Anthony de Jasay: Justice, in: The New Palgrave Dictionary of Economics and The Law, Macmillan, London 1998

The concept of justice informs our sense of justice, rather than being formed by it. The concept escapes circularity, resting as it does on foundations that are independent of notions of justice.


Answers to questions about what justice~ demands are commonly, and misleadingly, called judgments. Misleadingly, because the same word is used to denote two radically different types of statement, and much hinges on the difference.
One type deals with points of fact. They are true or false, for facts are ultimately ascertainable, and once ascertained, exclude bona fide disagreement. They are intersubjectively valid. Points of fact in a strict sense include ‘points of law’, for what a text states, what the custom of a place is, or what conduct a social norm dictates, are also ascertainable. The world being what it is, texts may be poorly drafted, evidence may be blurred or missing, custom and convention may be falling into disuse. Findings in the face of such imperfect information may not be reached without a degree of help from judgments, which then appear as the product of personal intuition about what the facts would show if only they were fully accessible. For example, at one time interpersonal comparisons of utility were represented as `objective’, more like findings to support advocacy of income redistribution. (Later, these efforts came to be seen as misdirected because the problem was no longer conceived as one of facts but of value judgments.) In their role as substitutes for information, judgments function as surrogate findings.

Judgments properly speaking are expressions of moral intuitions, sharply distinct from surrogate findings which seek to express empirical conjectures. Judgments reflecting moral intuitions answer questions of justice that are deemed not to be matters of fact. Such judgments may seem right or wrong to another, perhaps rival moral intuition, but they are neither true nor false. Nothing in the rules of logic or epistemology compels two bona fide persons to agree on the same judgment; there can be no question of intersubjective validity. A shared religion, shared value judgments, shared interests may bring different persons’ moral intuitions closer to each other, and the contrary may occur when these influences affect different persons differently. Genuine judgments are intrinsically personal, enjoying a latitude that has no very obvious limits.

Latitude in judgment means discretionary justice. The task of a firm, well-defined concept of justice is to reduce discretion in which my say-so can oppose your say-so. It accomplishes this, in a seeming paradox, by constricting the scope of judgment in justice.

It will be argued in this essay that the world of justice is neatly divided into two adjacent realms, with no overlap between them. The two realms are ruled by two regulating maxims, ‘suum cuique’ (to each, his own) and `to each, according to…’ (i.e. one reference variable). In the realm of `suum cuique’, the concept of justice leaves little space for judgments. Findings do nearly all the work. Where, on the other hand, `to each, according to…’ is the master rule, there is an irreducible role left to judgment. This is perhaps no very bad thing as long as discretion is kept in its irreducible place.
If the thesis presently to be unfolded is anywhere near right, the concept of justice requires that where findings furnish complete answers, they should be left to do so. The persistent invasion of one realm of justice by the regulating maxim of the other brings incoherence, confusion and discredit to justice.
Disorderly minds, discretionary justice. Conceptual analysis often has recourse to two sources, the behavioral and the linguistic. They furnish evidence of how people understand a concept as witnessed by their reactions and by their use of the corresponding words in ordinary speech.

By the first source, then, a just stale of affairs is one that people agree to. However, outside actual contracts, agreement may not be signified by any positive act. It may be tacit acceptance, which in turn may reveal nothing more definite than acquiescence born of indifference or impotence. An enormous latitude in states of affairs is compatible with tacit acceptance. Moreover, one and the same mind can manifest tacit acceptance of mutually incompatible states of affairs. Many theories of justice `finesse’ the problem of tacit acceptance by postulating hypothetical agreements that would be reached under specified conditions, e.g. under ignorance of one’s particular identity, endowments and interests, under uncertainty, under a desire to agree to anything reasonable, under mutual insight into each other’s intentions, and so on. However, what hypothetical persons would agree to in hypothetical circumstances, while perhaps an interesting subject of speculation, cannot reliably predict what actual people will agree to in actual circumstances, and reveals only one of many possible versions of what, they understand by `just’.
Is usage of words like `just’ or `fair’ in ordinary speech a better witness to the common understanding of the concept? People will unhesitatingly say `Due process of law should be observed (but) the wife-slayer should not get off by hiring high-priced lawyers’; `People are free to spend their own money as they choose (but) should not get better medical care by paying more for it’; or `The given word must be kept and contracts honoured (but) great inequalities of income and wealth are unjust’. The near-impossibility of reconciling the two members of these and similar pairs does not prevent them from being uttered in the same breath.

It is well known from social choice theory that it is not, in general, possible to obtain a coherent hierarchical ordering of states of affairs by aggregating, according to a plausible rule of aggregation, disparate individual orderings of such states. We may find X ranked above Y, Y above Z but Z above X. If collective entities, Such as a society, were supposed to have a collective mind, we should have to say that it was a disorderly one, unable or unwilling to sort out its own incoherences. When one and the same individual is found to hold incoherent preferences, opinions, judgments, unconsciously and without embarrassment disagreeing with himself, the disorderly mind is not a symptom of the impossibility of adding up orderly ones, but a fairly predictable feature of human imperfection.

Disorderly minds can harbour mutually incompatible judgments, and discretionary justice. This is not to suggest that people should wake up and tidy up their minds. It is to suggest that the concept of justice needs more coherent foundations than most people’s understanding of what is just.
Before justice: some foundations. Manifestly, justice is not a self evident concept that reveals itself to the intuition of everyman. To what extent can one derive it from elements more basic, more self evident than itself, and that ara independent of, and prior to, any notion of justice? If it is a composite, what are its components; if an edifice, what are its foundations?

Five foundation stones seem to beat most of its weight, doing so unobtrusively, beneath the surface. They are verifiability and falsifiability; feasibility; harm; and trust. Only briefly identified here, their role in carrying the edifice will emerge in detail in subsequent sections.
`There is a hippopotamus in the room’ is an affirmation we can both verify and falsify with the same ease. `There is a needle in the haystack’ is a statement that the person affirming it can verify, but that we can only falsify with great difficulty and, in the limiting case of the infinitely great haystack, not at all. All data statements, i.e. statements with empirical content that, if confirmed, can pass for findings, are either verifiable, or falsifiable, or both. Which they are (and if they are both, which confirmatory operation is less onerous) appears, in an obvious but little noticed manner, to govern the placing of the burden of proof on one of a pair of contradictory answers to a question of justice. The allocation of the burden of proof, in turn, gives rise to such crucial principles as the presumptions of liberty, of innocence, of title and of equal treatment.

Feasibility sets limits to the requirements of justice in the same truistic way as `ought implies can’ does to the commands of morals. A person’s feasible set of acts – what his environment and his possessions and claims on others permit him to choose – will generally contain some acts that sufficient reasons show to be inadmissible. These acts are unjust because they are inadmissible on independent grounds, and not inadmissible because they are unjust. No circularity is involved, as there would be if ruling an act admissible depended on prior knowledge of what justice required. All residual acts that have, for the time being, no sufficient reason speaking against them, are by implication, and for the time being, admissible. They are `liberties’. Infringing them is unjust because they are liberties in the absence of reasons to the contrary, rather than being liberties because it would be unjust to infringe them. More concisely, liberty is the baseline; it is the restrictions upon it that need to be justified.

Sufficient reasons for ruling feasible acts inadmissible cannot be exhaustively classified, but some can be catalogued with some certainty. One class of such reasons has to do with harm, another with trust, both of which are foundation stones in their own right of the concept of justice.

Whoever cripples me, slanders me or takes away what is mine and what I value harms me in a non-trivial way. This is not a trite proposition, and some may even consider it far from self evident, if only because neither `non-trivial’ nor `mine’ are non-controversial. As to the former, it is perhaps fair to say that when a harm is really on the borderline, the question whether it just passes muster as non-trivial is itself relatively trivial. As to the latter, what is duly `mine’ is clearly a central preoccupation of any analysis of justice. It cannot be resolved right at the outset, and must for a little while be held in abeyance. Notwithstanding any reservations on this score, harm is a robust enough idea to serve in the argument from feasibility to admissibility and hence to justice.

Trust enters into justice in a fundamental, constituent way through what one might call the belief-soliciting nature of statements. When making a statement, a speaker conveys words to a listener with the general objective of inducing in the latter a degree of belief – and the higher the better – that the statement is true or that he, the speaker, holds it to be true. The intrinsic purpose of saying these words must be to induce such a belief, for otherwise it would not matter whether it was these words or some others that were conveyed. The intrinsic purpose of a promise, a particularly important type of statement, is to induce a belief in the promisee that the promise will be fulfilled, or failing that, that the promisor intends to fulfill it. The promise is essentially a trust-soliciting statement. Breach of the promise betrays the trust that the promise was intended to induce. Contract is reciprocal promise. Whatever other reasons may be marshaled in support of the institution of contract, a self-evident one is that default betrays the trust the defaulting party has solicited. As such, it is an unjust act. It is perfectly possible to accept this conclusion, derived from the responsibility assumed by the promisor when he solicits the trust of the promisee, even if the proper remedy for the breach of promise is in dispute. An influential view in modern jurisprudence holds that when the breach of a contract causes no assessable damage, no remedy is applicable. In this view, when a contract is partly executed, remedy is restitution of the defaulter’s gain or the damage caused by the default, while if the contract is unexecuted, remedy need only be brought for damage caused by the reliance of the plaintiff on its expected execution. This view appears to excuse the breach if it causes no damage to the promisee’s interest. It reflects `rights-based’ theories of justice that derive rights from interests and obligations from rights. Such theories commingle issues of justice with those of welfare, and are not conducive to a clear concept of justice.


Principles can be identified that rest on the foundations described in the preceding section and, perhaps together with other and compatible ones that are not specifically identified, constitute justice. Three such principles seem necessary, though we should not expect them to be sufficient, to constitute the essential concept: responsibility, presumption and convention.

Responsibility. The principle of responsibility results from the relation between a state of affairs and its putative cause – a relation that is normally a matter of empirical findings. A state of affairs in which we find disabled orphans, destitute old people, young people vainly looking for something useful and gainful to do, a region devastated by flood, an industry dying on its feet is readily described as an unjust world. Calling it unjust as distinct from sad or infuriating or crying out for some help, however, is implicitly to impute it to an unjust act committed by someone at some stage. Failing such imputation, one would in effect be saying that injustices can be self-generating – loose talk at best, nonsense at worst.

In the occasionally helpful language of game theory, Nature is a player whose moves have no discernible motive. Nature does not, as far as we can tell, seek to maximize anything. Nor does it deliberately help or binder the human players who do. To say that an unjust state of affairs has come about without anyone committing an unjust act that could have caused it is saying that Nature has committed an unjust act. People sometimes do say this of Nature or of God in ordinary speech. They may even call Nature, God, history, chance, `the system’ or some similarly elusive entity `responsible’. But they do not mean the responsibility that is a ground for being called to account, and if they did, they would sound as strange as the primitive tribe that bums the effigy of the river demon for causing the flood, or the medieval villagers who punished the cattle guilty of straying into the standing crop. Such doings would be thought laughable by the very thinkers who hold that what justice demands first and above all is the undoing of the effects of chance, the accidents of heredity and history and the acts of Nature on human lives.

It is not within our scope to speculate what the Herculean task of ceaselessly undoing what Nature keeps doing would require in terms of social and economic organization. On the other band, it is pertinent to the concept of justice that the talk of evening out the work of chance would entail arranging redress, not for injustices in the sense of acts that call for restitution and retribution, but for every human act that was either helped or hindered by chance. People would have to bear residual responsibility for the consequences of their own choices. Which consequence was imputable to choice and which to chance would have to be judged, offering almost infinite scope to judgmental justice and making it conceptually so vague as to be virtually indeterminate. This may not be the chief objection to it, but for the present purpose it seems telling enough.

It may be worth making it explicit that if an act of Nature, say a calamitous flood, is held to be an injustice to the flood victims, then the actor committing the injustice cannot be made responsible for repairing it. If the injustice is to be repaired just the same, the repair must be exacted from those who had the prudence or blind luck not to build their homes on the flood-plain; but making them repair the injustice they have not committed is an injustice, suggesting that a concept of justice that demands this is incoherent, a product of disorderly minds. If the nonvictims are to be made to help the victims of the flood, some other ground than justice, e.g. some notion of an interpersonal sum of welfare, must be invoked to defend the injustice involved.

Presumption. A hypothesis and its negation form a pair such that it suffices to verify one member of the pair to falsify the other, or vice versa. A feasible act is either sufficiently harmless, unobjectionable, hence free, or it is not sufficiently harmless, objectionable, liable to sanction, unfree. A person is either innocent, or guilty. A holder either has title to an asset in his possession, or he has not. A case (raising a question of justice) is either relevantly like another case, or it is not. It is obviously redundant to prove one member in such pairs and disprove the other. The burden of proof need only be placed on one of the two, but it must be placed on one of them for a conclusion to be reached at all. Whichever side the burden of proof is placed, the opposite side is treated as the privileged hypothesis; a presumption is established in its favour. Unless the contrary hypothesis is successfully proved, the privileged hypothesis is presumed to be true and for the time being treated as such. Just as the presumption of innocence means that a man is treated as innocent until he is proved guilty, so the presumption of guilt means that he is treated as guilty until he is proved innocent. The two alternatives look symmetrical, though of course they are not. It is tempting to suppose that in our ago and civilization we put the burden of proof where we do, because opting for the presumption of innocence, of freedom, of possession, of equal treatment of like cases, is intuitively just and ranges us on the side of the angels. Supposing this, however, is to suppose that we already know what is just and which side is that of the angels. There is, instead, an independent reason that makes these presumptions prevail, due to their being asymmetrical to their opposites in a way that makes them dominant quite regardless of ages, civilizations and intuitions about justice.

Consider feasible acts. An actor proposes to perform an act. There could be an objector, an authorized representative of society or just another party speaking for himself, who might affirm that the act would cause sufficient harm to a public or a private interest to justify its being stopped, prohibited, sanctioned. The presumption of liberty implies that the act will remain free and will not be stopped until the affirmation of its objectionable nature is not only made, but also verified. The opposite presumption would stop the actor doing the act in question until he could show it to be proof against possible objection. There is an indefinitely large number of potential objectors having a potentially infinite number of objections, some of which may be sufficiently strong. To falsify the hypothesis that the act is objectionable, and therefore not one of the actor’s liberties, is a needle-in-the-haystack type of task, very difficult and costly if the set of potential objections is largo, and logically impossible if the set is not finite (which, in a strict sense, it never is). Taking the haystack apart blade by blade to falsify the supposition of its harbouring a needle would take long enough to mean an indefinite suspension of the act whose free performance depended on there being no needle. Taken literally, the presumption that every act may be harmful and hurt some interest would freeze everything into total immobility; this could only be attenuated by society granting advance clearance to a specific list of demonstrably innocuous acts, like visiting one’s ailing aunt, whose safety and lack of threat to any conceivable aspiration, wish, interest or value would not have to be established before they could be undertaken. However, opting for the presumption of liberty is hardly a matter of ethics, of a liberal temperament, or even of efficient social and economic organization. It is a matter of epistemology, of how knowledge works and how verification differs from falsification.

The way knowledge works directs, perhaps imperatively so, the burden of proof to be placed on interdictions, on the assumption of guilt, on the supposition that possession is illicit or title has some vice, and on the claim that some case is relevantly different from some precedent, some other case. The presumptions of liberty, of innocence, of rule to possession and of like treatment of like cases, do not stand on their particular merits, genuine as those may be, but are the automatic consequences of where the burden of proof falls – a matter about which sensible thought leaves little choice.

Equal treatment, with the burden of proving that two cases are relevantly unlike, has a pivotal place among the presumptions that enter into the concept of justice. If equal treatment is to be presumed, and if one person benefits from the presumption of liberty, of innocence, of good title or other favourable presumptions if there are any, then all other persons must also benefit from the same presumptions unless it can be shown that their case differs in a relevant respect from that of the first person. It is intuitively appealing that the principle of presumption is general because it is just that it should be so. However, as was argued above, its generality has an independent source in knowledge that can be recognized before any knowledge of justice. Circularity is avoided; we have a constituent principle of justice, rather than one that is derived from the requirements of a justice that already exists in our intuitive understanding.

Convention. If there existed a truly pre-social state of affairs, or one where everyone was suddenly struck by amnesia, there would be no precedents to follow and no patterns of behaviour to fall in with. Could questions of justice be resolved with any degree of predictability, or for that master at all? Take two cases. In the first, two persons board a (pre-social) bus with only one empty seat. Which of them should sit and which stand? In the second, a man’s honour is mortally offended. May he kilt the offender, perhaps by giving him a chance in a duet? `Before justice’, there is no predictable answer to either question; letting only their moral intuition speak, some people will give one answer, others another. Yet in a firm concept of justice, there must be guidelines that will narrow down the just resolution of such cases to a fairly predictable range. Once we alt know what is just, the guidelines will be laid clown in jurt laws. But if a concept of justice must precede just law, what will precede justice itself?

It is such considerations that suggest that there is a place for at least one more principle, in addition to the principles of responsibility and presumption, to enable the concept to be constituted from elements not dependent on prior ideas of justice. This principle is simply that where social conventions guide behaviour, questions of justice should be resolved according to such guidance.

Conventions are most intriguing phenomena both as regards their origin and their enforcement. Lack of space forces us to take their emergence, their functionality and their limited but nonetheless real self enforcing features for granted here. For our purpose, they fall into two classes: conventions that foster civility, and those that discourage torts.

With two passengers aspiring to one vacant Beat, they are both worst off when they fight each other over it. The next-worse solution is that they both remain standing, each fearing to look boorish if he sat clown. If one sits and the other stands, at least one of them is better off and the other is no worse off. This is the Nash equilibrium solution, where neither can improve his own situation without the other accepting to worsen his. However, there are two equilibria, depending on who sits and who agrees to stand. This inequality of outcome is a source of conflict. Consequently, it is not self-evident that one or the other of the two equilibria will in fact be reached. Many if not most conventions of civility guide the parties in such conflicts round the issue and achieve an equilibrium. With the convention of first come, first served, both passengers know that he who boarded first is supposed to get the seat. We incorporate this into justice and have little hesitation saying that it is just that the first-comer should have first pick. For certain contingencies, a different and stronger convention may supersede the basic standard: for example, the aged and the infirm may get priority over the firstcomer. Lest this story should give too Panglossian a view of social coexistence, it should be noted that conventions of civility tend to have a rather weak self-enforcing capability. The brutish driver, the lout spoiling the party, the queue-jumper Pace only mild sanctions, perhaps nothing more severe than contemptuous looks. It is also relevant to note that breaches of those conventions that are weakly sanctioned tend not to be called unjust acts in ordinary speech, and vice versa. Which way the causation runs should provide food for deep and probably fruitless thought.

Conventions against torts discourage serious harms to person and property and, more generally, non-trivial violations of the liberties of others except where this is done as a sanction employed to enforce the convention itself.

These conventions are largely self explanatory. They are ancient, as old as society itself in the truistic sense that a society is formed by its members starting to adhere to such conventions in favour of each other. They are also cross-culturally stable, most societies treating much the same acts as torts. Doing one’s bit to enforce the conventions, outlawing murderers, helping to catch the fugitive thief, watching the neighbours’ property, putting peer pressure on defaulters and lending a hand in getting contracts duly executed, used to be norms of behaviour in conformity, as it were, to a convention to enforce the conventions. Though the enforcing function has to a great extent been taken over by state authorities, vestigial traces of it still subsist in many social groups.

The substance of the conventions against torts, including the mandatory character of the enforcement usually combining restitution and retribution, forms a stable and readily grasped part of the concept of justice.


Standard usage talks of two kinds of justice, commutative and distributive. The first sees to the full commutation of deserts into rewards and punishments. The second ensures that rewards and punishments are distributed as they ought to be. Assuming that `deserts’ means, broadly, all good grounds for rewarding and punishing, so that people in a just state of affairs get all they deserve, and deserve all they get, the supposed duality of justice is baffling. If commutative justice has been satisfied, it must have duly discharged the distributive function by the just allocation of rewards and punishments, benefits and burdens. Likewise, it is hard to sce how distributive justice can be satisfied if it is not by the execution of commutative justice. The confusion has irked many logical minds, provoking them to deny that there could be any such thing as distributive justice. While this seems to be a mistaken view for reasons to be put forward in Section V, the commutative-distributive distinction is not very helpful, and obscures another kind of duality in the concept of justice. This concerns the two types of situation in which justice is a pertinent consideration.

The two kinds of situations are regulated by two complementary maxims. First enunciated by Cicero, the maxim `suum cuique’ (`Render unto each his own’) has survived as the common nucleus of two golden rules put forward by the third-century Imperial Roman legal thinker Ulpian in his Digesta.

The other maxim of `to each, according to’ is the generalized schema of `to Bach, according to his needs, from euch, according to his abilitics’, proposed by the revolutionary writer Louis Blanc in 1839 and mado famous by Marx in the Critique of the Gotha Programme. Despite appearances, the two maxims are not rivals. They regulate two separate realms, and where one applies, the other does not.

Datum telling it what is a person’s own. It is an ascertainable fact, or failing that, a presumption, that he has certain liberties, good title to his possessions and valid claims to what is owing to him under outstanding contracts. He has his own, and gets his own, as long as he and others do no more than exercise their liberties and fulfill their obligations, with their interactions being confined to voluntary exchanges and the rendering of unilateral benefits. `Suum cuique’ is breached when a person’s liberties are violated, his possessions are taken from him or trespassed upon, when his obligor defaults or when he is forced to render involuntary benefits to others. Under `to each, according to’ each person in some defined class must get some benefit, or carry some burden, according to some defined criterion. For Marx, the criterion for benefits was need, and for burdens ability, while he left opon and undefined the class of persons who shall benefit or carry burdens. But the maxim in its general form leaves undefined both the class of persons and the reference criterion, to be decided on the merits of the case. All it lays down unambiguously is that all who benefits should do so according to the same reference criterion, as should all who are made to suffer or carry burdens. In `lot all guilty persons be punished according to their crime’, both the class of persons and the criterion according to which a punishment is to be meted out to them suggest themselves and appear to make the choice obvious enough. Yet even in this case, the choice remains a judgmental one, while in many other cases it is much less obvious and leaves great latitude to judgment.

Manifestly, both `suum cuique’ and `to Bach, according to’ cut right across the somewhat questionable distinction between commutative and distributive justice. Each has to do with just distributions and the rectification of unjust ones, where `distribution’ must of course be understood in a general, all-inclusive sense.

The distinction between the two realms of justice ruled by the two maxims is a different, and very much sharper, one. Under `suum cuique’, the exercise by everyone of their liberties and the fulfillment of their obligations distributes benefits and burdens. A just state of affairs prevails unless an unjust act violates it.

Under `to each, according to’, certain benefits or burdens must be distributed to chosen persons according to a common criterion. The class of persons and the distributive criterion must be chosen justly. Unlike in the realm of `suum cuique’, the distribution here is the product of a deliberate act. Failing that act, there is no distribution; failing it being just, the distribution cannot be just. An unjust stale of affairs ensues unless a certain just act is performed.


What is a person’s own is fundamentally a question of what, within his set of feasible acts, he is at liberty to perform. By the same token, knowledge of what `suum cuique’ implies can be reached by knowing what particular persons’ liberties are. Most, if not all, of this knowledge derives from the foundations and constituting principles of justice which dictate the division of any feasible set into admissible and inadmissible acts.

Admissible acts. Before Friday arrived, any act Robinson Crusoe was capable of performing he was at liberty to perform. There was no potential objector, raising questions of freedom, no one to tell him what to do, raising questions of obligation and duty, and no one his acts would inconvenience or harm, raising questions of civility, nuisance and tort. Feasibility and admissibility coincide for the solitary person. This need not mean that all his feasible acts are morally irreproachable, that he cannot do any wrong that is not wrong on consequentialist grounds. What it does mean is that the solitary person may be restrained by ethics but need not restrain himself on grounds of justice; he cannot be unjust without being unjust to others. Nothing in the concept of justice suggests that doing oneself an injustice is anything but a figure of speech.

Admissible acts are either liberties we are free to perform unless sufficient cause is shown why we should not, or obligations we are not free not to perform if called upon to do so by the holder of the right that was created when the obligation was assumed, or duties that we have moral reasons to perform but are not obliged, and normally cannot be forced by others, to perform.

The simplest liberties are acts that are matters of indifference to others and create no externalities. My reading in my study is undoubtedly indifferent to everyone else. My walking in the wood could be a negative externality to some people who like to take solitary walks. My driving to work is almost certainly a negative externality to other road-users, as is their driving to me. However, society in its wisdom has not evolved conventions about not entering the wood when someone has gone for a walk there, nor about desisting from driving into dense traffic. In the absence of a broadly recognized convention, such acts, for all the negative externalities they generate, remain liberties. Smoking in public is an interesting borderline case; in some societies, a convention seems to be evolving against it, though the spontaneity of its emergence is questionable in view of the influence exerted by the public health authorities and by tort litigation. `Imposing a convention’, an oxymoronic idea on a par with inventing a tradition or decreeing a custom, apart from other and probably more fatal flaws, confuses the issue of how we know what is just. For an `imposed convention’, unlike ones that emerge unaided by what Edmund Burke called `artificial government’, presupposes a concept of justice and is not an independent source of knowledge contributing to it.

A different and more serious problem is posed by imposed obligations. Making a promise or entering into a contract is to assume an obligation and to solicit the trust of the promisee that his right to the performance thus promised will in fact be honoured. Not to fulfill the obligation is therefore prima facie unjust. However, certain schools of thought assert that rights can arise not only by promise or contract, but by the recognition that the right in question would serve a very important interest. Along these lines, a right to the satisfaction of basic needs, a right to work, a right to education are said to exist and may be decreed by government. For the rights to be exercised, others must be placed under the obligation to provide the wherewithal. Unless it can be successfully argued that the involuntary, coerced obligors are in fact responsible for the basic needs of others being unmet, employment opportunities or educational facilities lacking, it is an injustice to coerce them to provide redress and serve these putative rights, however important they are. If the coercion in question is to be justified, it must be on grounds other than justice, and more powerful than justice.

In addition to exercising one’s liberties and fulfilling one’s obligations, doing one’s duty completes the typology of admissible acts. It suits the structure of our argument to treat it presently in its negative form, i.e. as the breach of one’s duty.

Inadmissible acts. Acts are inadmissible on general and on specific grounds. The general ground for inadmissibility is the prevention, frustration or obstruction (without sufficient reason) of another’s admissible act – what is sometimes called interference with his liberty. Questions of deliberate intent, negligence or strict liability may enter here, affecting in particular the problem of redress and sanction. They cannot adequately be treated here. Deliberately raising the opportunity cost of an admissible act may or may not be inadmissible; this is part of the problem of coercion, and will be treated in that context. It is clear on the whole that to the extent that the reasons that make admissible acts admissible com to be integrated into the concept of justice, inadmissible acts that override these reasons are unjust.

The general ground for inadmissibility is that the act in question interferes with a liberty without sufficient reason. ‘This means that, at Ieast if sufficiency is so defined as to be empirically meaningful and verifiable, a reason may be furnished and may on occasion be found sufficient. Although the presumption of liberty prevails and therefore acts interfering with liberties are presumed to be inadmissible on general grounds, some such acts may nevertheless come to be admitted on examination of their merits. Inadmissibility on specific grounds forms part of the general case with, however, a particular reinforcing feature: the act in question is inadmissible both because it interferes with a liberty, and because it is in breach of a recognized rule, namely the social convention whose function is to protect that and other liberties. Thus, my stealing your money is inadmissible both because it deprives you of the chance to do certain things you would have otherwise been at liberty to do, and because it breaks the convention that forbids stealing. Roughly speaking, the first ground is consequential, the second deontological. Specific grounds for inadmissibility are, as it were, fully catalogued in the set of conventions in defence of civility and against torts. Anyone living in the society in question has ready access to the catalogue. If there can be sufficient reasons for interfering with a liberty, it seems immaterial whether they test on general or on specific grounds. All one can say to reaffirm the difference between them, then, is that specific norms of conduct including specific interdictions, are not meant to be examined case by case on their merits.

Coercion calls for particular attention among inadmissible acts. Suppose racketeers offer protection against burglary to all shops in a street. Some agree to pay protection money. Those who do not become more attractive targets for the freelance burglars, and suffer increased losses. This raises for them the opportunity cost of not paying protection money, but contrary to how this predicament will be loosely described, does not `force’ or `coerce’ them to pay it. Coercion begins when the racketeers threaten the recalcitrant shopkeepers with burglary or worse unless they start paying. Short of threatening to commit a tort, the racketeers have not exercised coercion.

Property as a liberty. It seems self-evident enough that an act a person can perform and which is not inadmissible is one of `his’ liberties. `Suum cuique’ applies unambiguously. It is equally easy to grasp how and why a person’s life and limbs, his good name, his personal belongings are `his’ and what justice requires with respect to them. `Property’ in a wider sense, going beyond personal chattels, and including the freedom of disposition over valuable assets, the freedom to exclude all others from access to them and to the income they produce, while perhaps no harder to grasp, is notoriously controversial, if only because many people feel less compunction about coveting, and wanting to redistribute to themselves or to others, the `impersonal’ assets of some owner than about taking his personal belongings or the money in his wallet. What `to each, his own’ dictates becomes apparently less obvious and unequivocal as the property becomes larger, more abstract and more distant from the putative owner’s daily life. Yet there is nothing intrinsic in property that would make it more or less legitimate according to its kind or size. The concept of justice speaks with equal clarity to its various manifestations, private or public, large or small, tangible or intangible, earned or inherited.

Owners acquire title to assets by two main means. One is by not consuming current income – what an earlier generation of economists approvingly called `abstinence’ – the other by voluntary exchange or the acceptance of bequests and gifts. We must leave the third means, the finding of unowned resources, discoveries, inventions on one side for a moment.

Not consuming current income is prima facie not unjust, but is it perhaps unjust to have the income in the first place? Many would say that it is unjust if largo enough to permit so much saving. However, since current income is the produce of voluntary exchanges including the exchange of one’s personal exertions, and of income-producing assets, if these can be justified, then income need not separately be justified. Assets acquired from a previous owner in voluntary exchange, giving value for value received, is an exercise of their liberties by both and as such fully respect the command of `suum cuique’. What, however, of the previous owner? And what of the owner before him? There is an old and familiar argument that at the end of a long chain of voluntary exchanges, each of which is as legitimate and just as the one preceding it, we ultimately arrive, at the end of the chain, to an acquisition that was not an exchange with a previous owner. It is then argued, though this does not follow at all from the chain having an end, that unless the first acquisition at the far end of the chain was justified, the whole chain unravels and property loses its justification. It will be borne in mind through, what follows that this reasoning rests on the presumption that possession, far from permitting suppositions about title, places the burden of proof of title on the possessor. In the chain argument, title is presumed to have a vice until the contrary is proven. This remains a requirement through the length of the chain.

Along these lines the Western, and especially the Anglo-American, doctrine of property is wedded to the Lockean tradition, in which first acquisition is by taking first possession of a resource, say land, and where this is justified only on condition that the well-known Lockean proviso, namely that `enough and as good’ is left to others, is satisfied. Locke himself had no doubt that his proviso would in fact always be satisfied, both by virtue of the open frontier whose closing he did not foresee, and by virtue of enclosed land, as he put it, producing ten limes or more of the produce of common waste. Perhaps comforted by his belief that what is valuable is not necessarily scarce, he did not seem to appreciate the logical quandaries his proviso creates.

In the first place, if resources are finite, any discovery, any finding and appropriation of a resource by the first individual reduces the probability of an equally valuable finding by Bach subsequent seeker. This tan be Been cither in terms of a declining marginal value of finds, or an increasing marginal finding cost of finds of equal value. With finite resources, the Lockean proviso cannot possibly be satisfied. The plain man, who: finds that there is not `enough and as good left’ for him as some others possess, need not be able to appreciate the fine logical point to be convinced.

In the second place, however, one must ask the more fundamental question: why ever should one accept the Lockean proviso as a condition of the justice of property? Speaking in terms of land, if an individual must safeguard the interests of others before taking first possession of a tract of land, this means that the others had some interest in that tract, i.e. that is was not unowned, but had already been appropriated by them in some way. In equity, he must compensate the others; yet the others cannot object to his appropriating the tract if, in equity, they are compensated. The situation is manifestly one of a joint tenancy. The first individual is one of the joint tenants, and the tract he takes out of the joint tenancy must be small enough to leave the remaining joint tenants as well off on their remaining land as they were before. It is only this situation that makes the Lockean proviso jurt.

However, imputing some kind of prior ownership to a joint tenancy just shunts the conundrum a little way off, to be faced again. For how does the joint tenancy justify its appropriation of the land? Only two ways tan be tried. One is by showing that it has satisfied the Lockean proviso and taken the land out of an antecedent, and more encompassing, joint tenancy by leaving to the latter `enough and as good’ as it took. This, in turn, begs the question of how that more encompassing joint tenancy came to own the land, and leaves the Lockean proviso dangling from the end of an infinite regress. The other way to try is to assert that while the individual, taking first possession, owes a liability to a joint tenancy, the joint tenancy owes no liability to anyone because the land it had taken possession of was unowned, unclaimed, unencumbered by anybody’s prior interest. Accepting this purely hypothetical assertion is to give the benefit of presumption of title to the joint tenancy while refusing it to the individual who had taken first possession. The joint tenancy is presumed to have good title, while the individual is expected to prove his good title by showing that he has satisfied the Lockean proviso. If this is its last line of defence, the doctrine of property had better not be based on the Lockean proviso.

Intellectual property merits a minor digression at this juncture. If one creative act, e.g. an invention, does not reduce the probability of an equally valuable one by someone else in the future, the Lockean proviso with respect to intellectual property is meaningless, for how can you fail to leave `enough and as good’ of infinity?

For property of all kinds to find its place in the scheme of justice, it needs a doctrine that is neither impossible to observe when resources are finite, nor nonsensical when they are infinite, nor unjust in distributing the burden of proof.

`Finders, keepers’ is a starting point that, perhaps clone among starting points, is a question of ascertainable facts and does not beg questions of justice in order to draw conclusions about justice. To find something valuable is an admissible act as long as no one can show a valid prior claim to the find; it benefits from the presumption of liberty. It also benefits from chance, over which justice neither can nor must pretend to rule, and it is protected against arbitrary challenges by the convention of `first come, first served’. Where valid prior claims exist, acquisition is justified by the voluntariness of exchange. Incremental acquisition of property means saving from income and is justified as an admissible act, a liberty, if the income itself is justified and if income is understood at largo, to include windfalls. Bequests raise problems when they are mandatory, but if both the leaving and the accepting of them are free, they can be assimilated to voluntary exchanges, as can gifts for the same reason. If finding and appropriating what is unowned is a liberty, abstaining from consumption is a liberty, and voluntary exchange is a liberty, then property is a liberty. Under `suum cuique’ a person’s property is his if and because the acts that led to his possessing it were his liberties.

If this is established, the problem of property without possession is easy to fit into the concept of justice. Property is a liberty, or more precisely a set of liberties that the owner can choose to exercise. He is also at liberty to surrender such liberties, by voluntarily exchanging them for value received. He may, for example, lot the use of his house to another in exchange for rent. By entering into such contracts, others acquire certain property rights and assume certain obligations, and so does he. His former liberties become pairs of obligations-and-rights, which are the reciprocals of the rights-and-obligations of his contract parties. Speaking of `property rights’, instead of simply of property, draws attention to the presence of as yet unexecuted or partly executory contracts of this kind, signifying undischarged obligations.

Duty and its breach. An obligation is a relation between two persons, the obligor and the obligee or right-holder, on The one hand, and an (onerous) act on the other that the obligor must perform at the right-holder’s option. The relation is a consequence of a special kind of reciprocal promise, the contract whose existence and terms are normally verifiable. Breach of contract is inadmissible and an injustice. The conventions against torts provide for remedies against breaches of contract at least within the bounds of unconscionability. What, however, is a duty and how does it arise? One of our less errant and more uniform mora) intuitions leads us to say that parents owe certain duties to their dependent children, employers to their employees, officers to their men, judges to the accused, kings to their subjects, and that these duties are matters of justice in a different way from obligations. Duty appears to be a complex consequence of a relation between a person in authority and others depending from it, with the authority exercised in the distribution of certain benefits or burdens, rewards or punishments among the dependents in a jurt manner. What inserts duties into the concept of justice is that only some distributions pass for just.

There is good reason, in the presumption of like cases requiring like treatment, why we hold that it is unjust for a parent to overfeed one child and starve another, for the teacher to give teacher’s pet better marks than his course work deserves, or for the officer to assign the dangerous or unpleasant mission always to the same men. The parent, the teacher, the officer are in breach of their duty, yet the victims of the unjust use of their authority, unlike parties to a contract who have, so to speak, bought rights to the performance of certain obligations, have little recourse beyond grumbling, sullen resistance and the hope of one day getting their own back.

At best, the dependent recipients of a distribution by authority have a legitimate expectation that the authority will be exercised in a just manner. But if it is not, since they are not being deprived of anything that was theirs to begin with, `suum cuique’ does not take care of their case.

In an attempt to repair this gap, it is sometimes said that relations of authority and dependence, command and obedience must really be understood as tacit contracts between the parties concerned, one agreeing to respect authority in exchange for the other exercising it in a non-arbitrary manner, though what the latter condition is meant to entail may call for further definition. A classic version of this attractive metaphor is the social contract between ruler and ruled. Evidently, if such relations were as good as explicit contracts, the requirements of justice could be derived from them; all distribution would fall into the realm of `suum cuique’ and ensuring its justice would be reduced to contract-enforcement. With voluntary cooperation and command-obedience brought under the same rule, the concept of justice would be a simpler one than it proves to be when it is stripped of such pleasing fictions.

Where `suum cuique’ stops, ceasing to offer reliable or indeed any guidance, `to Bach, according to’ takes over to fili the void, though its manner of doing so is neither above bona fide contestation nor always predictable.


When the parent has pocket money to give his children, the teacher papers to mark, the officer duties to assign and the judge sentences to hand down and the government taxes to Ievy, the salient common feature is that without certain persons in authority performing certain acts within their competence, the children would get no pocket money, the papers would not get marked, the fatherland would not be defended, criminals would not be sentenced, and the tax burden would not be allocated. Under the rule of `suum cuique’, at the end of a day of liberties exercised and obligations discharged, everybody has got his own and nothing is left to distribute. Under `to each, according to’ things fail to get distributed unless someone sees to their distribution. He can see to this in infinitely many different ways, for the rule he must follow, namely `to each member of some class of persons, distribute the benefit or the burden according to something they have in common’, is so general as to be virtually useless as a constraint. Can one be any more definite about which distributions will be more just than others?

Is there a presumption of equality? The same questions of justice must have the same answers; it is self evident that like cases must be treated alike. However, every case is liko every other case (for otherwise they would not all be cases), yet every case is also unlike every other (for otherwise they would not be other cases but one and the same case). Likeness in some respect is coupled with unlikeness in some other respect. Which respect is relevant?

The general proposition that the case of one person differs from the case of another in some respect that is relevant to what share each should get in some distribution is a needle-in-the-haystack hypothesis, unfalsifiable if the number of respects that could be relevant in a distribution is undenumerable. Therefore the burden of proof cannot be borne by the negation and must be placed on the assertion of a relevant difference between cases. This appears to establish the presumption of equality, interpreted to mean that until some difference between the cases of two persons is proved relevant, they should get the same share. Such a presumption, in turn, may partly explain why the idea of equality carries a connotation of justice, albeit in a somewhat vague and diffuse way.

However, there is some misfit between this presumption and justice. `To each (child), the same (pocket money)’, `to each (student), the same (marks)’, `to each (athlete), the same (laurels)’, `to each (criminal), the same (sentence)’, `to each (patient), the same (medical attention)’, `to each (poor family), the same (allowance)’, `from each (taxpayer), the same (tax)’ makes for an implausible and incongruous series of recommendations that radical egalitarians might be the first to protest against.

Equiproportionality. Clearly, if the presumption of equality means `to each, the same’, its general application would yield bizarre results. We would all want athletes to get laurels according to their prowess, and at least some of us would want good students to get better marks than bad ones. We would usually think it just that sentences should be the heavier the graver the crime, and with that serious illness should get more medical attention than a head cold. What appears to have gone wrong with `to Bach, the same’ is that it is a special case of `to each, according to’ where the relevant reference criterion, that triggers off the. distribution of a reward or a penalty, is not a quality, condition or circumstance of each person’s case, but the person himself. It is the athlete and not his prowess, it is the patient and not his illness, the family and not its need, the criminal and not his crime. In this special case, the chosen reference criterion is the person, and since the reference criterion is a constant as between the cases to be treated, the result is also a constant: `to each, the same’.

Where the reference criterion a variable, such as prowess, scholastic achievement, illness, need, crime, exertion at work, capacity to pay taxes, or any other among the variety that may be thought suitable as the `according to’ of a distribution rule, the result would of course also be a variable. For the general form of the rule, to which the presumption of equality guides us, is that in each distribution there shall be an equal proportion between the reference variable and the distributive share it triggers off. The generalized form of the title produces equiproportionality (`Aristotlean equality’), of which absolute equality is a special case that will obtain when the reference variable is held invariant between persons.

Nothing in `to each according to’ determines how the blanks shall be filled, who shall be included in the distribution and what reference variable shall be taken as relevant. The title leaves the filling of the blanks ultimately to moral intuition.

To find that intuition enters dominantly into deciding what is just, is to see the flashing .of warning lights. Self evident propositions, findings of ascertainable facts, the logic of the burden of proof, agreements and conventions go to shape a concept of justice that is largely unequivocal in resolving questions arising from `suum cuique’. Given the facts of the case, there is a vast area of determinacy about who owns what and what others owe to him. Just acts lead to just states of affairs without seeking to do so. Manifestly, however, this ceases to be true when distributions, instead of emerging as the by-products of the sum of just acts serving other purposes, must be consciously chosen by a distributor. The choice is left indeterminate in two of its variables, leaving it to more) intuitions, value judgments and perhaps also to partisanship, ideological fashion or sheer opportunism to decide what shall be deemed the jurt distribution.

When a distributor distributes something, all the concept of justice tells him is that he must treat like cases alike, different cases differently, i.e. follow the presumption of equality. However, for each set of possible recipients of a distribution of a benefit, positive or negative, there are indefinitely many ways of following the presumption of equality. Should everybody do military service, or only the young, or only able-bodied young males? Should family allowances be means-tested? Should all students get scholarships, and should they all get the same% – and if not, should grants vary according to parental income, some measure of honest plodding, or innate ability and the prospect of glory to the school? Formally, equality is satisfied by each and every one of these alternatives, and by none more than any other. And how else can one satisfy equality, if following the title satisfies it only formally and doing it formally fails to warm the heart?

The latitude left to decisions about who shall share in a distribution, and according to what his share should be fixed, leads to a predictable consequence. The value judgments, and related uses of moral and emotional discretion that enter into the rival answers different people finally give to these questions, are mistaken for parts of the concept of justice itself.

All is there to be distributed. The vast bulk of the world’s goods, tangible and intangible, is produced and distributed as a matter of course as and when liberties are exercised and mutually agreed obligations are discharged. In this, the realm of `suum cuique’, no distributive decisions as such need, or indeed can, be taken. The question of the justice of distributive acts can properly speaking only arise in the realm of `to each, according to’, where judgment and discretion are dominant and where theories of justice may play a role in educating, and perhaps explaining if that is indeed possible, our moral intuitions that inspire them.

Contemporary theories of justice, however, tond to maximize their scope by obliterating `suum cuique’. The space thus vacated becomes one where everything worth having or escaping, and that is transferable among persons, is necessarily the subject of a deliberate distributive decision that is either just or unjust. Where `suum cuique’ is tacitly or overtly laid aside and distributions are not prejudged by it, they can and must be judged. All things function like cakes waiting to be sliced and shared out by respecting the presumption of equality. One theory may favour one reference variable as the basis of distribution, the other another, but all have in common the starting point o.f the cake that is just there. Nobody had to provide the wherewithal for it, nobody had to bake it, no prior claims attach to it, and its distribution would be unjust or fail to take place altogether unless it was effected according to a just act.

The move to take an unowned cake, baked without anybody having been responsible for baking it, as the natural starting point may be overt or veiled. An example of the overt move is to propose to deduce justice from the solution of the problem of an agreed sharing out of manna from heaven. Covert moves are less blunt. Perhaps the best-known one is to take it that fair-minded people will accept, as a matter of fairness, principles of justice they would agree on if they ignored all the advantages with which nature or luck had endowed them. Given such a veil of ignorance, it is not known who contributed how much to producing the cake, and it is equally unknown who must contribute how much to reproduce it once it is eaten. Finally, `suum cuique’ was set aside by some theories by claiming that the past and present exercise of liberties and rights does not establish filie to property. For one thing, it is contended that it is society, by a collective effort, that provides for individuals the legat framework they need for exercising liberties and rights. No one could own property if society did not protect it from predators, and no one could earn income if society did not enforce contracts. Any distribution is really willed by society, and it is incumbent upon society to bring about a just distribution by a just act. For another, pace the marginal productivity theory of income distribution and Euler’s theorem, it is argued that what people get in current income is no indication of what they really contribute to the social product, nor is the pattern of wealth ownership evidence of past contributions to the stock of wealth. Everybody’s share is owed to everybody else’s contributions in effort, saving, invention and the transmission of experience from Day One to the present. The stock of material and moral wealth is one vast positive externality, as is the current product. It is said to be meaningless to pretend that anybody has contributed any particular quantity of it and as a result owns any part of it: there is no `suum cuique’. Instead, there is a social obligation to distribute goods and bads according to proper criteria conforming to the presumption of equality what would otherwise be distributed randomly and unjustly.

It is in these various ways that all that is valuable and transferable is assimilated to the basic fiction of the cake that nobody baked and that needs cutting into just slices. `To Bach, according to’ is extended over all aspects of life where justice can have any relevance. Being judgmental and indeterminate in specific content, it is invested with criteria that would indeed have plausibility in a fictitious situation, or to which fictitious persons might indeed agree in such situations. The interest and significance of the theories that set aside `suum cuique’ and make `to each, according to’ the universal rule is the greater, the less importance is attached to real facts and notably to real, as distinct from hypothetical, agreements. However, the importance we do attach to such things is not altogether a matter we decide at our pleasure. Its short anchors in logic and epistemology, conventions and agreements allow justice but little leeway.