As a political theory, liberalism is known for the emphasis it places on the status and dignity of the individual. As long as liberal theory has existed, however, there have been those who reject this individualist picture and advance instead the claims of the community. These communitarians object to the liberal emphasis on voluntary attachments and on individual freedoms (of speech, of movement, of association), the exercise of which tends to break down community bonds and produce separation, alienation, and loneliness. Far from enabling the ‘pursuit of happiness,’ the critics argue, these rights in fact tend to enable individuals only to lead emptier and less fulfilling lives.[1]

But some critics of liberalism argue at a much more fundamental level. They are concerned, not with the empirical evolution of liberal societies, but with the metaphysical assumptions of liberalism as they concern the nature of the self. Is the self prior to its ends and attachments, existing separately from them and capable of detaching itself when necessary? Or do these ends and attachments to some extent constitute the self, making decisions to abide by them a matter of maintaining one’s coherent identity—and enabling the self to combine its identity, united in a nation or society, with others who share its ends?

The latter position is forwarded by Michael Sandel in Liberalism and the Limits of Justice. Sandel’s work is directed specifically at the theory of John Rawls, whose account he takes to be the best expression of the foundations of liberalism. The purpose of this essay is not to defend Rawls from Sandel’s attack, but rather to determine whether Sandel’s critique serves to defeat any liberal theory that would meet Rawls’ goals. Its approach is inspired by a suggestion made by Rawls, that the difference principle could be used not only as a means of assessing institutions but also of judging individual actions. Rawls terms this theory “rightness as fairness,” to correspond with the doctrine of “justice as fairness” which he describes at length.[2]

I contend that the criticisms made by Sandel against a liberal theory of justice can potentially be translated into criticisms of a liberal theory of rightness. At the very least, such a theory would make two claims: first, that rightness enjoys a certain priority in our considerations (just as justice “is the first virtue of social institutions”), and second, that rightness is appropriately considered as applying to the actions of an individual human being—in Sandel’s terms, an “antecedently individuated” self. If Sandel’s denial of the status of justice as a “first virtue” and of the “antecedently individuated” self are successful in the context of justice, I see no reason why they should not be equally successful in the context of rightness. If, however, the arguments fail in a discussion of rightness, then it may be possible to re-translate their failure back to the discussion of justice, to find that the liberal claims of justice still stand.

The fruitfulness of this approach, and the connections between theories of rightness and those of justice, can immediately be seen in addressing Sandel’s response to the claim of priority. According to Sandel, virtues partly depend for their value on the circumstances in which they are exercised. Physical courage, for instance, is more valuable on the field of battle than on a traffic-clogged highway. Similarly, the virtue of justice need not always be primary, but might instead be a conditional virtue, depending on the existence of such “circumstances of justice” as limited generosity and moderate scarcity. In some conditions (within a loving family, perhaps, or among friends) the virtues of benevolence or generosity of spirit might be preferred to a strict equality of what is received and what is due; in others (a world of unlimited resources), justice is simply unneeded. On this account, justice is only a “remedial” virtue, there to step in when social bonds are eroded or other virtues fail. How, then, can we agree with Rawls’ pronouncement that justice “is the first virtue of social institutions”?

Sandel thinks Rawls would answer this argument through what he calls a “deontological rejoinder.” It does not matter whether or not the circumstances of justice hold in the actual, empirical world; they are assumed to hold in the original position, and thus the principles that it produces will call for justice as the first virtue. However, if this rejoinder is accepted, it raises the further questions of why the circumstances are assumed in the original position, and why an original position containing such assumptions can make binding claims about social institutions. These questions are pursued further by Sandel to challenge the possibility of an “Archimedean point” from which institutions can be impartially and non-arbitrarily assessed.

Yet when viewed from the standpoint of rightness, rather than justice, no such “deontological rejoinder” is needed. Translated into the sphere of rightness, the liberal makes only the considerably less controversial claim that rightness is the first virtue of actions. (Indeed, it seems to be part of the very definition of rightness that it be the first virtue.) Moreover, any account of the circumstances of rightness will not identify situations where rightness is or is not appropriate, but rather situations where there is or is not something right to be done. Consider Christine Korsgaard’s discussion of Kant’s categorical imperative, that one should act only on maxims that one could at the same time will as a universal law. How could it be moral to give money to the poor, one might ask, since if everyone did so there would be no poor to whom one could give money? The answer is that there is no contradiction; if the poor were already well provided for, one’s aim in adopting the maxim of benevolence would not have been frustrated, but rather would have already been fulfilled.

This is why Kant held, with Aristotle, that there is no need for justice between friends; not because the circumstances that provide for moral interests do not apply (for Kant, there are no such circumstances), but because the relevant moral interests are already fulfilled.[3] A friendship is an arena within which individuals already value each other as ends in themselves. When such relationships begin to break down, and the individuals no longer value each other as ends, specifically “moral” instructions might then be needed; such instructions are unnecessary in the ideal case not because they are inappropriate, but because their aims have already been achieved. Thus, there can be no argument from the “circumstances of rightness” to the superiority of other virtues in individual actions.

This argument, which is given in the context of rightness, can then be translated back into the context of justice by noting that justice is a moral assessment of a social institution. One has a moral duty to promote just institutions and to reform or abolish unjust ones; this duty does not arise from some novel and external assessment of what makes for justice, but rather the compatibility of that institution’s functioning with one’s pre-existing moral obligations. Suppose that we advance a conception of justice whereby certain institutions are just if and only if they have some identifying property P (say, an equal distribution of resources). It then follows that we have a natural duty to promote P-institutions and to reform or abolish non-P-institutions. Yet what is the source of this duty? Either P-institutions should be promoted in virtue of more basic, “pre-institutional” moral duties, or they should be promoted in virtue of the fact that they are P. Yet the latter claim seems entirely heteronomous and arbitrary, no matter how pleasant a description of P that we give; unless P is such that it instantiates some of our pre-existing obligations, there is no reason why its claims should be binding on us. If justice within a certain institution requires an equal distribution of resources, this fact does not find its ground in an elegant but external conception of justice (discovered through revelation, perhaps?), but in the fact that given the circumstances, our moral obligations would already preclude our giving the participants unequal shares. Indeed, Rawls places unusual conditions on the original position—a position from which all information considered morally irrelevant has been removed—in an attempt to show that our pre-institutional moral obligations require us to act a certain way in the context of institutions.

Moreover, by the contrapositive of this principle, an institution that satisfies our pre-existing moral obligations is presumptively just. Thus, to the extent that one’s moral obligations are already fulfilled within the context of, say, a loving family, there is no reason to consider that institution unjust. More detailed principles of justice, much like the specific “moral” instructions as to how to treat one’s friends, are only necessary to the extent that individuals are no longer valuing each other as ends; in this case, just institutions are needed to restore the system’s moral status.[4] A “remedial” justice can still be the first virtue of social institutions, because those institutions, where morally relevant, are purely remedial themselves.

Thus, we can see that the consideration of rightness as well as justice has the potential to defuse some of Sandel’s criticisms of the liberal account. The next task, and the more difficult one, is to apply this process to the central focus of his discussion, namely the self and its ends. To Sandel, a liberal theory of justice (and thus of rightness as well) presupposes a certain metaphysics, according to which the self is logically prior to its ends. Rather than being distinguished from other selves based on its attachments and commitments, the liberal self is “antecedently individuated.” Individual human beings, whether in the original position or out of it, can choose a conception of the good without it defining their character; the self is “standing always at a certain distance” from its interests, and can change them without losing its identity. “No project,” writes Sandel, “could be so essential that turning away from it would call into question the person I am. Given my independence from the values I have, I can always stand apart from them.”[5]

The thoroughgoing separation of the self from its ends, however, raises the specter of a “radically disembodied” self: a purely rational agent stripped of all desires, attachments, or conceptions of the good. Such an agent has not only lost everything that might make him or her recognizably human, but also everything that might potentially ground his or her moral conduct. This inability to make meaningful choices is what Sandel finds so concerning about the resident of the original position:

To imagine a person incapable of constitutive attachments such as these is not to conceive an ideally free and rational agent, but to imagine a person wholly without character, without moral depth. For to have character is to know that I move in a history I neither summon nor command, which carries consequences none the less for my choices and conduct.[6]

In contrast, Sandel describes a view of the self as “radically situated.” On this picture, the self does not exist prior to its ends—it is its ends, it is composed of them. These community attachments and principles are not merely something that the self ascribes to, but are rather constitutive of the self. And to the extent that others share our ends, our identities merge with theirs in a larger entity—a family, a class, a nation—that is uniquely able to form and pursue a common good.[7]

If Sandel’s vision of the “situated” self is accepted, one can easily see how a critique of liberal theories of justice might proceed. If the self is situated, rather than separate from its ends, there is a possibility of individuals being incorporated into a larger, communal self. This larger self may then have a good appropriate to it that is essentially common; and in that case, it will no longer be relevant to adopt the liberal perspective and examine the society from the viewpoint of the individual. Moreover, a commitment to the situated self will undermine any potential claim for justice as the “first virtue.” For if a certain communal attachment or commitment forms part of one’s very identity, then how could the abstract requirements of justice have priority?

When applied to a theory of rightness, however, Sandel’s arguments appear far less persuasive. As will be shown below, the conditions under which the self is “antecedently individuated” are in fact the only conditions under which moral action can be understood. Attempting to identify the self with its ends makes it impossible to explain how these ends can be normative for us; it also prevents us from explaining how the same individual might have ends which change over time. Ultimately rightness cannot be understood without reference to detachment from ends; and if this is true, then the same can be said of justice as well.

Although Sandel’s language is rhetorically powerful, his portrayal of the liberal self as “disembodied” is somewhat extreme. If the self is not radically situated, it is at least situated in a field of action. Even the most hard-nosed Kantian would admit that there is more to life than practical reason; people are born and live their lives in a complicated web of attachments and social norms, all of which help shape their view of the world and their role within it. Nor is there any doubt that human beings often find themselves in circumstances where their freedom appears constrained. A young man in Nazi-occupied France, forced to choose between caring for his aged mother and serving his country by joining the resistance, may very well lament his situation; he might regret having been forced by fate into such a conflict of duties. No one has the opportunity to choose the circumstances of his or her birth, or the particular kinds of moral questions they will face (“a history I neither summon nor command”). A liberal theory of rightness whole-heartedly accepts that we may be faced unwillingly with difficult choices; it is concerned only with what we as moral agents do with the choices that are given to us.

In this context, let us consider the moral obligations of the radically situated self. The commitments and attachments of which it is composed are not merely descriptive (“John is the father of Jane”), but normative, encoding certain obligations and social roles. Part of what it means to be a father, or a daughter, or a fellow-countryman, is to recognize and accept the force of these conventions. Otherwise, it is hard to see how these commitments could stand in as strong a relation to us as Sandel insists. Of course it is true of John that he, and no other, is the male genetic parent (as well as socially-identified legal guardian) of Jane; yet if his fatherhood is to be part of his identity, it must imply a certain claim on his thoughts and actions. Moreover, to the extent that these normative commitments are part of our very identity, they are, in a sense, self-executing. We are not beings who exist prior to our ends, who can adopt or disown them through an act of will. To use Sandel’s “cognitive” account, we do not choose our commitments, but discover them through reflecting on our identity.

The concept of a self-executing attachment may sound unusual, but it is the only way for Sandel to explain how a radically situated self can find itself under obligations. To the extent that our social commitments and attachments are unchosen, acting under them is quite obviously heteronomous. How, then, can they be normative for us? If John asked why he should act towards Jane in such-and-such a way, an answer that “you are the father of Jane, and your society expects fathers to do so” would not be a compelling response. For there is nothing inherent in society’s expectation that creates in us a moral obligation; according to Kant, not even the divine will has this power. Sandel also cannot respond by saying that John has identified as a father under the terms set out by his community, that he has adopted the obligations of a father as his own—for that would imply that he could have chosen not to adopt them, that he can exist distinct from them and has a self that is prior to its ends. Instead, Sandel can only argue that an attachment forming part of John’s identity is, on that basis, inherently normative. We are creatures whose nature it is to act on our commitments; we simply cannot, without compromising our identity, do otherwise.

Yet however intriguing such a portrayal might sound in the abstract, it is quickly revealed as implausible, perhaps even bizarre, when it is applied to the actual process of decisionmaking. Consider, for example, the practice in some societies of ‘honor killing.’ The argument might go as follows: in a society whose mores allow—indeed, mandate—such violence, part of what it means to have a daughter is to recognize and accept the necessity of murdering her should she bring dishonor on the family by being raped. Now suppose that John, who lives in such a society but has never thought very much about the practice before, learns that his daughter Jane has been raped, and suddenly finds himself very hesitant to abide by the community mores. Perhaps this is because he knows the practice to be a relatively recent historical development, or because he is aware of other societies (even those of his co-religionists) that have interpreted the father-daughter relation differently. But regardless of his qualms, he cannot fail to murder her, on pain of compromising his identity. In fact, he cannot even coherently imagine protecting his daughter—were he to try, it would not be him protecting her, but some alien intelligence distant from his thoughts and ways of life. Such a portrayal is incompatible with any reasonable notion of agency. We would be, on Sandel’s account, forced to think of ourselves as community-mores-obeying robots.

Some might charge that this example is unfair; no one ever considers the practice of honor killings as part of their identity. At most, John identifies with his status as a father, and then decides whether or not to participate in the practice on that basis. But if John’s self-identification is to have any weight in deciding this question, it must come with some normative baggage attached. And if this baggage is to constitute part of his identity, then any obligation it includes must be self-executing—and, like instincts and other psychological compulsions, morally non-reviewable.

If Kant’s arguments against heteronomy are accepted, however, such attachments could not be self-executing for any person to whom we ascribe even a modicum of practical reason. We must be detached from our values at least to the extent that we can assess them morally and do not consider them brute and natural facts. John might know that he is the father of Jane, and know that his society has certain expectations with regard to his conduct, but he is only bound by such expectations to the extent that he adopts them as his own—to the extent that he views himself as prior to his ends. As moral agents, our attachments are only self-executing if we do not think about them. As soon as we subject them to practical reason, their supposedly inherent bindingness disappears, and we are free to reject or condemn them—even if, psychologically, we could not do without them. If a liberal account of rightness is accused of offering a false picture of our social attachments, then the communitarian account of rightness may be accused, perhaps with more justice, of offering a false picture of our moral psychology. We are born into attachments, to be sure, but we construct their normative value for ourselves.

On Sandel’s picture of the self, even such moral scrutiny is too great a detachment from our ends. Yet his account fundamentally rests on an equivocation as to the nature of this detachment. Consider Sandel’s final expression of the type of detachment liberalism requires:

The independence of self [required by liberal theory] does not mean that I can, as a psychological matter, summon in this or that circumstance the detachment required to stand outside my values and ends, rather that I must regard myself as the bearer of a self distinct from my values and ends, whatever they may be.[8]

Now compare with this the description of a desire forming part of one’s identity:

Imagine that a desire, held tentatively at first, gradually becomes more central to my overall aims, until finally it becomes an overriding consideration in all I think and do. As it grows from a desire into an obsession, I possess it less and it possesses me more, until finally it becomes indistinguishable from my identity.[9]

It is crucial to note that the latter process is portrayed as entirely psychological. One’s desire is held with greater and greater strength; it becomes an “obsession,” a factor in all of one’s thinking; it eventually becomes something that one could not imagine losing. But what one could not imagine is having the psychological detachment to act against the desire; no one is asserting that life without this desire would be metaphysically impossible. Indeed, we can imagine ourselves committing all manner of acts—brutal murder, bank robbery, howling at the moon in polite company—and from the standpoint of practical reason can assess them, no matter how contrary they are to our deeply-held attachments and commitments. (Otherwise, there would be no grounds for moral discussion of such commitments, for such discussion would require at least the conceivability of suspending them.)

But there is no reason to believe that these psychological factors ever grow to metaphysical strength, in part because this is a process that takes place in time. If on the basis of a sufficiently strong desire I claim that the self does not exist prior to its ends, how do I explain the fact that I did not always think this way? What existed prior to when the self lost its prior existence? Sandel claims that moral thinking is more a matter of reflection than will, an attempt to answer the question “Who am I?” rather than “What ends shall I choose?” Yet his cognitive account, for all of its supposed virtues, is remarkably unreflective. How is it, exactly, that we discover who we are? Do we dig up the surface layers of the mind, and find commitments underneath? And how is this different from the psychological process—which Sandel examines at length, and which he portrays, against Rawls, as insufficient to generate normative principles—of merely identifying our unchosen wants and desires?

On the cognitive picture, it is not clear how we may account for the fact that identities, as well as communities, tend to change over time. We may regard communities and cultures as self-sufficient entities for the purposes of anthropology, preserved like a specimen in formaldehyde for future study. But when we examine how human beings actually recognize and accept a community as part of their lives, the process is very different. Social commitments and attachments do not enter people’s lives in neat anthropological packages, but haltingly and in fragments.[10] Certain commitments seem unnecessary at one time and vital at another—and how can we deny that choice sometimes plays a role in such revisions?

The cognitive account admits for change only insofar as we progressively eliminate erroneous self-conceptions; we might have been confused as to our true nature, and then discover it later on. But not all change to our identity is like this; old attachments and commitments often pass away, and new ones take their place. Consider a man who has been kidnapped from his home in Africa and brought to the antebellum South on a slave ship; certainly a grave injury has been done him, but he does not die, or lose the capacity for moral judgment. Or, for a less extreme example, consider what happens to parents when a child is born. Must we really believe that the parents’ obligations to their new baby exist timelessly and are merely discovered through a process of reflection, rather than being constructed in a historical process that may involve some degree of choice?[11]

More concerningly, if the self is simply composed of various ends, then there is no criterion by which we can identify the woman who was pregnant with the woman who now has a newborn, for the former set of ends and attachments is distinct from the latter. Sandel notes, but then conveniently forgets, that Hume himself recognizes the picture of the self as a “bundle” of perceptions and attachments as an ultimately incoherent account of personal identity.[12] This incoherence removes the need for any serious consideration of the radically situated subject. And if the self is only partially situated, if it exists prior to at least some of its ends, the question will immediately arise as to why it does not exist prior to the others. If we have a technique of individuating selves antecedent to certain ends, why will that same technique not succeed in individuating them entirely? And to the extent that Sandel limits moral agency to those ends to which the self is prior, we are forced to give up our notions of moral agency with regard to the others, and view ourselves (at least partly) as community-mores-obeying robots.

Sandel’s account eschews a historical analysis and does not search for the origins of attachments. Instead, it proceeds as if it is already obvious that these loyalties exist and are morally binding:

But we cannot regard ourselves as independent in this way without great cost to those loyalties and convictions whose moral force consists partly in the fact that living by them is inseparable from understanding ourselves as the particular persons we are—as members of this family or community or nation or people, as bearers of this history, as sons or daughters of that revolution, as citizens of this republic. Allegiances such as these are more than values I happen to have or aims I “espouse at any given time.”[13]

However, in the course of identifying aspects of our identity as morally binding, Sandel fails to show why it is important that these particular aspects of our identity or moral psychology are intersubjective, rather than entirely random. If occupying a given social role can be morally obligatory for me simply because it is an overriding desire of mine, could not other, similarly strong desires create other moral obligations? Without grounding them in an autonomous agent, Sandel cannot give an account of why these convictions have “moral force.” At best, they can have psychological force—a result that should not surprise us at all, given that human psychology is profoundly shaped by experience.[14] We do not need any moral theory to explain our feelings of attachment to a family or community; what we need is an explanation of why—or, indeed, whether—the actions we take based on those feelings will be morally acceptable. If no such explanation can be provided, or if the acts are in fact not morally acceptable, then Sandel cannot possibly be advancing a non-liberal theory of rightness. At most, his discussion of attachments can have predictive power as to how individuals do act, or how their societies expect them to act, but not how they should act, or which acts are right.

To obtain an account of rightness, we must look elsewhere. The disembodied self, disassociated from all attachments that might bind it; the free and unconstrained noumenon, acting as first cause while yet uncaused in itself—these are not the types of characters one is accustomed to meeting in daily life. Yet these are exactly the types of entities that one must identify from the standpoint of practical reason. When I deliberate and make a decision as a moral agent, I do so under the presumption that my will is undetermined by external factors. Even if I do not believe (according to theoretical reason) that my actions are free, it is a requirement of the practical standpoint that I act under the idea of freedom. One cannot regard one’s own actions, “from the inside,” as externally caused; rational action is only possible on the assumption that we are not robots, and that our desires, however strong, are not self-executing. Once one assumes the role of a self-aware agent endowed with practical reason—once one recognizes that one’s attachments are not in themselves binding, that they may be tested against other commitments or before the bar of practical reason—then all of Kant’s arguments against heteronomy appear with full force to deny the self-executing nature of attachments. One may still act on these attachments, of course, but that does not mean that they are regarded as controlling,

Such a picture of our moral psychology justifies the liberal reliance on the individual. The liberal self is antecedently individuated because moral agency is antecedently individuated. When we speak of the moral responsibilities of a nation, we are using shorthand for the moral responsibilities of its individual citizens, who are moral agents; we do not suppose that there is a larger metaphysical entity with the faculties of reflection and will. But we do believe that individual human beings have these faculties, and that they are appropriately regarded as moral agents—and conversely, to the extent that we regard ourselves and others as moral agents, we must regard them as possessing individual selves.

Thus, the communitarian account fails to generate an acceptable theory of rightness. As a source of obligations, rightness must be the first virtue of actions; and to provide for normative and binding principles, one must adopt a view of the self as antecedently individuated. Thus, liberal theories of rightness seem far more promising than any alternative. Additionally, if the “natural duty” interpretation of justice advanced above is accurate, then Sandel’s argument must similarly fail in the case of justice. Given that the just society is defined in terms of rightness, it will therefore be defined with reference to an antecedently individuated self; given that rightness must be the first virtue of actions, justice must therefore be the first virtue of social institutions. The liberal account of justice is sound.

Yet one final line of response remains open: Sandel could still challenge the link between justice and rightness. Perhaps rightness does require a liberal conception of the self; but if justice is not merely rightness as applied to institutions, perhaps it will still require something more. In fact, Sandel adopts this argument; our moral obligations, he argues,

. . . go beyond the obligations I voluntarily incur and the “natural duties” I owe to human beings as such. They allow that to some I owe more than justice requires or even permits, not by reason of agreements I have made but instead in virtue of those more or less enduring attachments and commitments which taken together partly define the person I am.[15]

The difficulties with a novel or external conception of justice have already been discussed above. But in this context, Sandel’s response seems to take the form of a critique of rationalism: no theory of justice that does not rely on self-executing community attachments could ever explain why it is so important to love one’s children, or why it could ever be moral to treat one’s aged mother differently from her equally aged next-door neighbor. Sandel’s argument here seems to resemble G.E. Moore’s “refutation” of skepticism: our intuitions—that we owe certain things to our mothers, or (in Moore’s case) that we have hands—are simply too strong for any contrary theory, no matter how well argued, to be correct.

Yet this response contains two significant flaws. First, it is not clear that our intuitions as to social obligations are in fact quite as strong as Sandel thinks. We can imagine ourselves maintaining, in the face of all logic, that parents have certain special obligations with regard to their children. But imagine an Oxford don who maintained, in the face of all logic, the rightness of honor killings; he would not be considered noble, but insane. Because Sandel’s picture of the situated self is general, equally applicable in our own and other societies, we are unable to pick and choose which society’s attachments are uniquely capable of generating special obligations. Yet whatever kind of intuitions we have that support our social attachments, they do not support everyone’s social attachments, unless we are willing to lapse into a complete moral relativism. And our intuitions opposing such relativism (as well as those opposing honor killings, at least in Oxford) are very strong indeed.

Second, the ‘natural duties’ account of justice may have a flexibility that Sandel does not recognize. It is certain that the kinds of attachments to which human societies give rise are extraordinarily important to those whom they bind. To be loved by one’s family, for instance, is a good whose loss cannot easily be compensated for by the attentions of others. Yet there is no reason why a non-self-executing view of morality cannot take this importance into account. These attachments clearly find part of their force in their effects on those who are bound. Lear was not driven mad by the betrayal of persons he met on the street; what he found “sharper than a serpent’s tooth” was to have a thankless child. The kind of solace that Matthew Arnold seeks in “Dover Beach” (“Ah, love, let us be true / To one another!”) cannot be provided by contracting with strangers.[16] These effects provide powerful resources for a purely autonomous theory of right—one denying the self-executing nature of attachments in themselves—to find grounds for our familial or community obligations. One may recognize such attachments as situating the subject in a field of action without situating him or her wholly outside the practical standpoint; one’s power to do good or harm is vastly expanded in the context of such relationships, and this fact need not go ignored by moral theory.[17]

Thus, without fleshing out a complete conception of justice founded in natural duties, we can have some confidence that our understanding of just institutions is inherently dependent on our understanding of moral obligation and right action. Sandel’s communitarian account gives a false picture of the latter; by removing the necessary detachment of the self from its ends, the theory prevents any well-grounded moral assessment of one’s actions. From the standpoint of practical reason, we are required to regard our desires from a distance; we are forced to act under the idea of a freedom that the communitarian viewpoint denies. And if the self, in a theory of rightness, must be viewed as separate from its ends, there is no reason to think the two inseparable in a theory of justice.

[1] This may be a persuasive argument for some; but it should be noted in passing that it is an empirical argument, and must stand or fall with the available psychological evidence. Alternatively, some communitarian critics might argue that community attachments are a (non-empirically) necessary part of a good life, and that even individuals who consider themselves fulfilled outside of them remain, in an important way, impoverished. Yet this position is not so much a critique of liberalism as an alternative theory of the good, and a rather thick theory at that; its success as a criticism will depend entirely on these critics’ success in establishing their positive proposal and demonstrating its incompatibility with liberal freedoms.

[2] A Theory of Justice, p. 111.

[3] Cf. the title essay in Korsgaard’s Creating the Kingdom of Ends.

[4] As an example, suppose that two friends regularly dine together at restaurants. On one analysis (and ignoring distributional aspects), it might be most “just” for them to pay only for their own meals, or perhaps to pay for exactly the same number of meals when the bills cannot be easily split. But justice need not require such penny-pinching exactitude; if one of them tends to pay more or more often than the other, no injustice arises, so long as the two value each other as friends and are not merely using the other a source of free food. It is only when one of them fears that they might be so used that it becomes necessary for them to identify and adhere to a set of specifically just allocations of payment.

[5] Liberalism and the Limits of Justice, p. 62.

[6] Liberalism and the Limits of Justice, p. 179.

[7] It may be, on Sandel’s account, that the self need not be radically situated, but only partially so. Sandel’s claims are fulfilled as long as some of one’s ends are constitutive of one’s identity; the others might be ends one has rather than is.

[8] Liberalism and the Limits of Justice, p. 182.

[9] Liberalism and the Limits of Justice, p. 56.

[10] Cf. Jeremy Waldron’s account of folk stories in “Minority Cultures and the Cosmopolitan Alternative,” 25 U. Mich. J.L. Ref. 751 (1992): “If we were making an anthropological study of each item, we would want to explore the detail of its context and provenance; we would look at the tale of the prodigal son in the context of Aramaic storytelling, and we would confine the children lost in the wood to the Germanic villages from which the Grimm brothers drew their collection of folklore. But that is absurd as an account of how cultural materials enter into the lives and choices of ordinary people. For that purpose, the materials are simply available, from all corners of the world, as more or less meaningful fragments, images, and snatches of stories. Their significance for each person consists in large part in the countless occasions on which they have been (from the anthropological purist’s point of view) misread and misinterpreted, wrenched from a wider context and juxtaposed to other fragments with which they may have very little in common.”

[11] One possibility is that the self is in part constituted by a wide variety of conditional attachments (e.g., “if I had a child, then I would be obliged to act towards him or her in such-and-such a way”) which are then instantiated by life experiences. But the sheer number of conditional attachments that would be necessary (“if I were kidnapped and sold into slavery on a foreign continent, then . . .”) seems far too large for them all to be plausibly pre-included in an individual’s identity.

[12] Sandel, pp. 12-13; P.F. Strawson, The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason, pp. 169-170.

[13] Liberalism and the Limits of Justice, p. 179.

[14] Why else would psychologists always ask patients about their mothers?

[15] Liberalism and the Limits of Justice, p. 179.

[16] Similarly, forgetting a child’s birthday present is less serious as a moral issue when the child is not particularly bothered by the oversight.

[17] A non-self-executing account can also provide an explanation of how such attachments may retain their force even in societies where the nature of the attachments differ. In a society where, say, the relationship between uncle and nephew is generally seen as far more important than that between father and son, the obligations of an individual would be different from what we are used to. To put it crudely, the failure of an uncle to buy his nephew a birthday present in a society that placed great value upon that relationship would mean far more disutility than could be compensated for by the impersonal provision of the present. I do not mean to suggest that all, or even most, of our intuitions regarding familial obligations can be given by such simple analysis; but it is certainly something worth noting that one can discuss the value that individuals place on their attachments without abandoning a liberal framework or conceding that such attachments are necessarily self-executing.