embarrassing consequence of the inclusion of possible people among the POPs, if Rawls' own normative principles are adopted, and if it is assumed that to have any life at all is better than not to be born. The unborn will then be his least advantaged class; and so his difference or maximin principle (see below) will require him to say that before anything is done for the rest of us, we ought to secure the birth of all these possible people.4 This would lead us to a duty of procreation on a vast scale; we could stop only when the earth would support no more people above the starvation level. But Rawls could reasonably escape this consequence by rejecting the assumption that any life is better than none; he does not need for this purpose to disfranchise merely possible people. The classical principle also has been thought by some to require a very expansive population policy; but this too will depend on what weight is put upon quality of life as opposed to mere life.

(ii) The Thickness of the Veil

We saw in Part I that Rawls does not adopt the simplest and most economical version of the "veil of ignorance" to which his POPs are to be subject -- namely that which deprives them only of knowledge of which individual each of them is to be. He prefers to say that they are not to know what properties of various sorts they are to have, and that they know only the "general" facts about society.5 That is, besides, or instead of, being ignorant of what individual each of them is to be, the POPs are ignorant of everything that cannot be described in general terms. So, for example, they are allowed to understand "political affairs and the principles of economic theory" (137/30) and "the laws of moral psychology" (138/6). There is one exception to the ban on particular knowledge: the POPs are allowed to know the "particular fact" that the POLs' society is subject to the "circumstances of justice" -- a phrase under which Rawls covers such facts as that human beings are vulnerable to attack and that natural resources are limited (127/3, 6), as are also people's powers of reasoning, memory and attention (127/27).

I suggested in Part I that §28 indicates that Rawls thinks that his thick veil of ignorance helps him to avoid utilitarianism. It is not easy to see why he thinks this; but the key to the understanding of the reasoning seems to lie in the relation between ignorance of particular facts and refusal to use the "principle of insufficient reason" (IR, see below). Rawls says that the POPs "discount estimates of likelihood that are not based on a knowledge of the particular facts" (173/6, cf. 155/25 ff.). This seems to imply that if they did have knowledge of particular facts (even if they did not know their individual places in the world constituted by those facts), they would be able to work out relative frequencies of sorts of events, and thus "the objective probabilities" of occurrences in the POL society (cf. 168/15, 171/31). Thus they would have no need to use IR, but could base their predictions of how each individual POP-turned-POL was likely to fare, given the adoption of any one set of principles of justice, on "objective probabilities". And this, he may have thought, would lead to utilitarianism (see below). I am far from certain that this is Rawls' argument. But if it is, it is fallacious. For even if the POPs had knowledge of objective probabilities of sorts of occurrences in the world, they still might not have knowledge of the probability of each of them being a particular individual POL, and so might not know how likely it was that any individual POP-turned-POL would fare well, or ill. For this, they might have to rely on IR. In other words, if the sortition which results in this POP being this POL is not subject to objective probabilities, no amount of objective probability in the distribution of welfare among the POLs will help any individual POP to know how likely it is that he will get a certain degree of welfare, unless he is allowed to use IR. So, since the knowledge of particular facts would not by itself make the use of IR unnecessary, Rawls does not gain anything by refusing to allow knowledge of these particular facts. The upshot is that everything in his argument for the rejection of utilitarianism depends on his refusal to allow the POPs to use IR. To this, therefore, we must now turn.

(iii) Insufficient Reason and Aversion to Risk

The principle of insufficient reason (IR) requires us to assign equal probabilities to two or more outcomes when we have no reason to suppose that the probability of one is greater than that of another. In some forms it leads to paradox: if we are drawing blindfold from an urn containing black, red and yellow balls in unknown proportions, it gives us different values for the probability of drawing a black ball according as we state the alternatives, which by IR will be equiprobable, as black, red and yellow or as black and coloured. If, therefore, the situation of the POPs were such that they had no basis for listing the members of the set of outcomes uniquely, then it would be hard to quarrel with Rawls' insistence that rational POPs would not use IR. However, (to use a Rawlsish phrase) it seems reasonable to suppose that a POP, knowing that he will be one of n individuals, all in some respects faring differently, will say that there are, correspondingly, n different outcomes, and will, if he uses IR, conclude that the probability of getting any one of them is 1/n. At any rate, we have at least as much reason to suppose this as Rawls has to suppose anything else.

He admits (165/35) that "if the parties are viewed as rational individuals who have no aversion to risk and who follow the principle of insufficient reason in computing likelihoods . . ., then the idea of the initial situation leads naturally to the average [utility] principle". It is interesting that in this passage he seems to imply that it is possible to have no aversion to risk and to use IR, without thereby ceasing to be rational; if so, it cannot be the rationality of the POPs that he relies on. But his more considered view seems to be that it is not rational to use IR (cf. 172/30).

It is not clear whether he thinks that the POPs have an aversion to risk. On 172/13, 26 he says, "The essential thing is not to allow the principles chosen to depend on special attitudes towards risk. . . . What must be shown is that choosing as if one had such an aversion is rational given the unique features of that situation irrespective of any special attitudes towards risk" (my italics). This seems inconsistent with 169/15 ff., where he implies that the POPs actually have, and ought to have, a reluctance to take great risks, on the inadequate ground that their descendants may reproach them if they do (might not their descendants be just as likely to say "Nothing venture, nothing win"?). It also seems inconsistent with 165/35, quoted above, if that implies that one who was not averse to risk and used IR would not be, on that score, irrational. In 172/16 he says "the parties do not know whether or not they have a characteristic aversion to taking chances"; and the reader is not to know either.6

It looks, at any rate, as if the principal weight is being placed on the rejection of IR. I am sure that reviewers more competent than me in the theory of probability will have discussed this topic at length. But the important thing to notice is that the answer to the question of whether it would be rational for the POPs, in default of objective probabilities, to use IR does not matter. Suppose we grant that it would not. The important point would then be: Why are they denied knowledge of objective probabilities? Rawls says that each of his assumptions about the POPs "should by itself be natural and plausible "(18/15), but this particular feature seems to be quite arbitrary. It is only there because it may help to lead by arguments which Rawls finds acceptable to conclusions which he finds acceptable (cf. again the remark on 141/23, "We want to define the original position so that we get the desired solution"). It is obvious that, if Rawls had wished to reach utilitarian conclusions, he could have so arranged it that the POPs were going to have their place as POLs assigned to them by means of a well-conducted mechanical or electronic lottery of the usual kind. In that case, the POPs, knowing this, would have known the objective probabilities of their getting any particular POL-role, and would have known that they were equal, and utilitarianism would have resulted. So the very most that Rawls may have done towards setting up a non-utilitarian theory of justice is to show that it is possible, if one desires, so to rig the assumptions of the theory that it does not lead straight to a utilitarian conclusion.

We next have to examine the conclusions that Rawls thinks it does lead to. He thinks that the POPs would, in choosing their principles, "maximin" -- that is to say, choose the course which has the best worst outcome. They will seek to maximize the welfare of the least-advantaged representative members of their society. It is important to distinguish this strategy from another, which I will call the "insurance" strategy; for Rawls uses arguments in favour of maximining which are really only arguments in favour of insuring against utter calamity (cf. 156/30,163/18,176/23). We insure our houses against fire because we think that a certain outcome, namely having one's house burnt down and having no money to buy another, is so calamitous that we should rule it out. This is not at all the same strategy as maximining. If the POL society were going to be affluent enough to provide a more than just acceptable standard of living for even the least advantaged, the insurance strategy would allow the POPs to purchase a very great gain for the more advantaged at the cost of a small loss for the least advantaged; but the maximin strategy would forbid this. Maximiners would end up refusing to let the man with, say, ten thousand pounds a year have any more if the man with the minimum income of nine thousand pounds received in consequence a pound or two less; but the follower of the insurance strategy would by that time have lost interest.

POPs following an insurance strategy would fix a minimum and frame their principles of justice to secure that. They would not have to know whether the minimum was feasible; all they would have to do would be to say that below it, the interests of the more advantaged were always to be sacrificed to those of the less (as in wartime rationing). Rawls has not actually given us anything to determine what minimum the POPs would fix (might they not differ on this?). But neither has he given us anything to determine the minimum which he himself requires for his argument. It is what they "can be sure of by following the maximin rule" (154/30); but how do they know what this amounts to? And if they do not know, how can they tell that anything less is "intolerable" (156/24)?

The POP game is in effect played by imagining ourselves in the original position and then choosing principles of justice. Rawls' POPs come to the decisions that they come to simply because they are replicas of Rawls himself with what altruism he has removed and a veil of ignorance clapped over his head. It is not surprising, therefore, that they reach conclusions which he can accept. If I myself play this game, I import into the original position my prejudices and inclinations, which in some respects are different from Rawls'. I have some inclination to insure against the worst calamities, in so far as this is possible.7 But I have no inclination to maximin, once the acceptable minimum is assured; after that point I feel inclined to take chances in the hope of maximizing my expectation of welfare, as I do in actual life (for example, I do not entirely refrain from investing my spare cash because I might lose it). And in certain cases I do not feel inclined to maximin even in very reduced circumstances. If, when I was a prisoner of war, a benevolent and trustworthy Japanese officer had said that he would play poker with me and, if I won enough, allow me to buy myself a ticket home through neutral territory with a safe conduct, then I should have accepted the invitation, in order to give myself a chance, however small, of freedom (the priority of liberty!) rather than forgo this chance and husband my money to buy smokes with as I languished on the Burma railway.

Thus the maximin strategy does not appeal to me as in general a good one for choices under uncertainty. Even Rawls does not go so far as to claim that. He states three features of situations which give "plausibility" to the maximin strategy (154/11). The first is the ignorance of objective probabilities; but we have seen that the imposition of this condition is entirely arbitrary. The second is that the chooser has to have "a conception of the good such that he cares very little, if anything, for what he might gain above the minimum stipend" (154/28); but this condition is clearly inapplicable, for the POPs "do not know their conception of the good" (155/15).8 The third feature is that some outcomes are "intolerable" (156/24); but this justifies only insurance, not maximining. It looks, therefore, as if Rawls has not succeeded in making his choice of strategy even "plausible". But in spite of this he says "the original position has been defined so that it is a situation in which the maximin rule applies" (155/22). We can only say "Amen ".

I do not claim to have shown that the maximin strategy is a bad one for POPs, only that Rawls has given no good reason for holding that it is a good one. The truth is that it is a wide open question how the POPs would choose; he has reduced the information available to them and about them so much that it is hard to say what they would choose, unless his own intuitions supply the lack. Rawls, however, has one recourse, and that is that the results of his theory have to tally with his "considered judgments". But they do not tally with mine. A maximin strategy would (and in Rawls does) yield principles of justice according to which it would always be just to impose any loss, however great, upon a better-off group in order to bring a gain, however small, to the least advantaged group, however affluent the latter's starting point. If intuitions are to be used, this is surely counterintuitive; at least, not many of us are as egalitarian as that.

It is to Rawls' credit that he does not avail himself of some well-worn but fallacious polemical arguments against utilitarianism. It is true that on 156/24 he insinuates, without stating, that utilitarianism could justify "if not slavery or serfdom, at least serious infractions of liberty for the sake of greater social benefits". But he very fairly admits on 26/13 that "certain common precepts of justice, particularly those which concern the protection of liberties and rights" can be accounted for in a utilitarian system as "those precepts which experience shows should be strictly respected and departed from only under exceptional circumstances if the sum of advantages is to be maximized". It would be unfair on our part to expect Rawls to explore more fully the possibilities of showing the place of common notions of justice in a utilitarian system -- that is not his enterprise. But he has not shown that there are no such possibilities; and until this has been shown, philosophers would do well to go on looking for them. It may be that they could be found, without anything like so much intuitive scaffolding as Rawls needs for his own system; it may be that the world is so constituted that to fail to inculcate and strenuously pursue principles of justice fairly closely related to some of the commonly accepted ones will result in a diminution of utility. I am indeed inclined to think that this is so; but obviously the question calls for further investigation.

In concluding this not very sympathetic notice, it must be said that a reviewer with more ample patience and leisure might possibly have done better for Rawls. I have taken a great deal of pains (and it really has been painful) trying to get hold of his ideas, but with the feeling all the time that they were slipping through my fingers. Very often, when I had found what looked like a statement of his opinion on some issue, I later found another remark which seemed to say something different. The book is extremely repetitious, and it is seldom clear whether the repetitions really are repetitions, or modifications of previously expressed views. I have drawn attention to some of these difficulties, and there are all too many others. Rawls is not to be blamed for failing to keep the whole of this huge book in his head at the same time (the only way to avoid inconsistencies when writing a book); and still less are his readers. He is to be blamed, if at all, for not attempting something more modest and doing it properly.

Many years ago he wrote some extremely promising articles, containing in germ, though without clarity, a most valuable suggestion about the form and nature of moral thought. It might have been possible to work this idea out with concision and rigour (Rawls' disciple Mr Richards has made a tolerably good job of it in his book A Theory of Reasons for Action, which is much clearer than Rawls' own book as an exposition of this type of theory). If Rawls had limited himself to, say, 300 pages, and had resolved to get his main ideas straight and express them with absolute clarity, he could have made a valuable contribution to moral philosophy. The discussions of other topics (which contain much that is of interest) could have been published separately; and that would have given him more room and more time to tighten up the main argument, which is all I have been able to consider in this review. As it is, his book is likely to waste a great deal of a good many people's time -- though they will also gain some insights, and will at any rate get some exercise. Over the years he has collected a mass of criticisms of his views, but has thought it possible to insulate himself from the effects of them by folding each in a little piece of cotton wool. If he had been more self-confident and less rigid, or even if he had had a greater sense of style, he would have either stated the criticisms clearly and answered them concisely, or else revised the views. What he has done instead is to try to incorporate the criticisms without changing the views; and that is what has made the views so hard to catch.

Corpus Christi College, Oxford

1 Part I appeared in this journal, April 1973, pp. 144-155.

2 References are to pages/lines of Rawls' text.

3 A. J. Richards, A Theory of Reasons for Action (Oxford, 1970), pp. 88-9, 310.

4 I owe this point, and much else, to Mr. Parfit.

5 It will perhaps not be necessary to enquire precisely what he means by "general" (let alone such minutiae as whether he means the same by "general" on 197/9 as he does on 137/1, or whether, rather, as seems necessary in order to preserve consistency, what are called "general facts" on 197/9 are what have been called "particular circumstances" on 137/11).

6 Nor whether the referent of 'they' is the POPs or the POLs. For if it is the POPs, Rawls is saying that a veil of ignorance can conceal from someone an aversion which he presently has (a rather steep assumption); but if the POLs, how is this condition relevant to the present motivation of the POPs as regards risk-taking?

7 It is a difficult question (too difficult for discussion here) to what extent an insurance strategy on the part of the POPs is compatible with utilitarianism. Utilitarian POPs could insure against calamities if the premium were such as to maximize their expectation of utility; and this is what many POLs try to do. To be willing to pay a high premium may simply be an indication that one attaches a high negative value to the calamity in question, and less value to the possible affluence one is sacrificing. The diminishing marginal utility of affluence is relevant here. So, on the other hand, is the sheer impossibility of insuring against some of the worst calamities (for example, that of being a person whose temperament simply prevents him being happy). A utilitarian POP might well achieve the results of an insurance strategy without the strategy, if he assigned a high acceptance-utility to a level-one (prima facie) principle enjoining compassion -- a sentiment which is perhaps better able than "justice as fairness" to make us look after the unfortunate.

8 They therefore make do with assumptions about "primary goods" (which seem in this passage, though not always, all to have monetary values). The fate of a man who was made miserable because he lacked something which he valued very much, but which was not on Rawls' list of primary goods, is therefore not even insured against.