II. Smith’s Main Account of Justice

Smith’s only sustained discussion of justice comes in Part II, Section II of TMS. There he writes that “because the mere want of beneficence tends to do no real positive evil,” it follows that (1) beneficence therefore “cannot be extorted by force” and (2) “the mere want of it exposes to no punishment” (TMS II.ii.1.3). According to Smith, if someone acts with justice toward you, it leaves neither you nor anyone else worse off than you already were, though by itself it may not make you or anyone else better off. For this reason, Smith calls it “a negative virtue” (TMS II.ii.1.9), claiming, remarkably, that “We may often fulfil all the rules of justice by sitting still and doing nothing” (TMS II.ii.1.9). The person sitting still and doing nothing is not acting with positive virtue — that is, is not generating any improvement — “But,” Smith contends, “still he does no positive hurt to anybody. He only does not do that good which in propriety he ought to have done” (TMS II.ii.1.3).

So failing in proper beneficence ^{Footnote 3} — a category that for Smith includes things like charity, compassion, generosity, and “humanity” — might give us reason to disapprove of and be reasonably disappointed by another’s behavior, but Smith believes it does not license coercive punishment like retribution, jailing, or fines. If I do not do you the good office you hoped or expected I would, you may be disappointed, even justifiably so. Because I have done you no “real positive hurt,” however, meaning I have not worsened your ex ante position, you may not take positive action to punish me. By contrast, if I fail to act with justice toward you, that means I acted maleficently and did indeed do “real positive hurt” to you. I left you worse off than you were before, and that, according to Smith, gives rise to justified resentment, which licenses punishment.

Smith claims that there are only three rules of justice, which he calls, interestingly, “sacred”: (1) “the laws which guard the life and person of our neighbor,” (2) “those which guard his property and possessions; and last of all come [3] those which guard what are called his personal rights, or what is due to him from the promises of others” (TMS II.ii.2.2). We act justly, then, according to Smith, when (1) we do not kill, enslave, or molest others; (2) we do not steal from or defraud others; and (3) we do not break our voluntary contracts or promises. By contrast, beneficence involves making at least one person better off, while — assuming justice was respected — no one was made worse off. If, however, such improvement was not required of me by law, contract, promise, or any other specific obligation, on Smith’s account you cannot have had an enforceable expectation of my improving your position.

III. Smith’s Other Uses of Justice

Smith repeats or uses the same account elsewhere. For example, later in TMS, when discussing competing schools of moral philosophy, Smith canvases different conceptions and uses of justice. There he reaffirms that the sense in which he is using it is his “negative” account:

“In one sense we are said to do justice to our neighbor when we abstain from doing him any positive harm, and do not directly hurt him, either in his person, or in his estate, or in his reputation. This is that justice which I have treated of above, the observance of which may be extorted by force, and the violation of which exposes to punishment” (TMS VII.ii.1.10). In his Lectures on Jurisprudence (LJ), Smith is reported as claiming: “The first and chief design of every system of government is to maintain justice; to prevent the members of a society from encroaching on one another’s property, or siezing what is not their own. The design here is to give each one the secure and peaceable possession of his own property.”^{Footnote 4} In The Wealth of Nations (WN), Smith repeats that the “first duty of the sovereign” is “that of defending the society from the violence and injustice of other independent societies”;^{Footnote 5} he adds that the “second duty of the sovereign” is “that of protecting, as far as possible, every member of the society from the injustice or oppression of every other member of it, or the duty of establishing an exact administration of justice” (WN V.i.b.1). The most straightforward interpretation of these passages is that Smith understands “justice” as the protection against maleficence. In the many places in WN in which Smith mentions or discusses the “administration of justice” or even “a well-governed society,”^{Footnote 6} it is this conception he has in mind.

In several places, however, Smith’s descriptions reveal complexities. Note, first of all, the word “often” in the sentence quoted earlier: “We may often fulfil all the rules of justice by sitting still and doing nothing” (TMS II.ii.1.9). Often is not always, which apparently implies that there are times when justice does indeed require us to do something beyond “sitting still and doing nothing.” Some such cases seem obvious and consistent with Smith’s account. For example, if I have made a voluntary promise or voluntarily assumed an obligation, then justice might require me to take positive action under relevant circumstances. If I am a parent, I have indefinitely many obligations to my children, which would seem reasonably to count as voluntary and thus under the heading of justice; if I am the paid lifeguard of a pool or the paid security guard for a building with a manmade pond, then if someone is drowning in either, I am required by justice to take action. We might also extend the obligation to exigent circumstances in which my particular expertise is called for or would help. If I am a medical doctor and there has been an accident or if I am a police officer and a crime is being committed, then, even if I am officially off duty, it seems reasonable to assume I have a positive obligation to act because doing so under such circumstances is part of what I accepted when I became a doctor or police officer — and thus it would be an injustice, not merely a lack of proper beneficence, for me not to do so. If, however, I have no such obligation, then I can fulfill justice simply by not acting maleficently toward others — as Smith says.

Smith can accommodate these extensions of justice by reference to the proper motivation to address injustice that he indicates in TMS, namely, natural resentment. Smith writes, “Resentment seems to have been given us by nature for defence, and for defence only. It is the safeguard and the security of innocence” (TMS Ii.ii.1.4). The emphatic clause “and for defence only” might seem to limit the appropriate application of Smithian justice to only refraining from or punishing maleficence. Yet Smith’s inclusion of the requirement to fulfill voluntary promises and obligations under “justice” means that he believes that failing to fulfill a voluntary promise also violates “the security of innocence.” How? Perhaps because he believes we would naturally feel resentment toward a person who failed to fulfill the obligation. But in that case we seem to have not one but two criteria for “justice”: (1) making another’s position worse off (that is, maleficence), and (2) inducing natural resentment in disinterested observers. What is the relation between these two? I suggest that natural resentment is the more fundamental. Maleficence is a failure of justice because it makes another worse off, but we are justified in punishing it because the natural resentment we feel toward it — a resentment that is implanted in us “by nature” (by God?) — warrants our belief that it is indeed wrong and requires punishment. Smith says something similar in LJ: “From the system I have already explain’d, you will remember that I told you we may conceive an injury was done when an impartial spectator would be of opinion he was injured” (LJ, 17).^{Footnote 7} The reference to an “impartial spectator” suggests that Smith imagines we would ask ourselves whether a disinterested but fully informed spectator would feel resentment in the case at hand: if so, it counts as an “injury” and thus injustice; if not, then not. This line of reasoning further suggests that it is ultimately the presence or absence of natural resentment — though perhaps only if sought through the vehicle of a proper “impartial spectator” — that determines whether an injustice occurred.^{Footnote 8}

In LJ, Smith concentrates his discussion of justice on the injuries others may do to our property, but he includes as property our physical persons, our liberty to act, and even our reputations (see LJ, 13). He elaborates that the government is charged with protecting justice, and he distinguishes between “commutative justice,” or the protection of our “perfect rights,” which are those to “which we have a title to demand and if refused to compel an other to perform”; and “distributive justice,” or the protection of our “imperfect rights,” which are those “which correspond to those duties which ought to be performed to us by others but which we have no title to compel them to perform” (LJ, 9). The former, Smith says, are proper objects of governmental concern and “force,” while the latter belong instead “to a system of morals as they do not fall under the jurisdiction of the laws” (LJ, 9).

Thus, beneficence is frequently morally required of us, but because it is not part of justice, it is therefore not a justifiable cause for state or other third-party coercion. Almost all of the mentions of justice in The Wealth of Nations exhibit or assume this same negative conception, holding it to apply to people’s security in their persons, property and possessions, and voluntary promises. Smith also occasionally uses the word “equity,” although usually as a synonym for “justice.” For example, Smith writes in LJ that it “is evident that by the law of equity such possessions should be restored to the right owner” (LJ, 24); he refers to what both “the civil law of all countries and naturall justice and equity” hold and to what both “Justice and equity plainly require” (LJ, 100 and 101); and he writes of what “all justice and equity in the world” hold (LJ, 436).

Yet in one place in WN, Smith writes, “No society can surely be flourishing and happy, of which the far greater part of the members are poor and miserable. It is but equity, besides, that they who feed, cloath and lodge the whole body of the people, should have such a share of the produce of their own labour as to be themselves tolerably well fed, cloathed and lodged” (WN I.vii.36). Some commentators take this passage to imply that Smith endorsed more than a merely negative conception of justice, perhaps one that incorporates into the notion of justice at least some positive duties to benefit others.^{Footnote 9} Samuel Fleischacker argues that either we must interpret Smith’s conception of justice as including some duties of beneficence — according to Fleischacker, the correct interpretation, which he claims justifies “the modern notion of distributive justice, the notion used to justify socialism and welfare state liberalism”^{Footnote 10} — or “we must simply recognize a contradiction” in Smith’s account.^{Footnote 11}

Let me raise two further complicating passages. First, in TMS, aside from duties like those discussed earlier of parents to children, people in their official and even unofficial capacities, and so on, Smith writes: “A superior may, indeed, sometimes, with universal approbation, oblige those under his jurisdiction to behave, in this respect, with a certain degree of propriety to one another. [ . . . ] The civil magistrate [ . . . ] may prescribe rules, therefore, which not only prohibit mutual injuries among fellow-citizens, but command mutual good offices to a certain degree” (TMS II.ii.1.8). Second, in WN, Smith argues that in order to mitigate the “mental mutilation” (WN V.i.f.60) to which extreme division of labor can lead in workers who perform the same rote tasks over and over, “The publick can facilitate this acquisition [of “the most essential parts of education,” namely, “to read, write, and account”] by establishing in every parish or district a little school” (WN V.i.f.54–55). In these cases Smith seems to be imputing duties to government that go beyond the negative, “defence only” conception of justice (from here on, Smith’s “NDO” conception) that he elsewhere defends. Is Smith’s discussion of justice ultimately coherent?

IV. Smith’s Reasons for His “NDO” Conception of Justice

Smith’s NDO conception seems rather thin. Yet I suggest that the way to understand Smith’s conception of justice is as providing a robust default — reflected by his calling the rules of justice “sacred” — from which, however, special exceptions can be made under specific circumstances. The default is his NDO conception, which he articulates carefully in TMS and which informs almost all of his references to and uses of both “justice” and “equity” in TMS, WN, and LJ. Yet Smith arrives at this conception not by deduction from first principles or an interpretation of natural law — the way, for example, Locke, whose work Smith knew well,^{Footnote 12} might — but rather by induction on the basis of his historical observations. His argument is that the NDO conception is what history has proved is crucial for society to function and what people’s natural resentment actually recommends. But because he is trying to survey the often messy and even inconsistent historical record of human experience, he recognizes that there are circumstances under which sound judgment will require departure from those principles. When he suggests repairing to the judgment of an impartial spectator, it is to help adjudicate these potential exceptions. Yet he affirms that such circumstances will be exceptional and rare.^{Footnote 13}

To support this interpretation of Smith’s account, let us look at some of the reasons Smith offers in support of the NDO conception of justice. The first is Smith’s claim that no society can subsist unless its members respect these rules of conduct. Even a society of “robbers and murderers” must at least “abstain from robbing and murdering one another” (TMS, II.ii.3.3). On the other hand, a society can subsist if its members respect these rules of justice but do not act with beneficence toward each other. Because beneficence “is less essential to the existence of society than justice,” Smith concludes that “Society may subsist, though not in the most comfortable state, without beneficence; but the prevalence of injustice must utterly destroy it” (TMS, II.ii.3.3). Justice, therefore, “is the main pillar that upholds the whole edifice” of society, while beneficence “is the ornament which embellishes, not the foundation which supports the building”; for that reason, Smith argues, it is “sufficient to recommend, but by no means necessary to impose” beneficence (TMS II.ii.3.4). This is consistent with Smith’s claim in LJ that the protection of “perfect rights” (that is, against maleficence) is the duty of legislation, while the protection of “imperfect rights” (that is, the benefits others should provide us) falls instead within the province of personal moral relations. Thus, for Smith, justice is both a necessary and a sufficient condition for the existence of society, but beneficence is neither necessary nor sufficient. That means that they enjoy a lexical priority — justice first, beneficence only thereafter — and, if we assume that the state is justified in providing only what is necessary for human society, then it follows that it is justified in providing only NDO justice.

A second reason Smith supports this thin conception of justice is its relative ease of administration. He argues that it is relatively easy (1) to capture its essence in simple rules, (2) to detect infractions of it, and (3) to remedy infractions. By contrast, beneficence is far more difficult to describe in rules, far more difficult to detect in its absence, and far more difficult to remedy.^{Footnote 14} Unlike infractions of justice, improper beneficence can be detected and adjudicated only on the basis of detailed, context-specific knowledge of the situation, persons, and matters involved in particular cases. We might all agree, for example, that we should be charitable and that charity is a virtue; nevertheless, it would be very difficult to generate a set of precise rules that will allow us to determine what charity requires of any particular person in specific circumstances. In practice, we have to rely on practical judgment, which Smith, like Aristotle, believes does not operate by mechanical execution of general rules.

This suggests a final reason Smith endorses his negative conception of justice: it allows a proper sensitivity to individual circumstances. What counts as being sufficiently charitable depends on the particular circumstances of the case in question. The history and situation of the people involved, their available means and tradeoffs and opportunity costs, and even their goals and ambitions are all material considerations that are difficult to assess from afar and in advance. There is also typically a range of behaviors or actions that might qualify as properly beneficent, which means that no single course of action will be required to satisfy one’s obligations. Thus, beneficence cannot plausibly be incorporated into the definition of justice, which, because it can license coercion, requires predictable and relatively precise application of clear rules. Smith’s NDO concept of justice restricts it, therefore, to those few areas of conduct that it can plausibly and effectively address, and leaves to localized judgment the determination of what positive beneficence requires in light of particular circumstances, while at the same time allowing for the possibility of departures in exceptional circumstances when an impartial spectator who indeed possesses the relevant local knowledge judges such departures appropriate.

V. Social-Justice Objections to Smith

Let me now raise a cluster of related social-justice objections to Smith’s thin NDO account of justice. The first can be adapted from an influential argument Peter Singer has given to defend his position that the existence of starving people in the world imposes obligations on those of us who are not starving. Singer’s argument: (1) Starvation is very bad; (2) If you can prevent something very bad from happening, without sacrificing anything nearly as important, you should do so; (3) by donating to hunger relief agencies, you can prevent something very bad from happening without sacrificing anything nearly as important; therefore (4), you should donate to hunger relief agencies.^{Footnote 15}

Singer’s position presumes that a Smithian distinction between “negative” inaction and “positive” action does not hold, or at least is not dispositive in determining moral culpability or obligation. For Singer, I am equally morally blameworthy whether I fail to help you when I could or I deliberately make your situation worse. As he puts it, there is no moral difference between killing and letting die. ^{Footnote 16} Moreover, whereas Smith employs two variables in judging a person (justice and beneficence), Singer has only one (right/justice and wrong/injustice).^{Footnote 17}

This has important policy implications. Smith argues that the state can be charged with defending justice but not beneficence, leaving the latter to private persons or groups and absolving the state from responsibility for it. Because Singer denies the distinction, on his account if it is wrong not to give money to hunger relief agencies, then there is no reason in principle why the state should not enact policy to enforce that moral obligation. Indeed, for Singer justice requires state action when private initiative is not sufficiently forthcoming.^{Footnote 18}

I count this objection to Smith’s conception of justice under the heading of “social justice” because it relies on a conception of justice that requires us to take positive action to remedy at least some kinds of, as Singer puts it, things that are “very bad.” Although “social justice” is defined differently by different people,^{Footnote 19} one common feature is the incorporation into justice of at least some positive moral obligations — typically a positive requirement to help the poor, the hungry, the disenfranchised, the powerless, perhaps the undeservedly unlucky. John Tomasi, for example, argues that social justice, which on his conception pays attention not only to the rules regarding acquisition and transfer of property but also the resulting patterns of holdings, should be a robust concern of all political theorists, including those espousing market economies.^{Footnote 20} Tomasi argues that concern for proper distributions of holdings, which focuses in particular on the holdings of the poor, should be a central aspect of both the analysis and putative judgment in favor of market economies. Tomasi’s argument is that the Rawlsian “justice as fairness” argument should (1) be taken seriously even by (classical) liberals like Adam Smith and (2) that the market’s relative ability to realize the social justice implied by “justice as fairness” should be an integral component in the evaluation of market economies: “In a just society, institutions and rules should be crafted so that whatever broad patterns of inequality emerge reflect our commitment to respecting all citizens as valued members of a cooperative whole.”^{Footnote 21} Even more recently, Thomas Piketty takes a similar position when he argues that “social inequalities are acceptable only if they are in the interest of all and in particular of the most disadvantaged social groups.”^{Footnote 22}

These claims rely on a conception of justice that requires us to take positive action to remedy at least some kinds of, as both Tomasi and Piketty put it, inequalities that are inconsistent with social justice.^{Footnote 23} The social-justice argument hence suggests that Smith’s NDO conception of justice, including the limitations it places on justified state action, is indeed too thin because (1) it disallows too many important vehicles for the alleviation of misery and suffering, (2) it fails to acknowledge any corporate or public obligation toward those less fortunate, and (3) it seems unduly to privilege its “negative” conception of individual freedom over, for example, welfare or equality.^{Footnote 24}

VI. Smith’s Response to the Social-Justice Objections

Despite how apparently thin it is, I believe understanding Smith’s NDO conception of justice as a default that allows for the possibility of exceptions when an impartial spectator judges appropriate — rather than as an inflexible rule — strikes a plausible balance that allows it to address some aspects of social-justice arguments like Singer’s (and Tomasi’s and Piketty’s).

I base my claim on what I call Smith’s Local Knowledge Argument (LKA): Given that everyone has unique knowledge of her own “local” situation, including her own goals, desires, and opportunities, each individual is typically therefore the person best positioned to make decisions about what courses of action she should take to achieve her goals.^{Footnote 25} That does not mean that people are infallible in judging their own situations, but rather that individuals have a better chance of knowing how best to use their own resources and what courses of action to take to achieve their goals, including their positive moral obligations, than do third parties because individuals are more likely to possess the local knowledge required to make such determinations reliably. Because third parties typically do not know your schedule of value, your opportunity costs and tradeoffs, your tastes and preferences, or your voluntarily assumed positive obligations, the LKA implies that we should have less confidence in their ability to make accurate judgments about how you should allocate your scarce resources and energies.

Another set of examples from recent critics will illustrate how Smith’s argument would respond to the social-justice objection. Some have argued that justice requires us to take positive action to ensure that all people have their “basic needs” met, or to make sure they possess proper “basic goods.” As David Copp, for example, argues:

Similarly, Martha Nussbaum offers a list of ten “Central Capabilities” that she argues it is “the task of government” to “secure to all citizens.”^{Footnote 27} Among other things, her list includes: “Being able to have good health, including reproductive health; to be adequately nourished; to have adequate shelter”; “having opportunities for sexual satisfaction”; “Being able to have pleasurable experiences and to avoid nonbeneficial pain”; “Not having one’s emotional development blighted by fear and anxiety”; “Being able to form a conception of the good and to engage in critical reflection about the planning of one’s life”; “Having the social bases of self-respect and nonhumiliation”; “Being able to laugh, to play, to enjoy recreational activities”; and “being able to work as a human being, exercising practical reason and entering into meaningful relationships of mutual recognition with other workers”^{Footnote 28} As Nussbaum makes clear, for her these are not mere aspirations: they instead constitute “basic social justice,” and her list “ascribes an urgent task to government and public policy” to make her Central Capabilities a matter of “constitutional guarantees” and “fundamental human entitlements” enforced and secured by government.^{Footnote 29}

Even granting that these needs, goods, and capabilities are indeed basic, even necessary for happiness, the challenge lies not in knowing how to procure and provide these things in general but rather in knowing how to procure and provide it for specific persons — and the LKA suggests that this cannot be known by distant third parties, however expert. Consider “having opportunities for sexual satisfaction,” being “able to have pleasurable experiences,” having “the social bases of self-respect,” or being able “to enjoy recreational activities”: Even if one agrees that a happy or full life requires these, third-party policy makers cannot know wherein they consist for particular persons, or how policy could support it for particular persons. Experts in various relevant fields — economics, psychology, medicine, ecology, nutrition, and so on — possess only aggregate knowledge that is general and abstract, not tied to particular individuals. They may possess statistical generalities, but they cannot know what any individual should do, and hence Smith’s LKA would hold that they are ill-positioned to structure governmental policy to effectuate those goals in particular cases. According to Smith’s argument, then, “The sovereign is completely discharged from a duty, in the attempting to perform which he must always be exposed to innumerable delusions, and for the proper performance of which no human wisdom or knowledge could ever be sufficient; the duty of superintending the industry of private people, and of directing it towards the employments most suitable to the interest of society” (WN IV.ix.51).

A final example comes from recent neo-behaviorist work informing public policy — as found, for example, in Richard Thaler’s and Cass Sunstein’s Nudge, which argues that experts should arrange the way options appear so that people are more inclined to choose things they should. Thaler and Sunstein argue for what they call “libertarian paternalism,” which they define as the strategy to devise policy that will “maintain or increase freedom of choice” (the libertarian part) and at the same time will “influence people’s behavior in order to make their lives longer, healthier, and better” (the paternalistic part).^{Footnote 30} Their stated goal is to help people make the decisions they would have made “if they had paid full attention and possessed complete information, unlimited cognitive abilities, and complete self-control.”^{Footnote 31} This seems like a high standard, and as a result some critics have charged them with erring on the side of paternalism versus that of libertarianism,^{Footnote 32} but they insist that their intention is only “to influence choices in a way that will make choosers better off, as judged by themselves.”^{Footnote 33} This is exactly the point at which Smith’s LKA applies. When Smith writes, “[W]hat is the species of domestick industry which his capital can employ, and of which the produce is likely to be of the greatest value, every individual, it is evident, can, in his local situation, judge much better than any statesman or lawgiver can do for him” (WN IV.ii.10), his argument applies not only to legislators but to the regulators Thaler and Sunstein envision as well. Arguing that regulators should nudge us to make decisions we ourselves would make if we “paid full attention and possessed complete information, unlimited cognitive abilities, and complete self-control” assumes that those regulators know, or could plausibly know, what the choices are that we would make under such conditions. Because no one is ever in such conditions, however, it is unclear what concrete guidance such an ideal can give.

Because in fact we are all imperfect reasoners with incomplete knowledge, the Smithian would predict that in practice regulators with such a broad mandate would end up nudging us to make choices consistent with the regulators’ own preferences, goals, and ambitions, rather than with those of individual citizens, whose situations and circumstances they do not know. That, Smith argues, would constitute not only an infringement of citizens’ liberty but also, because the regulators’ preferences, goals, and ambitions are not identical to those of others, a threat to others’ well-being as well. Smith writes that the “statesman, who should attempt to direct private people in what manner they ought to employ their capitals, would not only load himself with a most unnecessary attention, but assume an authority which could safely be trusted, not only to no single person, but to no council or senate whatever, and which would nowhere be so dangerous as in the hands of a man who had folly and presumption enough to fancy himself fit to exercise it” (WN IV.ii.10). The “folly and presumption” Smith mentions here relates to his argument that I have elsewhere called the “Great Mind Fallacy,”^{Footnote 34} whereby one envisions policy or regulation with beneficent intent but that could be effectively created and administered only by someone with all the relevant knowledge of people’s localized situations. Unfortunately, no such “Great Mind” exists. Indeed, Smith claims that the knowledge that would be required to make wise decisions about how millions of people should allocate their time, talent, and treasure “exceeds all computation” (WN I.i.11). Thus, Smith’s argument holds that even policy carefully articulated by experts would likely not be able to generate the beneficence that leaving such matters to localized individuals could.

Returning now to Singer’s social-justice argument, let us ask how Smith would respond to it. Singer motivates his argument with a now famous example:

The answer, of course, is that you should wade in to save the child. But if you did not save the child, how should we characterize your inaction — as an injustice to the child, or merely as insufficient beneficence toward the child? If we hold it to be the former, then we might be justified in punishing you, even coercively; if we hold it to be the latter, however, then although we might be justified in condemning you and criticizing your decision, we would not be licensed to forcibly punish you. Smith’s NDO conception of justice would seem to disallow us from punishing you: Since you engaged in no maleficence toward the child — you did not worsen the child’s position or violate any express promise you had made to him or anyone else — Smith would say that you are “the object of hatred, a passion which is naturally excited by impropriety of sentiment and behaviour; not of resentment, a passion which is never properly called forth but by actions which tend to do real and positive hurt to some particular persons” (TMS II.ii.1.3). We may hate you, but we may not jail, fine, or otherwise coercively punish you. Singer’s denial of a distinction between killing and letting die suggests he would find Smith’s assessment unacceptable.

Singer goes on to argue that the lesson from the example of the drowning child also applies to those who fail to provide monetary assistance to children dying of hunger in developing countries, claiming that physical distance does not alter the moral character of the two scenarios.^{Footnote 36} I would like to suggest, however, that Smith’s full conception of justice can accommodate Singer’s drowning child example. As I have argued, Smith believes we should accept his NDO conception as a default, not an inviolable rule; he allows exceptions in cases when an impartial spectator would feel proper resentment (not mere hatred). The key for Smith is contained in his Local Knowledge Argument: because third parties are typically not in a position to know when or how one should provide positive benefit to others, Smith argues that legislatures are ill-equipped to delineate rules for benefit it is proper to demand. But Singer’s pond scenario is a case in which such local knowledge is in fact present: the child is right before us, we know we could save him, we know no one else will, and we know that if we do not wade in the child will drown. In a case like this, a Smithian impartial spectator might well feel not just hatred but resentment if we failed to help the child, and so this would be one of the exigent (and, thankfully, rare) cases in which an impartial spectator would allow an exception to the NDO default. But the NDO conception would deny the extension Singer wants to make of the drowning child example to cases like hungry people overseas because we do not possess in the latter case the requisite local knowledge that we do possess in the former case.

Allowing exceptions only in cases when localized knowledge is present and warrants the departure would comport with the other cases Smith describes as exceptions to his NDO default. Consider again, for example, the passage quoted earlier in which Smith suggests that a “superior may, indeed, sometimes, with universal approbation, oblige those under his jurisdiction to behave, in this respect, with a certain degree of propriety to one another” (TMS II.ii.1.8). For Smith, this is not a general obligation or license on behalf of legislators; it is, rather, a small and localized exception. This is indicated by the several ways Smith hedges and limits the obligation. Note, for example, that he says that such a superior “may” do this, but only “sometimes” and only “to a certain degree.” He goes on to specify that he is speaking about a “civil magistrate,” meaning a local political leader, not a distant legislator. And even for such a local leader, who presumably has familiarity with the relevant local situation that is demanded by the LKA, Smith contends that coercively obliging beneficence “requires the greatest delicacy and reserve to execute with propriety and judgment,” lest one “push it too far,” at which point it becomes “destructive of all liberty, security, and justice” (TMS II.ii.1.8).

Moreover, when Smith proposes in WN that the government might establish schools to mitigate the effects of extreme division of labor on workers, he is careful to argue: (1) that it should be primary schooling only — claiming that government is warranted to ensure only what all people, in any line of work, would need, which he holds to be only reading, writing, arithmetic, and perhaps some geometry (WN V.i.f.55); (2) that less than half of the cost should be subsidized from public monies, to ensure proper alignment of incentives (WN V.i.f.55); (3) that there should be a market with competition and choice among schools (WN V.i.f.4 and V.i.f.12); and (4) that, in any case, those parts of education that are privately funded “are generally the best taught” (WN V.i.f.16), whereas those that are publicly funded frequently contain material that is “useless, absurd, or fantastical” (WN V.i.f.47). All this strongly cautions against public funding for education except in those extraordinary cases when there is no other way to do it. If Smith’s NDO conception of justice would seem to rule out government-provided education — because it would constitute providing benefit to some at others’ expense, a violation of the second of Smith’s “sacred laws of justice” (TMS II.ii.2.2) — Smith’s hedges here would comport with my proposed interpretation of Smith’s NDO account that indeed allows exceptions but only in those extraordinary cases when an impartial spectator in possession of the requisite local knowledge would approve.

VII. Smithian Ultimate Justice

My suggestion, then, is that the social-justice objections to Smith’s admittedly thin conception of justice do not yet defeat it. Smith’s claims (1) that proper beneficence is possible only when one is in possession of detailed local knowledge and (2) that legislators are unlikely to possess such knowledge are plausible, as are his further claims that (3) infractions of justice are easier for third parties to perceive and rectify, (4) the government should restrict its coercive apparatus to those matters it can competently address, and (5) the government should therefore concern itself primarily with protecting NDO justice. Let me now offer an independent reason to think that Smith might have been on to something important in his discussion of justice.

Recall that Smith claims that the rules of justice constitute the “foundation which supports the building” of society, as opposed to the “ornament which embellishes” it (TMS II.ii.3.4); and that whereas society can subsist without beneficence, “the prevalence of injustice must utterly destroy it” (TMS II.ii.3.3). Smith also claims that “the rules of justice are accurate in the highest degree, and admit of no exceptions or modifications” (TMS III.6.10). But Smith also argues that the rules of justice must be held “sacred” — except when departures are required from the principles of justice themselves: no exceptions “but such as may be ascertained as accurately as the rules themselves, and which generally, indeed, flow from the very same principles with them” (TMS III.6.10). The preceding quotation indicates the dependence, for Smith, of any putative departure from his NDO conception on the presence of relevant local knowledge: the circumstances of such a departure must “be ascertained as accurately as the rules themselves.” What might the circumstances for a justifiable departure from NDO justice look like? Suppose that for a given individual A, if A’s position were unimproved by others who act with justice but not beneficence toward him, we can say that he has led a life that did not achieve the full complement of happiness he otherwise could have. I suggest that on Smith’s account, A’s life might therefore generate justified resentment in an impartial spectator who would hold that he has a reasonable general expectation that his life in human society would indeed provide him benefit, even if this expectation is not tied to any specific persons who violated specific obligations to him. The purpose of life in society, after all, is to improve upon the conditions in which we would be if we were on our own (see TMS II.ii.3.1). Hence if A’s life in society in fact provided no such benefits, it seems reasonable that an impartial spectator would conclude that his resentment toward society is justified. Yet how could this societal “indifference” or “insensibility” be rectified? Only by someone who has a perspective from which to judge the entire sweep of A’s life — God, for example.

This implicates the Smithian notion of theodicy, which he brings up several times in TMS. Here are some of the central passages: “For it well deserves to be taken notice of, that we are so far from imagining that injustice ought to be punished in this life, merely on account of the order of society, which cannot otherwise be maintained, that Nature teaches us to hope, and religion, we suppose, authorises us to expect, that it will be punished, even in a life to come” (TMS II.ii.3.12). In a passage discussing the fact that the “rules which direct” the “natural course of things” “sometimes produce effects which shock all [humanity’s] natural sentiments,” Smith writes:

Those passages address injustice that goes unpunished in this life, but Smith makes a similar argument with respect to the beneficence we properly should have received but did not. In the chapter “Of universal benevolence,” Smith writes:

Smith goes on to claim that the “wise and virtuous man” should be willing to subordinate “inferior interests” to superior ones, including ultimately “to the interest of that great society of all sensible and intelligent beings, of which God himself is the immediate administrator and director” (TMS VI.ii.3.3).

Note three central claims in these passages: (1) unpunished injustice in this life will be punished in the life to come; (2) God intends the “greatest possible quantity of happiness” in the fullness of time;^{Footnote 37} and (3) personal virtue requires taking an absolute reckoning of “interests,” across time and persons weighted equally. These claims suggest a final argument in favor of Smith’s thin conception of justice. Smith writes: “Our happiness in this life is thus, upon many occasions, dependent on the humble hope and expectation of a life to come: a hope and expectation deeply rooted in human nature. [ . . . ] That there is a world to come, where exact justice will be done to every man, where every man will be ranked with those who, in the moral and intellectual qualities, are really his equals; [ . . . ] is a doctrine, in every respect so venerable, so comfortable to the weakness, so flattering to the grandeur of human nature, that the virtuous man who has the misfortune to doubt of it, cannot possibly avoid wishing most earnestly and anxiously to believe it” (TMS III.2.33). His argument is that belief in a future life facilitates good human relations in this life, not only because it gives us comfort and solace but also because it can soften, even defeat, our felt need to punish every transgression, including each case when we were not done a good office someone should have done us.

Everyone’s life is full of instances of having been misjudged, of experiencing lack of charity, of innumerable actions that deserve disapprobation. Not all of these might rise to the level of injustice in the proper sense, but they may count as “injustices” in the broader sense that includes slights, under-appreciation, insufficient credit, unfairness, and so on. Thus, they may trigger the resentment of an impartial spectator (or Impartial Spectator — that is, God). If we were to believe that there were no hereafter, then all actions that fail to comport perfectly with virtue would have to be rectified either in this life, or not at all; if we believed that, we might be inclined to be far more confrontational, more demanding and entitled, more peevish and captious toward one another than we otherwise would be. If, by contrast, we believed that a full, accurate, and objectively correct reckoning of all virtues and vices, and our full and proper complement of happiness, would be eventually forthcoming, we would not need to be so punctilious in keeping a ledger of moral debits and credits. We could instead be forgiving, charitable, and merciful, which is far more conducive to successful relationships with others and thus far more conducive to human happiness in this life. We would, of course, also have the direct incentive to virtue that belief in (or fear of) divine retribution for vice provides, but the dynamic I am describing would apply not to the committer but to the recipient of vice; and while not, perhaps, affecting grave actions of vice, it might assuage resentment over smaller infractions, thereby smoothing human relations in a world of imperfect creatures.

If it turns out there is no God, still not all is lost. The belief in the relevant aspects of virtue contribute to peaceful and cooperative, even happy, human relationships in this life. If it is true, as seems to be the case, that injustices are not fully compensated on earth, then the Smithian position I have described of a default NDO conception that allows for rare exceptions enables us to understand what it would mean to compensate them and what it would mean to reconcile them. It would provide, then, for an ultimate justice. A proper theodicy would provide for an establishment of a full and complete distribution of justice and beneficence across all individuals and integrated over the fullness of time. It would also seem to require that God exist and be just: for otherwise there could be no reasonable expectation of full happiness. So Smith offers us the following dilemma: Either there will be a Tartarus in which the wicked are appropriately punished and an Elysium in which the just are appropriately rewarded by a just God who serves as a perfect and impartial adjudicator of all our actions (TMS II.ii.3.12); or this life is all there is. If we believe the latter, all manner of difficulties will arise both for us and for those we care about. By contrast, if we believe the former, numerous virtues will accrue to us, and numerous benefits to ourselves and others as well.

VIII. Conclusion

I have argued that Smith’s NDO conception of justice is his default conception, and that, even if it is thin, his Local Knowledge Argument not only enables it to survive social-justice objections but makes it more plausible than more expansive, social-justice conceptions of justice. If I have that right, Smith’s theory not only gives us relatively clear guidance for institutional and governmental policy, but it may, indeed, even give us a “complete” theory — one, that is, that gives us both the correct and proper guiding fundamental principles (NDO justice) and a clear criterion for determining justified deviations (when an impartial spectator in possession of the requisite local knowledge judges appropriate). Finally, it outlines a conception of ultimate justice that encourages mutually beneficial associations and even happiness in this life. The potential benefits attaching to the possibility that it might also be true make it all the more compelling. Perhaps, then, Smith’s conception of justice is not so thin after all.