Although Rawls says often enough that the two principles of justice govern only justice in basic structure he also says three things which tell against that restriction. This means that in each case he must either uphold the restriction and repudiate the comment in question or maintain the comment and drop the restriction.¹

First Rawls says that when the difference principle is satisfied society displays fraternity in a particularly strong sense: its citizens do not want “to have greater advantages unless this is to the benefit of others who are less well off.… Members of a family commonly do not wish to gain unless they can do so in ways that further the interests of the rest. Now wanting to act on the difference principle has precisely this consequence .”² But fraternity of that strong kind is not realized when all the justice delivered by the difference principle comes from the basic structure and therefore whatever people's motivations in economic interaction may be. Wanting not “to gain unless they can do so in ways that further the interests of the rest” is incompatible with the self-interested motivation of market maximizers which the difference principle in its purely structural interpretation does not condemn.³

Second Rawls says that the worst off in a society governed by the difference principle can bear their inferior position with dignity since they know that no improvement of it is possible that they would lose under any less unequal dispensation. Yet that is false if justice releas to structure alone since it might then be necessary for the worst off to occupy their relatively low place only because the choices of the better off tend strongly against equality. Why should the fact that no purely structurally induced improvement in their position is possible suffice to guarantee the dignity of the worst off when their position might be very inferior indeed because of unlimited self-seekingness in the economic choices of well-placed people?⁴ Suppose for example that (as politicians now routinely claim) raising rates of income taxation with a view to enhancing benefits for the badly off would be counterproductive since the higher rates would induce severe disincentive effects on the productivity of the better off. Would awareness of that truth contribute to a sense of dignity on the part of the badly off?

Third Rawls says that people in a just society act with a sense of justice from the principles of justice in their daily lives; they strive to apply those principles in their own choices. And they do so because they “have a desire to express their nature as free and equal moral person and this they do most adequately by acting from the principles that they would acknowledge in the original position. When all strive to comply with these principles and each succeeds then individually and collectively their nature as moral persons is most fully realized and with it their individual and collective good.”⁵ But why do they have to act from the principles of justice and “apply” them “as their circumstances require”⁶ if it suffices for justice that they choose as they please within a structure designed to effect an implementation of those principles? And how can they without a redolence of hypocrisy celebrate the full realization of their natures as moral persons when they know that they are out for the most that they can get in the market?

Now as I said these inconsistencies are not decisive against Rawls. For in each case he could stand pat on his restriction of justice to basic structure and give up or weaken the remark that produces the inconsistency. And that is indeed what he is disposed to do at least with respect to the third inconsistency that I have noted. He said⁷ that A Theory of Justice erred by in some respects treating the two principles as defining a comprehensive conception of justice;⁸ he would accordingly now drop the high-pitched homily which constitutes the text to note 5 above. But this accommodation carries a cost: it means that the ideals of dignity fraternity and full realization of people's moral natures can no longer be said to be delivered by Rawlsian justice.⁹

A major fault line in the Rawlsian architectonic not only wrecks the basic-structure objection but also produces a dilemma for Rawls's view of the subject¹¹ of justice—a dilemma from which I can imagine no way out. The fault line exposes itself when we ask the apparently simple question: What (exactly) is the basic structure? For there is a fatal ambiguity in Rawls's specification of the basic structure and an associated discrepancy between his criterion for what justice judges and his desire to exclude the effects of structure-consistent personal choice from the purview of its judgment.

The basic structure the primary subject of justice is always said by Rawls to be a set of institutions and so he infers the principles of justice do not judge the actions of people within (just) institutions whose rules they observe. But it is seriously unclear which institutions are supposed to qualify as part of the basic structure. Sometimes it appears that coercive (in the legal sense) institutions exhaust it or better that institutions belong to it only insofar as they are (legally) coercive.¹² In this widespread interpretation of what Rawls intends by the “basic structure” of a society that structure is legible in the provisions of its constitution in such specific legislation as may be required to implement those provisions and in further legislation and policy which are of central importance but which resist formulation in the constitution itself.¹³ The basic structure in this first understanding of it is so one might say the broad coercive outline of society which determines in a relatively fixed and general way what people may and must do in advance of legislation that is optional relative to the principles of justice and irrespective of the constraints and opportunities created and destroyed by the choices that people make within the given basic structure so understood.

Yet it is quite unclear that the basic structure is always to be so understood in exclusively coercive terms within the Rawlsian texts. For Rawls often says that the basic structure consists of the major social institutions and he does not put a particular accent on coercion when he announces that specification of the basic structure.¹⁴ In this second reading of what it is institutions belong to the basic structure whose structuring can depend far less on law than on convention usage and expectation; a signal example is the family which Rawls sometimes includes in the basic structure and sometimes does not.¹⁵ But once the line is crossed from coercive ordering to the noncoercive ordering of society by rules and conventions of accepted practice then the ambit of justice can no longer exclude chosen behavior since at least in certain cases the prescriptions that constitute informal structure (think again of the family) are bound up with the choices that people customarily make.

“Bound up with” is vague so let me explain how I mean it here. One can certainly speak of the structure of the family and it is not identical with the choices that people customarily make within it; but it is nevertheless impossible to claim that the principles of justice which apply to family structure do not apply to day-to-day choices within it. For consider the following contrast. The coercive structure let us provisionally accept¹⁶ arises independently of people's quotidian choices: it is formed by those specialized choices which legislate the law of the land. But the noncoercive structure of the family has the character it does only because of the choices that its members routinely make. The constraints and pressures that sustain the noncoercive structure reside in the dispositions of agents which are actualized as and when those agents choose to act in a constraining or pressuring way. With respect to coercive structure one may perhaps fairly readily distinguish the choices which institute and sustain structure from the choices that occur within it.¹⁷ But with respect to informal structure that distinction though conceptually intelligible is compromised extensionally. When A chooses to conform to the prevailing usages the pressure on B to do so is reinforced; and no such pressure exists the very usages themselves do not exist in the absence of conformity to them. Structure and choice remain distinguishable but not from the point of view of the applicability to them of principles of justice.

Now since that is so since appropriately conforming behavior is (at least partly) constitutive of noncoercive structure it follows that the only way of sustaining the basic-structure objection against my claim that the difference principle condemns maximizing economic behavior (and more generally of sustaining the restriction of justice to the basic structure against the insistence that the personal too is political) is by holding fast to a purely coercive specification of the basic structure. But that way out is not open to Rawls because of a further characterization that he offers of the basic structure: this is where the discrepancy adverted to in the second paragraph of this section appears. For Rawls says that “the basic structure is the primary subject of justice because its effects are so profound and present from the start.”¹⁸ Nor is this further characterization of the basic structure optional: it is needed to explain why it is primary as far as justice is concerned. Yet it is false that only the coercive structure causes profound effects as the example of the family once again reminds us:¹⁹ if the “values [that] govern the basic [political] framework of social life” thereby govern “the very groundwork of our existence”²⁰ so too do the values that govern our nurture and conduct in the family. Accordingly if Rawls retreats to coercive structure he contradicts his own criterion for what justice judges and he lands himself with an arbitrarily narrow definition of his subject matter. So he must let other structure in and that means as we have seen letting chosen behavior in. What is more even if behavior did not as I claim it does partly constitute the noncoercive structure it will come in by direct appeal to the profundity-of-effect criterion for what justice governs. So for example we need not decide whether or not a regular practice of favoring sons over daughters in the matter of providing higher education forms part of the structure of the family to condemn it as unjust under that criterion.²¹

Given then his stated rationale²² for exclusive focus on the basic structure—and what other rationale could there be for calling it the primary subject of justice?—Rawls is in a dilemma. For he must either admit application of the principles of justice to (legally optional) social practices and indeed to patterns of personal choice that are not legally prescribed both because they are the substance of those practices and because they are similarly profound in effect in which case the restriction of justice to structure in any sense collapses; or if he restricts his concern to the coercive structure only then he saddles himself with a purely arbitrary delineation of his subject matter. I now illustrate this dilemma by reference to the two items that have already figured in Lectures 8 and 9: the family and the market economy.

Family structure is fateful for the benefits and burdens that redound to different people and in particular to people of different sexes where “family structure” includes the socially constructed expectations which lie on husband and wife. And such expectations are sexist and unjust if for example they direct the woman in a family where both spouses work outside the home to carry a greater burden of domestic tasks. Yet such expectations need not be supported by the law for them to possess informal coercive force: sexist family structure is consistent with sex-neutral family law. Here then is a circumstance outside the basic structure as that would be coercively defined which profoundly affects people's life-chances through the choices people make in response to the stated expectations which are in turn sustained by those choices.²³ Yet Rawls must say on pain of giving up the basic-structure objection that (legally uncoerced) family structure and behavior have no implications for justice in the sense of “justice” in which the basic structure has implications for justice since they are not a consequence of the formal coercive order. But that implication of the stated position is perfectly incredible: no such differentiating sense is available.

John Stuart Mill taught us to recognize that informal social pressure can restrict liberty as much as formal coercive law does. And the family example shows that informal pressure is as relevant to distributive justice as it is to liberty. One reason why the rules of the basic structure when it is coercively defined do not by themselves determine the justice of the distributive upshot is that by virtue of circumstances that are relevantly independent of coercive rules some people have much more power than others to determine what happens within those rules.

The second illustration of discrepancy between what coercive structure commands and what profoundly affects the distribution of benefits and burdens is my own point about incentives. Maximinizing legislation²⁴ and hence a coercive basic structure that satisfies the difference principle are consistent with a maximizing ethos across society which under many conditions will produce severe inequalities and a meager level of provision for the worst off; yet both have to be declared just by Rawls if he stays with a coercive conception of what justice judges. And that implication is surely perfectly incredible.

Rawls cannot deny the difference between the coercively defined basic structure and that which produces major distributive consequences: the coercively defined basic structure is only an instance of the latter. Yet he must to retain his position on justice and personal choice restrict the ambit of justice to what a coercive basic structure produces. But so I have (by implication) asked: Why should we care so disproportionately about the coercive basic structure when the major reason for caring about it its impact on people's lives is also a reason for caring about informal structure and patterns of personal choice? To the extent that we care about coercive structure because it is fateful with regard to benefits and burdens we must care equally about the ethic that sustains gender inequality and inegalitarian incentives. And the similarity of our reasons for caring about these matters will make it lame to say: Ah but only the caring about coercive structure is a caring about justice in a certain distinguishable sense. That thought is I submit incapable of coherent elaboration.²⁵

My original criticism of the incentives argument ran in brief as follows:

The basic-structure objection to that criticism is of this form:

My preliminary reply to the basic-structure objection says:

Let me emphasize that my rebuttal of the basic-structure objection does not itself establish that the difference principle properly evaluates not only state policy but everyday economic choice. The argument for that conclusion is given in my “Incentives” lectures and summarized in section 5 of Lecture 8 above. I do not say that because everyday choice cannot be as the basic-structure objection says it is beyond the reach of justice simply because it is everyday choice it then follows that everyday economic choice is indeed within its reach; that would be a non sequitur. I say rather that it is no objection to my argument for the claim that justice evaluates everyday economic choice that everyday choice is (in general) beyond the reach of justice since it is not.

This point about the structure of my argument is easily missed so let me explain it in a different way. I have not tried to show that a robust structure/choice distinction cannot be sustained in the case of the economy—that claim is false. What I argued is that choices within the economic structure cannot be placed outside the primary purview of justice on the ground that the only thing (quite generally) which is within its primary purview is structure. The family case refutes that argument. That refutation doesn't I would agree exclude treating economic choices like the choices of a game player who obeys the rules (and therefore plays not unjustly) while trying to score as many points as he can.²⁶ What excludes that what defeats that analogy is the argument summarized in section 5 of Lecture 8 above.

But that raises a huge question with respect to blame. The injustice in distribution which reflects personal choices within a just coercive structure can plainly not be blamed on that structure itself nor therefore on whoever legislated that structure. Must it then be blamed in our two examples on men²⁷ and on acquisitive people respectively?

I shall presently address and answer that question about blame; but before I do so I wish to explain why I could remain silent in the face of it—why that is my argument in criticism of Rawls's restricted application of the principles of justice requires no judgment about blaming individual choosers. The conclusion of my argument is that the principles of justice apply not only to coercive rules but also to the pattern in people's (legally) uncoerced choices. Now if we judge a certain set of rules to be just or unjust we need not add as pendant to that judgment that those who legislated the rules in question should be praised or blamed for what they did.²⁸ And something analogous applies when we come to see that the ambit of justice covers the pattern of choices in a society. We can believe whatever we are inclined to do about how responsible and/or culpable people are for their choices and that includes believing that they are not responsible and/or culpable for them at all while affirming the view on which I insist: that the pattern in such choices is relevant to how just or unjust a society is.

That said let me now face the question of how blameable individuals are. It would be inappropriate to answer it here by first declaring my position if indeed I have one on the philosophical problem of the freedom of the will. Instead I shall answer the question about blame on the pre-philosophical assumptions which inform our ordinary judgments about when and how much blame is appropriate. On such assumptions we should avoid two opposite mistakes about how culpable chauvinistic men and self-seeking high fliers are. One is the mistake of saying: there is no ground for blaming these people as individuals for they simply participate in an accepted social practice however tawdry or awful that practice may be. That is a mistake since people do have choices: it is indeed only their choices that reproduce social practices; and some moreover choose against the grain of nurture habit and self-interest. But one also must not say: look how each of these people shamefully decides to behave so badly. That too is unbalanced since although there exists personal choice there is heavy social conditioning behind it and it can cost individuals a lot to depart from the prescribed and/or permitted ways. If we care about social justice we have to look at four things: the coercive structure other structures the social ethos and the choices of individuals; and judgment on the last of those must be informed by awareness of the power of the others. So for example a properly sensitive appreciation of these matters allows one to hold that an acquisitive ethos is profoundly unjust in its effects without holding that those who are gripped by it are commensurately unjust. It is essential to apply principles of justice to dominant patterns in social behavior—that as it were is where the action is—but it doesn't follow that we should have a persecuting attitude to the people who display that behavior. We might have good reason to exonerate the perpetrators of injustice but we should not deny or apologize for the injustice itself.²⁹

On an extreme view which I do not accept but need not reject a typical husband in a thoroughly sexist society—one that is in which families in their overwhelming majority display an unjust division of domestic labor—is literally incapable of revising his behavior or capable of revising it only at the cost of cracking up to nobody's benefit. But even if that is true of typical husbands we know it to be false of husbands in general. It is a plain empirical fact that some husbands are capable of revising their behavior since some husbands have done so in response to feminist criticism. These husbands we could say were moral pioneers. They made a path which becomes easier and easier to follow as more and more people follow it until social pressures are so altered that it becomes harder to stick to sexist ways than to abandon them. That is a central way in which a social ethos changes. Or for another example consider the recent rise in environmental awareness. At first only a few people bother to save and recycle their paper plastic and so forth and they seem freaky because they do so. Then more people start doing that and finally it becomes not only difficult not to do it but easy to do it. It is pretty easy to discharge burdens that have become part of the normal round of everybody's life. Expectations determine behavior behavior determines expectations which determine behavior and so on.

Are there circumstances in which a similar incremental process could occur with respect to economic behavior? I do not know. But I do know that universal maximizing is by no means a necessary feature of a market economy. For all that much of its industry was state-owned the United Kingdom from 1945 to 1951 had a market economy. But salary differentials were nothing like as great as they were to become or as they were then in the United States. Yet so I hazard when British executives making five times what their workers did met American counterparts making fifteen times what their (anyhow better paid) workers did many of the British executives would not have felt: we should press for more. For there was a social ethos of reconstruction after war an ethos of common project that moderated desire for personal gain. It is not for a philosopher to delimit the conditions under which such—and even more egalitarian—ethican prevail. But a philosopher can say that a maximizing ethos is not a necessary feature of society or even of a market society and that to the extent that such an ethos prevails satisfaction of the difference principle is prejudiced.

In 1988 the ratio of top-executive salaries to production-worker wages was 6.5 to 1 in West Germany and 17.5 to 1 in the United States.³⁰ Since it is not plausible to think that Germany's lesser inequality was a disincentive to productivity since it is plausible to think that an ethos which was relatively friendly to equality³¹ protected German productivity in the face of relatively modest material incentives we can conclude that the said ethos caused the worst paid to be better paid than they would have been under a different culture of reward. It follows on my view of the matter that the difference principle was better realized in Germany in 1988 than it would have been if its culture of reward had been more similar to that of the United States.³² But Rawls cannot say that since the smaller inequality that benefited the less well off in Germany was a matter not of law but of ethos. I think that Rawls's inability to regard Germany as having done comparatively well with respect to the difference principle is a grave defect in his conception of the site of distributive justice.

The legally coercive structure of society functions in two ways. It prevents people from doing things by erecting insurmountable barriers (fences police lines prison walls and so forth) and it deters people from doing things by ensuring that certain forms of unprevented behavior carry an (appreciable risk of) penalty.³³ The second (deterrent) aspect of coercive structure may be described counterfactually in terms of what would or might happen to someone who elects the forbidden behavior: knowledge of the relevant counterfactual truths motivates the complying citizen's choices.

Not much pure prevention goes on within the informal structure of society; not none but not much. Locking errant teenagers in their rooms would represent an instance of pure prevention which if predictable for determinate behavior would count as part of a society's informal structure: it would be a rule in accordance with which that society operates. That being set aside informal structure manifests itself in predictable sanctions such as criticism disapproval anger refusal of future cooperation ostracism beating (of for example spouses who refuse sexual service) and so on.

Finally to complete this conceptual review the ethos of a society is the set of sentiments and attitudes in virtue of which its normal practices and informal pressures are what they are.

Now the pressures that sustain the informal structure lack force save insofar as there is a normal practice of compliance with the rules they enforce. That is especially true of that great majority of those pressures (beating does not belong to that majority) which have a moral coloring: criticism and disapproval are ineffective when they come from the mouths of those who ask others not to do what they do themselves. To be sure that is not a conceptual truth but a social-psychological one. Even so it enables us to say that what people ordinarily do supports and partly constitutes (again not conceptually but in effect) the informal structure of society in such a way that it makes no sense to pass judgments of justice on that structure while withholding such judgment from the behavior that supports and constitutes it; that point is crucial to the anti-Rawlsian inference presented in section 2 above.³⁴ Informal structure is not a behavioral pattern but a set of rules yet the two are so closely related that so one might say they are merely categorially different. Accordingly so I argued to include (as one must) informal structure within the basic structure is to countenance behavior too as a primary object of judgments of justice.

Now two truths about legally coercive structure might be thought to cast doubt on the contrast that I allowed between it and informal structure in section 2 above. First although the legally coercive structure of society is indeed discernible in the ordinances of society's political constitution and law those ordinances count as delineating it only on condition that they enjoy a broad measure of compliance.³⁵ And second legally coercive structure achieves its intended social effect only in and through the actions which constitute compliance with its rules.

In light of those truths it might be thought that the dilemma I posed for Rawls (see section 2 above) and by means of which I sought to defeat his claim that justice judges structure as opposed to the actions of agents was misframed. For I said against that claim that the required opposition between structure and actions works for coercive structure only with respect to which a relevantly strong distinction can be drawn between structure-sustaining and structure-conforming action but that coercive structure could not reasonably be thought to exhaust the structure falling within the purview of justice. Accordingly so I concluded justice must also judge (at least some) everyday actions.

The truths rehearsed two paragraphs back challenge that articulation of the distinction between coercive structure and action within it. They thereby also challenge the contrast drawn in section 2 between two relationships: that between coercive structure and action and that between informal structure and action. And to the extent that the first relationship is more like the second the first horn of the dilemma I posed for Rawls becomes sharper than it was. It is sharp not only for the reason I gave namely the consideration about “profound effect” but also for the same reason that the second horn is sharp namely that everyday behavior is too germane to the very existence of (even) coercive structure to be immune to the principles of justice that apply to the coercive structure.

The distinction vis-à-vis action between coercive and informal structure so I judge is more blurred than section 2 allowed—not of course because informal structure is more separable from action than I originally claimed but because coercive structure is less separable from it than I originally allowed. Accordingly even if the dilemma constructed in section 2 was for the stated reasons misframed the upshot would hardly be congenial to Rawls's position—that justice judges structure rather than actions—it would rather be congenial to my own rejection of it. But I wish to emphasize that this putative strengthening of my argument is not essential. In my opinion the argument was strong enough already.³⁶