Before beginning to mount this critique I should like to contrast how certain matters that are relevant to it appeared to me once with how they appear to me now. Having been raised in a family devoted to the Canadian Communist Party² I was in my teens a pretty orthodox Marxist; I had enthusiastically embraced the theory that was offered to me. From the perspective of that theory as I then understood it I was contemptuous of various then current and now still familiar apologies for or defenses of economic inequality. Some of those defenses of economic inequality can be called normative and the others can be called factual. The normative defenses endorse inequality. They represent it as just. The factual ones by contrast do not deny (or affirm) that inequality is unjust. They say that inequality whether it be just unjust or neither is unavoidable.

A prominent factual defense of inequality traces it to a supposedly ineradicable human selfishness. This defense says that inequality is ensured by something as original to human nature as sin is on the Christian view of original sin: people are by nature selfish whether or not that is like being a sinner a bad thing to be and inequality is an unavoidable result of that selfishness whether or not inequality is unjust.

Being “selfish” here means desiring things for oneself and for those in one's immediate circle and being disposed to act on that desire even when the consequence is that one has (much) more than other people do and could otherwise have had. In a strong version of the selfishness hypothesis it is in the very nature of the relevant desire that it is a desire to have more than other people—that is it is a desire both that be on one of the higher rungs of the ladder of inequality and that others be on lower rungs. If I am (in this sense) strongly selfish then I want to have more than another does not (merely) because I'll then have more than I otherwise would but because I (at least also) fundamentally want to be above him. In a weaker version of the selfishness hypothesis what a person desires to have will merely as a matter of fact (in virtue that is of the fact that resources are finite) put her above others. Being above as such is not as it is in the stronger version the goal of our self-seeking but it remains the outcome of that quest for those who are gifted enough or lucky enough to enjoy success in the pursuit of self-interested desire.

The selfishness defense of inequality has two premises. First a human-nature premise: that people are by nature selfish. And second a sociological premise: that if people are selfish (whether by nature or otherwise) then equality is impossible to achieve and/or to sustain.

Now I used to reject both premises of the selfishness defense of inequality. But I have become sympathetic to the sociological premise. On that more in a moment First let me explain why I rejected both premises in the past.

I rejected the human-nature premise for the Marxist reason that as I believed social structures extensively shape the structure of motivation. There exists so I thought no underlying human nature which could be qualified as straightforwardly unselfish or selfish or selfish in some fixed degree. I do not mean that I thought there was no underlying human nature at all; I would have considered that to be a heretically unmaterialist proposition.³ (We are after all animals with a particular biology and a psychology induced by it.) But I did think that human nature was quite plastic with respect to motivation. There were in a sense I nevertheless recognized truths about how selfish human beings are by nature—such as for example the truth as to how propitious circumstances would have to be for humans to be unselfish in their attitude and behavior. Or more generally the truth about selfishness and human nature would be disclosed by the (variable) height of a line on a graph with circumstances on the horizontal and degrees of manifested selfish orientation on the vertical. The right way to characterize human nature would be as a function with circumstances as arguments and forms of behavior as values. And all you needed to believe in order to deny that human nature was selfish in a sense that threatened the egalitarian project was what I indeed believed: that people would be relevantly unselfish in propitious circumstances and that such circumstances were accessible.⁴

So I rejected the human-nature premise of the selfishness defense of inequality. But I independently rejected its sociological premise. I thought that even if people were by nature selfish the conclusion that inequality was inescapable would not follow—and was in fact false—because just as social structure was sovereign over motivation to the detriment of the first premise of the selfishness argument so too structure was sovereign over the upshot of motivation: even if and when people were indeed selfish be it in virtue of their unvarying nature or otherwise the rules governing their interaction could nevertheless prevent their selfishness from issuing in inequality.⁵ Such rules might for example be enforced by a great majority who were themselves selfishly or at any rate not altruistically motivated and who would be at the short end of inequality in the absence of such rules.

I remain skeptical of the human-nature premise of the selfishness defense of inequality for something like the old reasons. But I am no longer so skeptical of the sociological premise. I no longer think that even granting selfishness in motivation structure can block inequality in upshot.⁶ And this change of view is highly consequential. Thus for example if people are by now irreversibly selfish (not by nature but) as a result of capitalist history then so I now think structure alone could not suffice to deliver equality in the face of that selfishness.⁷ Even on reasonably sunny views about the limits of human nature itself capitalist history would have thrown us into a cul-de-sac from which we could not exit and regain the road to socialism.

My present greater friendliness to the claim of the italicized nostrum which we can call the Christian social nostrum is supported by reflection on work that I have done in the past few years on the Rawlsian justification of economic inequality. Rawls says that inequality is justified when it has the effect that those who are worst off are better off than they would be if the inequality were removed. Inequality is (not only justified but) just for Rawls when and because it is necessary to enhance the position of the worst off and he thinks it typically is necessary to that end in virtue of the benign influence on productive motivation of the material incentives associated with economic inequality.

Rawls presents that as a normative justification of inequality—the sort of justification of it that is which seeks to present (certain form of) inequality as just. But so I shall argue a close investigation of the principal mechanism that makes it true when it is true that economic inequality benefits the badly off reveals that the Rawlsian case for inequality is better characterized as a merely factual defense of it. Despite what Rawls himself says he does not show that incentive-based inequality is just on his own conception of justice but at most that it is regrettably unavoidable if not tout court⁸ then at least if we do not want to depress everyone's condition.⁹ Rawls's purportedly normative defense of inequality exposes itself on properly insistent interrogation as a merely factual defense of it. That is because as we shall see an anti-egalitarian selfishness must be attributed to the more productive as part of the explanation for why inequality is necessary to the extent that it is indeed necessary.

In his early essay “Justice as Fairness” we find the following passage:

There are in my view a number of obscurities and infelicities in this passage and further ones in the paragraph from which it is drawn.¹¹ But I am at present concerned only to remark on the extremely interesting fact that Rawls deleted the clause I have placed in italics when fourteen years later he published an otherwise substantially (and nearly verbally) identical paragraph in A Theory of Justice. The Theory sentence that corresponds to the first sentence in the passage above reads as follows: “If for example these inequalities set up various incentives which succeed in eliciting more productive efforts a person in the original position may look upon them as necessary to cover the costs of training and to encourage effective performance.”¹² It will follow from my argument against Rawls that if unequalizing incentives are truly necessary from the point of view of the interests of the badly off¹³ then they are necessary only because of infirmities in human nature that are more or less affirmed in the “Justice as Fairness” passage but which gain no mention as such in the corresponding Theory passage.¹⁴

It is as though both the Rawls of 1957 and the Rawls of 1971 agree with Bernard Mandeville (and Adam Smith) that “private vices” make for “publick benefits”—that (in other words) human selfishness can be made to benefit everyone—but that the Rawls of 1971 is unwilling to acknowledge that it is indeed vices which are in question.¹⁵ I agree with Mandeville—and Adam Bede¹⁶—that that's what they are.

Now that slogan as it stands is vague but I mean something reasonably precise by it here: that principles of distributive justice—principles that is about the just distribution of benefits and burdens in society—apply wherever else they do to people's legally unconstrained choices. Those principles so I claim apply to the choices that people make within the legally coercive structures to which so everyone would agree principles of justice (also) apply. (In speaking of the choices people make within coercive structures I do not include the choice whether or not to comply with the rules of such structures—to which choice one again so everyone would agree principles of justice (also) apply). I mean rather the choices left open by those rules because neither enjoined nor forbidden by them.)

The italicized slogan that I have appropriated here is widely used by feminists.¹⁷ More importantly however the idea itself which I have here used the slogan to formulate and which I have tried to explicate above is a feminist idea. Notice however that in briefly explaining the idea that I shall defend I have not mentioned relations between men and women in particular or the issue of sexism. We can distinguish between the substance and the form of the feminist critique of standard ideas about justice and it is the form of it which is of prime concern to me here¹⁸ even though I also endorse its substance.

The substance of the feminist critique is that standard liberal theory of justice and the theory of Rawls in particular unjustifiably ignore an unjust division of labor and unjust power relations within families (whose legal structure may show no sexism at all). That is the key point of the feminist critique from a political point of view. But the (often merely implicit) form of the feminist critique which we get when we abstract from its gender-centered content is that choices not regulated by the law fall within the primary purview of justice and that is the key lesson of the critique from a theoretical point of view.

Because I believe that the personal is political in the specified sense I reject Rawls's view that principles of justice apply only to what he calls the “basic structure” of society. Feminists have noticed that Rawls wobbles across the course of his writings on the matter of whether or not the family belongs to the basic structure and is therefore in his view a site at which principles of justice apply. I shall argue that Rawls's wobble on this matter is not a case of mere indecision which could readily be resolved in favor of inclusion of the family within the basic structure; that is the view of Susan Okin¹⁹ and in my opinion she is wrong on this count. I shall show (in section 2 of Lecture 9) that Rawls cannot admit the family into the basic structure of society without abandoning his insistence that it is to the basic structure only that principles of distributive justice apply. In supposing that he could include family relations Okin shows a failure to grasp the form of the feminist critique of Rawls.

Sections 1 and 2 of Lecture 9 offer independent replies to that basic structure objection. I show in section 1 that the objection is inconsistent with many statements by Rawls about the role of principles of justice in a just society. I then allow that the discordant statements may be dropped from the Rawlsian canon and in section 2 I reply afresh to the basic-structure objection by showing that no defensible account of what the basic structure is allows Rawls to insist that the principles which apply to it do not apply to choices within it. I conclude that my original criticism of Rawls rests vindicated against the particular objection at issue here. (Section 3 of Lecture 9 comments on the implications of my position for the moral blame ability of individuals whose choices violate principles of justice. Section 4 explores the distinction between coercive and noncoercive institutions which plays a key role in the argument of section 2.)

My criticism of Rawls is of his application of the difference principle. The principle says in one of its formulations²¹ that inequalities are just if and only if they are necessary to make the worst off people in society better off than they would otherwise be. I have no quarrel here with the difference principle itself²² but I disagree sharply with Rawls on the matter of which inequalities pass the test for justifying inequality that it sets and therefore about how much inequality passes that test. In my view there is hardly any serious inequality that satisfies the requirement set by the difference principle when it is conceived as Rawls himself proposes to conceive it²³ as regulating the affairs of a society whose members themselves accept that principle. If I am right affirmation of the difference principle implies that justice requires (virtually)²⁴ unqualified equality itself as opposed to the “deep inequalities” in initial life chances with which Rawls thinks justice to be consistent.²⁵

It is commonly thought for example by Rawls that the difference principle licenses an argument for inequality which centers on the device of material incentives. The idea is that talented people will produce more than they otherwise would if and only if they are paid more than an ordinary wage and some of the extra which they will then produce can be recruited on behalf of the worst off.²⁶ The inequality consequent on differential material incentives is said to be justified within the terms of the difference principle for so it is said that inequality benefits the worst off people: the inequality is necessary for them to be positioned as well as they are however paltry their position may nevertheless be.

Now before I mount my criticism of this argument a caveat is necessary with respect to the terms in which it is expressed. The argument focuses on a choice enjoyed by well-placed people who command a high salary in a market economy: they can choose to work more or less hard and also to work at this occupation rather than that one and for this employer rather than that one in accordance with how well they are remunerated. These well-placed people in the foregoing standard presentation of the argument are designated as “the talented” and for reasons to be given presently I shall so designate them throughout my criticism of the argument. Even so these fortunate people need not be thought to be talented in any sense of that word which implies something more than a capacity for high-market earnings for the argument to possess whatever force it has. All that need be true of them is that they are so positioned that happily for them they do command a high salary and they can vary their productivity according to exactly how high it is. But as far as the incentives argument is concerned their happy position could be due to circumstances that are entirely accidental relative to whatever kind of natural or even socially induced endowment they possess. One need not think that the average dishwasher's endowment of strength flair ingenuity and so forth falls below that of the average chief executive to accept the argument's message. One no doubt does need to think some such thing to agree with the different argument which justifies rewards to well-placed people in whole or in part as a fair return to exercise of unusual ability but Rawls's theory is built around his rejection of such desert considerations. Nor does Rawls believe that the enhanced rewards are justified because extra contribution warrants extra reward on ground of proper reciprocity. They are justified in his view purely because they elicit more productive performance.

I nevertheless persist in designating the relevant individuals as “the talented” because to object that they are not actually especially talented anyway is to enter an empirical claim which is both contentious and in context misleading since it would give the impression that it should matter to our assessment of the incentives argument whether or not well-placed people merit the contestable designation. The particular criticism of the incentives argument that I shall develop is best understood in its specificity when the apparently concessive word “talented” is used; it does not indicate a concession on the factual question of how top people in a market society get to be where they are. My use of the argument's own terms show the strength of my critique of it: the critique stands even if we make generous assumptions about how well-placed people secured their powerful market positions. It is moreover especially appropriate to make such assumptions here since the Rawlsian difference principle is lexically secondary to his principle that fair equality of opportunity has been enforced with respect to the attainment of desired positions; if anything ensures that those who occupy them possess superior creative endowment that does. (Which is not to say that it indeed ensures that. It is consistent with fair equality of opportunity that what principally distinguishes top people is superior cunning and/or prodigious aggressivity and nothing more admirable.)

Now for the following reasons I believe that the incentives argument for inequality represents a distorted application of the difference principle even though it is its most familiar and perhaps even its most persuasive application. Either the relevant talented people themselves affirm the difference principle or they do not. That is either they themselves believe that inequalities are unjust if they are not necessary to make the badly off better off or they do not believe that to be a dictate of justice. If they do not believe it then their society is not just in the appropriate Rawlsian sense for a society is just according to Rawls only if its members themselves affirm and uphold the correct principles of justice. The difference principle might be appealed to in justification of a government's toleration or promotion of inequality in a society in which the talented do not themselves accept it but it then justifies a public policy of inequality in a society some members of which—the talented—do not share community with the rest.²⁷ Their behavior is then taken as fixed or parametric a datum vis-à-vis a principle applied to it from without rather than as itself answerable to that principle. That is not how principles of justice operate in a just society as Rawls specifies the concept: within his terms one may distinguish between a just society and a just government—one that is which applies just principles to a society whose members may not themselves accept those principles.

So we turn to the second and only remaining possibility which is that the talented people do affirm the difference principle—that as Rawls says they apply the principles of justice in their daily life and achieve a sense of their own justice in doing so.²⁸ But they can then be asked why in the light of their own belief in the principle they require more pay than the untalented get for work which may indeed demand special talent but which is not specially unpleasant (for no such consideration enters the Rawlsian justification of incentives-derived inequality). The talented can be asked whether the extra they get is necessary to enhance the position of the worst off which is the only thing according to the difference principle that could justify it. Is it necessary tout court—that is independently of human will so that with all the will in the world removal of inequality would make everyone worse off? Or is it necessary only insofar as the talented would decide to produce less than they now do or to not take up posts where they are in special demand if inequality were removed (by for example income taxation which redistributes to fully egalitarian effect)?²⁹

Talented people who affirm the difference principle would find those questions hard to handle. For they could not claim in self-justification at the bar of the difference principle that their high rewards are necessary to enhance the position of the worst off since in the standard case³⁰ it is they themselves who make those rewards necessary through their own unwillingness to work for ordinary rewards as productively as they do for exceptionally high ones an unwillingness which ensures that the untalented get less than they otherwise would. High rewards are therefore necessary only because the choices of talented people are not appropriately informed by the difference principle.³¹

Apart then from the very special cases in which the talented literally could not—as opposed to the normal case where they (merely) would not—perform as productively as they do without superior remuneration the difference principle can justify inequality only in a society where not everyone accepts that very principle. It therefore cannot justify inequality in the appropriate Rawlsian way.

Now this conclusion about what it means to accept and implement the difference principle implies that the justice of a society is not exclusively a function of its legislative structure of its legally imperative rules but is also a function of the choices people make within those rules. The standard (and in my view misguided) Rawlsian application of the difference principle can be modeled as follows. There is a market economy in which all agents seek to maximize their own gains and there is a Rawlsian state that selects a tax function on income that maximizes the income return to the worst off people within the constraint that because of the self-seeking motivation of the talented a fully equalizing taxation system would make everyone worse off than one which is less than fully equalizing. But this double-minded modeling of the implementation of the difference principle with citizens inspired by justice endorsing a stat policy which plays a tax game against (some of) them in their manifestation as self-seeking economic agents is wholly out of accord with the (sound) Rawlsian requirement on a just society that its citizens themselves willingly submit to the standard of justice embodied in the difference principle. A society that is just within the terms of the difference principle so we may conclude requires not simply just coercive rules but also an ethos of justice that informs individual choices. In the absence of such an ethos inequalities will obtain that are not necessary to enhance the condition of the worst off: the required ethos promotes a distribution more just than what the rules of the economic game by themselves can secure. And what is required is indeed an ethos a structure of response lodged in the motivations that inform everyday life not only because it is impossible to design rules of egalitarian economic choice conformity with which can always be checked³² but also because it would severely compromise liberty if people were required forever to consult such rules even supposing that appropriate applicable rules could be formulated.³³

To be sure one might imagine in the abstract a set of coercive rules so finely tuned that universally self-interested choices within them would raise the worst off to as high a position as any other pattern of choices would produce. Where coercive rules had and were known to have such a character agents could choose self-interestedly and be confident that the results of their choices would satisfy an appropriately uncompromising interpretation of the difference principle. In that (imaginary) case the only ethos necessary for difference-principle justice would be willing obedience to the relevant rules an ethos which Rawls expressly requires. But the vast economics literature on incentive-compatibility teaches that rules of the contemplated perfect kind cannot be designed. Accordingly as things actually are the required ethos must as I have argued guide choice within the rules and not merely direct agents to obey them. (I should emphasize that this is not so because it is in general true that the point of the rules governing an activity must be aimed at when agents pursue that activity in good faith. Every competitive sport represents a counterexample to that generalization. But my argument for the conclusion stated above did not rest on the stated false generalization.)³⁴

Call that “the basic-structure objection.” Now before I develop it further and then reply to it I wish to point out that there is an important ambiguity in the concept of the basic structure as that is wielded by Rawlsians. The ambiguity turns on whether the Rawlsian basic structur includes only coercive aspects of the social order or also conventions and usages that are deeply entrenched but not legally or literally coercive . I shall return to that ambiguity in section 2 of Lecture 9 and I shall show that it shipwrecks not only the basic-structure objection but also the whole approach to justice that Rawls has taught so many to pursue. But for the time being I shall ignore the fatal ambiguity and I shall take the phrase “basic structure” as it appears in the basic-structure objection as denoting some sort of structure be it legally coercive or not but whose key feature for the purposes of the objection is that it is indeed a structure—that is a framework of rules within which choices are made as opposed to a set of choices and/or actions. Accordingly my Rawlsian critic would say whatever structure precisely the basic structure is the objection stands that my criticism of the incentives argument misapplies principles devised for a structure of individual choices and actions.

In further clarification of the polemical position let me make a back-ground point about the difference between Rawls and me with respect to the site or sites at which principles of justice apply. My own fundamental concern is neither the basic structure of society in any sense nor people's individual choices but the pattern of benefits and burdens in society—that is neither a structure in which choice occurs nor a set of choices but the upshot of structure and choices alike. My concern is distributive justice by which I uneccentrically mean justice (and its lack) in the distribution of benefits and burdens to individuals. My root belief is that there is injustice in distribution when inequality of goods reflects not such things as differences in the arduousness of different people's labors or people's different preferences and choices with respect to income and leisure but myriad forms of lucky and unlucky circumstance. Such differences of advantage are a function of the structure and of people's choices within it; so I am concerned secondarily with both of those.

Now Rawls could say that his concern too is distributive justice in the specified sense but that for him distributive justice obtains just in case the allocation of benefits and burdens in society results from actions which display full conformity with the rules of a just basic structure.³⁸ When full compliance with the rules of a just basic structure obtains it follows on Rawls's view that there is no scope for (further) personal justice and injustice which affects distributive justice whether it be by enhancing it or by reducing it. There is Rawls would of course readily agree scope within a just structure for distribution-affecting meanness and generosity;³⁹ but generosity though it would alter the distribution and might make it more equal than it would otherwise be could not make it more just than it would otherwise be for it would then be doing the impossible—to wit enhancing the justice of what is already established as a (perfectly) just distribution by virtue merely of the just structure in conformity with which it is produced. But as I have indicated I believe that there is scope for relevant (relevant that is because it affects justice in distribution) personal justice and injustice within a just structure and indeed that it is impossible to achieve distributive justice by purely structural means.

In discussion of my claim (see section 5 above) that social justice requires a social ethos which inspires uncoerced equality-supporting choice Ronald Dworkin suggested⁴⁰ that a Rawlsian government might be thought to be charged with a duty under the difference principle of promoting such an ethos. Dworkin's suggestion was intended to support Rawls against me by diminishing the difference between Rawls's position and my own and thereby reducing the reach of my criticism of him. I do not know what Rawls's response to Dworkin's proposal would be but one thing is clear: Rawls could not say that to the extent that the indicated ethos-promoting policy failed society would as a result be less just than if the policy had been more successful. Accordingly if Dworkin is right that Rawlsian justice requires government to promote an ethos friendly to equality it could not be for the sake of making society more disributively just that it was doing so even though it would be for the sake of making its distribution more equal. The following threefold conjunction which is an inescapable consequence of Rawls's position on Dworkin's not unnatural interpretation of it is strikingly incongruous: (1) the difference principle is an egalitarian principle of distributive justice; (2) it imposes on government a duty to promote an egalitarian ethos; (3) it is not for the sake of enhancing distributive justice in society that it is required to promote that ethos. Dworkin's attempt to reduce the distance between Rawls's position and my own threatens to expose the former as incoherent.

Now before I mount my two replies to the basic-structure objection a brief conceptual digression is required in clarification of the relationship between a just society in Rawls's (and my own) understanding of that idea and a just distribution in my (non-Rawlsian) understanding of that different idea. A just society here is one whose citizens affirm and act upon the correct principles of justice but justice in distribution as here defined consists in a certain egalitarian profile of rewards It follows that as a matter of logical possibility a just distribution might obtain in a society that is not itself just.

To illustrate this possibility imagine a society whose ethos though not inspired by a belief in equality nevertheless induces an equal distribution. An example of such an ethos would be an intense Protestant ethic which is indifferent to equality (on earth) as such but whose stress on self-denial hard work and investment of assets surplus to needs somehow (despite the asceticism in it) makes the worst off as well off as it is possible for them to be. Such an ethos achieves difference-principle justice in distribution but agents informed by it would not be motivated by the difference principle and they could not therefore themselves be accounted just within the terms of that principle.⁴¹ Under the specifications that have been introduced here this Protestant society would not be just despite the fact that it displays a just distribution. We might say of the society that it is accidentally but not constitutively just. But whatever phrasings we may prefer the important thing is to distinguish “society” and “distribution” as candidate subjects of the predicate “just.” (And it bears mentioning that in contemporary practice an ethos that achieves difference-principle equality would almost certainly have to be equality-inspired; the accident of a non-equality-inspired ethos producing the right result is at least in modern times highly unlikely. The Protestantism I have described in this paragraph is utterly fantastic at lest for our day.)

Less arresting is the opposite case in which people strive to govern their behavior by (what are in fact) just principles but ignorance or the obduracy of wholly external circumstance or collective action problems or self-defeatingness of the kinds studied by Derek Parfit⁴² or something else which I have not thought of frustrates their intention so that the distribution remains unjust. It would perhaps be peculiar to call such a society just and neither Rawls nor I need do so: justice in citizens was posited above as a necessary condition of a just society.

However we resolve the secondary and largely verbal complications raised in this digression the point will stand⁴³ that an ethos which informs choice within just rules is necessary in a society committed to the difference principle. My argument for that conclusion relied not on aspects of my conception of justice which distinguish it from Rawls's but on our shared conception of what a just society is. The fact that distributive justice as I conceive it (causally) requires an ethos (be it merely equality-promoting such as our imaginary Protestantism or also equality-inspired) that goes beyond conformity to just rules was not a premise in my argument against Rawls. The argument of section 5 turned essentially on my understanding of Rawls's well-considered requirement that the citizens of a just society are themselves just. The basic-structure objection challenges that understanding.