I. HART'S DISTINCTIONS

In his classic work, The Concept of Law,Footnote ³ Hart initially contrasts criminal laws, which can be considered “coercive orders,”Footnote ⁴ with “laws which give powers”Footnote ⁵ (such as rules for making wills,Footnote ⁶ contracts and marriages,Footnote ⁷ or city ordinancesFootnote ⁸), the second type of rule being, in Hart's view, facilitative rather than coercive.

This is all well and good with regard to acts that stay within the limits of the conferred legal powers but not always with regard to acts that exceed those limits. Hart misleads us when he goes on to suggest that there is always a “radical difference in function”Footnote ⁹ between the criminal law, which is “designed to suppress” conduct, and power-conferring rules, which “merely withhold legal recognition”Footnote ¹⁰ from acts that are ultra vires: “nullity can not . . . be assimilated to a punishment attached to a rule as an inducement to abstain from the activities which the rule forbids.”Footnote ¹¹ Hart fails here to mention that legal nullification or invalidation is capable of being used very effectively to control those who possess the power to generate legal norms.

Curiously, in his later discussion of how the highest rule of recognition (or constitution) can limit legislative powers, Hart focuses precisely upon what he previously ignored:

Invalidation serves here as a functional alternative to punishment, leading the legislature to stay within its prescribed sphere. The constitution makers have used nullity as a means of controlling the law makers.

Disabilities are in fact a widely effective way to restrict the exercise of powers, even though Hart never applies this insight to the lower levels of those empowered to generate legal norms (e.g., contract and city ordinance makers) that he discusses initially. The present essay fills this gap by showing the dominating impact of invalidation on all types of norm generators and then (in its last sections) suggests that Hart neglects to discuss this impact because he sometimes conflates restrictive invalidation and tolerant nonvalidation.

II. PUNISHMENT AND INVALIDATION

Let us begin with a simple model. Suppose Captain C has plenary legal power over Sergeant S, who in turn has plenary legal power over Private P. (Assume for simplicity that no moral or other nonlegal norms exist among C, S, and P.) Suppose further that C thinks P is becoming too exhausted from the daily hikes ordered by S and wishes to stop tomorrow's hike from taking place. C decides to do so by imposing a duty on S; C tells S, “You shall not order P out on a hike tomorrow.” Let us also assume that C and S understand that S will be punished if S does not comply with C's order. (Suppose also that the duties S imposes on P are backed by threats of punishment by S for noncompliance by P.) Note that, by hypothesis, C has here decided to control S solely by creating a duty for S, not by imposing any disability on S.

If S nevertheless orders P out on a hike, must P obey? The answer is yes, for S's power over P has been in no way limited. S has a duty not to send P out, but S has no disability that would prevent S from doing so. What will then happen if S disobeys C by ordering P out, and P disobeys S by refusing to go? Both S and P will have violated a legal duty, so both will be punished by their respective superiors. This in fact is probably how things would work out in the military.

What if C could have taken a different tack, more civilian than military in ethos? Let us imagine that instead of imposing any duty whatsoever on S, C decides to protect P solely by cutting back on S's power, imposing a Hart-type disability on S by saying, “You have no power to order P out on a hike tomorrow.” If S still orders P out, P has no duty to obey. So what happens now if P refuses S's invalid order? Neither S nor P will be punished, for neither violated any duty.Footnote ¹³ This, indeed, is how we ordinarily deal with state (or provincial) legislators who violate federal (or national) law when they purport to pass statutes. Since the state legislators are legally disabled from passing statutes in violation of federal legal rules, such statutes would be declared invalid. But neither the legislators nor any citizens who refuse to obey the legislators would be punished.Footnote ¹⁴

There appear to be many advantages—as well as some disadvantages—to disabilities-cum-invalidation over duties-cum-punishment as a method of controlling persons (or institutions) like S that are mid-level norm generators. For one thing, invalidation would seem to be cheaper in that imprisonment is avoided. Furthermore, effective invalidation requires a more open and participatory society in which P knows of the right to refuse S's order. Duties are enforced top-down while disabilities are made effective largely bottom-up. P may also prefer S to have a disability (rather than a duty) because a disability provides P with a secure “immunity” (Hohfeld's term) against any S-commanded duty to hike rather than just a freedom from such a duty, a freedom that is precarious because it is revocable at any moment by S (despite any duty C has imposed on S). On the other hand, a duty-only approach (without any imposed disability, as in our first scenario) may be better for efficiency, secrecy, and other military values. We shall see that the duty-only approach also has the advantage of securing legal independence or pluralism, as we later examine a legal world far removed from military precision. In the “dualist” conception, as we shall see, international law is a law that imposes only duties, not domestic disabilities, upon states.

However, rather than exploring all the advantages of imposing disabilities and/or duties, this essay focuses upon only one comparison. It seeks to understand whether punishment or invalidation tends logically toward greater domination of intermediate norm generators like S. At first sight, punishment may seem the harsher form of domination, because seemingly the most painful. Indeed, as we see above, Hart denies that “nullity” can even be considered a sanction, which of course is correct in the sense that it is ordinarily not a separable penalty added to the criminal act. Furthermore, Hart often minimizes any restrictive effects of disabilities.Footnote ¹⁵ And after all, the person solely under disabilities remains at liberty to behave as he wishes; his acts just do not generate duties for others (or for himself).

Yet does having duties not leave S with a greater degree of norm-issuing freedom? If no disabilities are imposed on him, S may still either create a duty for P (and maybe get punished) or else not create such a duty. S gets to choose. Increased certainty and severity of threatened punishment will, without any doubt, yield greater empirical compliance by S. However, threats of future punishment can never amount to control in the absolute sense of the physical forcing of a person's body to do someone else's will or otherwise making disobedience impossible. Even in the face of the maximum possible threat—certain and eternal damnation—the one choosing whether to sin can claim “I am the master of my fate: I am the captain of my soul.”Footnote ¹⁶

Invalidation, by contrast, restricts norm generation absolutely. If S is stripped of his power to create a duty for P, the creation of a duty for P becomes impossible for him. He is forced to conform to C's wishes. He can no longer choose to be an outlaw. This is a clear loss of freedom for S—and perhaps of dignity as well, even if P does not decide to act cockily and strut his new rights. Going back to religion again, when God told Adam and Eve that they would be punished if they ate the forbidden fruit, they were still left with free will—with the choice to sin or not. Invalidation is the equivalent of God putting a force field around the tree, stopping them absolutely from eating its fruit. The force field would have wholly deprived them of freedom to pick the fruit and of any dignity that comes with that freedom. In terms of liberty, it is less restrictive to have one's hands slapped than to have them tied.

Moreover, as a conceptual matter, punishment in itself (without invalidation) results in greater legal pluralism—more independent legal systems or spheres—than does invalidation in itself (without punishment). Where C subjects S only to duties (even backed by certain and severe punishments), one need look only to S in order to know P's duties. There may be two separate legal spheres quite independent of each other, one centered on C and the other on S. By contrast, if C uses disabilities to control S, S will have a reduced sphere within which S can shape the law for P. In order to know P's duties, one must now refer to the acts of C as well as to the commands of S, for some of the latter may have been invalidated by the former. For C to convert his orders from duties to disabilities would thus be to abrogate much of S's ability to create an autonomous legal sphere. The pluralism left intact by C not imposing disabilities on S should be favored by the military because instant obedience by P is an important good and P can more quickly and easily discover his duties if he need not look beyond the commands of S.

The preservation of such legal pluralism likewise must be a primary reason why countries such as the United Kingdom have traditionally adhered to the “dualist” understanding of international versus national law referred to above. Precisely insofar as international law can impose only duties, it can never in any way affect the British Parliament's monopoly of domestic lawmaking power. Under a dualist regime, international treaties have no domestic effect unless and until they are implemented by the national legislature. By contrast, a “monist” approach to international law allows international norms to supplement and even to invalidate national ones, thus creating a single legal universe with concomitant loss of national sovereignty.

Since some of us may not have much sympathy for sergeants, let us leave our simple C-S-P model and look further at the analogous ways in which invalidation can diminish political independence among nations. The legal history of the European Community offers us a powerful example of how the shift from duties to disabilities changed independent nations into subparts of a new legal entity. Such a seismic shift is by no means unthinkable in other regional economic associations as well and even on a world scale through organizations such as the United Nations and the World Trade Organization.

From its inception, the European Community treaty contained legal duties imposed on its members as well as various enforcement mechanisms against violators of those duties.Footnote ¹⁷ The fact that these enforcement mechanisms were originally envisioned by the signatories of the treaty shows clearly that the signatories assumed that member states might violate the treaties, that is, that violation was possible. However, in 1963 and 1964, in the Van Gend en Loos and Costa cases,Footnote ¹⁸ the European Court of Justice took it upon itself to convert certain of those duties into disabilities, making violation of the treaties impossible.

The specific duty involved in the Van Gend en Loos case was the duty not to raise tariffs on imports from other member states. The Court found that by reclassifying a certain good, the Netherlands in effect had increased its import tariff on that good. No problem so far. According to the treaties, the Netherlands was thus subject to being prosecuted for its treaty violation. But the Court decided that the rule against raising tariffs would be more “effective” if raising tariffs were converted from something prohibited to something impossible. The Court declared the increase in tariff to be simply invalid; the importer was told, in effect, that he need not comply with Dutch law, or rather, that the Dutch law in question did not really exist. The Netherlands did not have to be condemned as a lawbreaker because it was disabled from breaking the law no matter how hard it tried (and so the treaty procedures for a posteriori prosecution lost much of their point). The next year, in Costa, the Court reaffirmed its supranationalist position with greater clarity, and fourteen years later, in Simmenthal,Footnote ¹⁹ it made clear that even rules long found in national constitutions have no domestic legal effect if they contravene subsequent treaty norms.

Ordinary legal reasoning would have held that if there were a conflict between an earlier domestic constitutional norm and a later international treaty norm, the treaty norm would be the one that would be invalid under domestic law, although (according to dualist reasoning) the nation in question might still be subject to condemnation and possibly a penalty under international law.

One might pause here to note an outstanding feat of judicial activism: the European Court of Justice created a new polity by changing an international treaty into a constitution.Footnote ²⁰ But let us keep our attention fixed on the Netherlands. European Community law penetrated and took possession of Dutch law, making Dutch dissenters (here, those who objected to paying the tax) the agents of Europe. From that point on, European norms would be directly applicable and supreme in Dutch courts and in all other domestic institutions, and the European Court would defend the rights of Dutch individuals against the Dutch people.

Under this new legal regime, the European Community became able to dominate the Netherlands far more “effectively” (as the European Court aptly recognized) than it could have done by means of mere duties. Whether this is good or bad depends to a great extent on how much we care about Dutch political and cultural unity and independence, on the one hand, and the individual freedoms of Dutch traders, on the other.

We can, however, make some generalizations. It is no doubt true that the person on the ground, the individual citizen, has more freedom from duties under a system in which intermediate authorities are limited by invalidation, as compared to one in which they are limited only by punishment. Invalidation inherently favors individual liberty to behave as one wishes insofar as it reduces the total number of duties with which individuals have to comply by invalidating some of them, cutting down on the duty-generating power of some of their political superiors. It is tautologically true that limiting government is good for liberty from government.

At the same time, invalidation sharply curtails group and even individual autonomy, for to be auto-nomous (self-lawgiving) would seem to include precisely the ability to create duties for oneself. If only the threat of punishment is imposed on a group, it can still decide to resist its overlord as long as it is willing to chance suffering the prescribed sanction. If the duties the group (e.g., the Dutch people) decides to impose on itself can be invalidated from above, the group has lost its autonomy. Moreover, a degree of individual freedom has been lost as well: the freedom to effectuate legal change through the legal process. A Dutch voter can no longer aim to reclassify imports in a way that raises any tariffs, if such acts are legal nullities.

Individual autonomy (binding oneself by a rule) can suffer still more seriously from invalidation. Consider the power to contract, a form of private lawmaking. Even if contracts were punished, as long as they remained valid, the choice to create legal duties for oneself would remain open. For example, if contracts to pay gambling debts were no longer invalid (as they usually are now) but were instead merely punishable, individuals could still bind themselves by law to pay such debts. (Of course, if the punishment were severe and certain, they would be unlikely to exercise this lawmaking freedom.) Again, where marrying without a license (e.g., so-called “common-law” marriage) is illicit but still valid (e.g., in the law of equity), perhaps even punished but not nullified, individuals retain more ability to marry unofficially than they would if an unlicensed marriage were wholly invalid. Likewise for bigamy: if a second marriage were possible but punishable, the ability to have multiple spouses would be far greater than it is now, second marriages being null. Similarly, if same-sex marriages somehow were prohibited and punished (even by a term in jail) but remained valid once illegally begun, gays and lesbians would obviously have much more power to bind themselves in marriage than under today's regimes in which their marriages are an absolute nullity.Footnote ²¹

I conclude that invalidation is an enormously powerful tool of domination, more powerful than punishment alone as a means to limit the autonomy of groups and even of individuals. Insofar as we may wish to preserve pluralism in law and legal culture, we must be careful to distinguish invalidation from punishment and to use the former with caution as we build new legal orders in the world.

III. NONVALIDATION

Anyone familiar with Hart's initial discussion of legal powers and their limits will find it hard to imagine that he and I are describing the same legal phenomenon. As we see briefly above, Hart at first depicts legal powers as wholly beneficent, as simply intended to facilitate decentralized law-making. So he says that the purpose of requiring two witnesses for a will's validity or reciprocal consideration for contractual validity is just to give people easy-to-follow recipes for making rules that the government will enforce. Moreover, the disabilities implied here (no legal validity without two witnesses or without consideration) are not intended to stop a dying person from informally communicating her wishes to just one witness nor to stop informal promising without any reciprocal benefit.

Hart is indeed correct that his early examples do not have the suppressive aims or consequences discussed above in the hypothetical example of the military superior and the actual example of the European Community. It would be very strange to see the rules for wills or contracts as draconian means of making testation-without-two-witnesses and promises-without-reciprocity impossible. No one seeks to put an end to deathbed requests and extralegal promises. The law does not validate such informal acts but it has nothing against their having social force. Yet, as Hart later discusses, when a constitution imposes limits on legislative power, such as a rule against raising import taxes, its purpose is to make such ultra vires acts ineffective, to deprive them of all social force, just as C's disabling of S was meant to stop P from having to go out on a hike. The nature of this difference, I maintain, can be grasped by understanding the first discussion in Hart's classic work to deal with nonvalidation and the second to deal with invalidation.Footnote ²²

Invalidation occurs when one normative order declares the rules of another order to be no longer internally binding, to be nullities within that other order. Nonvalidation occurs when a normative order refuses within itself to recognize and effectuate the rules of another order but does not interfere with continued recognition and effectiveness of those norms within that other order.

Thus, for example, invalidation occurs when C declares that S has no power over P, perhaps even implying that P can come to C for a remedy if S nevertheless should attempt to use illicit force against P. Nonvalidation of S's order would mean only that C will not treat P's refusal to hike as an offense against C. Again, invalidation occurs when the European Court tells participants in the Dutch legal order that tariff increases are no longer to be treated by them as valid within the Netherlands. Nonvalidation would mean only that Dutch domestic tariff reclassifications do not count as valid accommodations of the Netherlands's international treaty obligations.

Nonvalidation of promises made without consideration lets them retain whatever normative force they have as a moral matter; they are simply not enforced in court. Another example: a secular state may refuse to validate religious marriage vows legally without necessarily seeking to invalidate them from a social perspective. Only if it sought to do the latter would the state aim to interfere radically with the autonomy of the moral orderings formed by such nonlegal vows.

Paradoxically, then, although the difference between invalidation and nonvalidation may seem but a nuance, it matters greatly. While invalidation is by its definition more effective than punishment in stopping intermediate norm generation, nonvalidation is more benign than punishment in letting such norm generation continue. Invalidation of matrimonial promises restricts them more absolutely than does even severe punishment for making such promises, but even a mild punishment would be more inhibiting than mere nonvalidation of such promises (i.e., simply ignoring them) would be.

This conceptual distinction between invalidation and nonvalidation may stand out still more clearly if we reverse for a moment the order of morality and legality, illuminating two differing ways in which morality may pass judgment on law.

We discuss above how law may fail to validate moral norms legally without thereby aiming to invalidate them morally. As a positivist, Hart makes a similar point with regard to moral invalidation of law. Against the radical claim that “an unjust law is not a law,” that is, that an unjust, immoral law is legally invalid, Hart asserts that a properly enacted law remains such even if it is nonbinding in a moral sense. For Hart, a legal rule cannot be invalidated by morality, though it may well be nonvalid in terms of moral obligation. Hart makes this distinction in important part for the sake of pluralist thought, seeking to preserve a conceptual separation between law and morality without disparaging the role of either.Footnote ²³

Here is a succinct example illustrating the inner logic of the possibilities we are discussing in order of increasingly greater restriction on intermediate norm generation: validation, nonvalidation, punishment, and finally invalidation.

Suppose that two parents are having trouble getting their child to go to bed by 9 p.m. They might wish for the state to validate the 9 p.m. duty they have imposed on their child in order to give that duty greater authority. But if state law refuses to back the parents up officially, this nonvalidation (the mere absence of official recognition) causes them no loss. The filial duty to obey remains intact as a matter of morality, religion, custom, and the like.

By contrast, the state could declare that making a child go to bed so early is child abuse and could forbid the parents to impose a 9 p.m. bedtime, backing up this state-imposed parental duty with a threat to punish them if they continue to tell their child to go to bed by 9 p.m. Yet at the same time, state law might still say and do nothing regarding the moral, religious, and other norms that would support any 9 p.m. duty upon which the parents might defiantly insist. The filial duty to go to bed could thus remain intact, albeit nonvalidated (and even punished) as a matter of state law.

Finally, the state could declare itself supreme and directly effective in all normative realms, invalidating every sort of norm requiring going to bed at 9 p.m. and informing the child of its moral and legal right to stay up late. Clearly, this last exercise of state power could be the most destructive of family autonomy. Yet there is a condition affecting its efficacy still to be discussed.

IV. A FURTHER CONDITION FOR NONVALIDATION

The very possibility of nonvalidation (rather than invalidation) of an act depends on the act in question having some force or purpose independent of the achievement of legal validity in the normative order that is granting or denying such validity.

If we look back to Hart's nonrepressive sorts of disabilities, such as the disability to make legally enforceable contracts in the absence of reciprocal consideration, we can see that the power-holders in question have no desire to have their nonvalidated informal acts become legally validated. No one is demanding that the set of “contracts without consideration” or the set of “wills with only one witness” become valid (as opposed to arguing that particular acts do not fall within these sets).Footnote ²⁴ Nonvalidation is an option because everyone is satisfied to have the law ignore such informal norm creation.

Consider, by contrast, the following report: in 2007, the New York Times ran a story about parents who had organized to obtain birth certificates for their stillborn children. They wanted the state to certify that their children had once existed. “It's about dignity and validity. It's the same reason why we want things like marriage licenses,” declared a leader.Footnote ²⁵ Insofar as parents come to depend upon state validation for the very existence of their children and marriages, they give nonvalidation the same effect over their lives as invalidation would have. Furthermore, by insisting upon validation, they tempt the state to extend its power. It would require great self-restraint and a kind of paternalism on the part of the state for it to resist this offer of overlordship—for the state to try to encourage these parents to seek the assurance they need from nonstate sources. After all, if the state does nothing, it will be blamed for the nonexistence of key family goods, so (unless it would incur large costs in doing so) it might as well get into the business or deciding which children and marriages deserve legal recognition. In at least a de facto sense, the option of nonvalidation no longer exists, because the parents have abdicated their extralegal autonomy.

In like manner, if the bedtime-concerned parents discussed above feared to exercise any extralegal authority so that they felt they could not ask their children to do anything except that which the state would recognize and support, they would have voluntarily relinquished their normative independence. There would no longer be any family sphere that could be left merely nonvalidated by the state (and thus still normatively binding in nonlegal ways).Footnote ²⁶

In other words, Hart's disabilities are stably nonrestrictive—open to continuing interpretation as nonvalidation rather than as invalidation—only insofar as two conditions obtain. First, the superior order must impose a disability only as to its own legal order; it must not negate the informal norms (or norms operative in a separate legal sphere) supporting the practice in question, such as the practice of informal promising or raising import taxes; it must not reach down to destroy an otherwise autonomous normative order. Second, that latter order must have internal strength; it must not come to depend for its internal validity on higher-level recognition or even on the absence of higher-level invalidation. Informal promising cannot remain fully autonomous if only noninvalidated promising comes to be felt binding, nor can Dutch law remain autonomous if it refuses to enforce any laws that have been rejected in the international realm. A nation whose own constitution is entirely monist with regard to international law depends for its independence upon the pleasure of the makers and interpreters of international law.

Indeed, if a lower-level normative community has sufficient strength, it may even be able to counter a higher legal authority's attempts to invalidate its norms. Lower-level actors may continue to abide by officially invalidated norms, that is, they may treat those norms as merely nonvalidated. Thus even common-law marriages and gambling debts that are wholly invalid according to state law may in fact retain significant force for the parties engaged in them. Likewise, a powerful country (e.g., the United States) may simply ignore the findings of some international legal tribunal that certain domestic norms lack validity (e.g., certain procedures for imposing the death penalty). In the same way, a fiercely filial child could treat an officially invalid bedtime rule as merely nonvalid and insist on observing it. Indifference to official invalidation can turn it practically into nonvalidation.

A diverse or plural legal world thus does not depend only on making national and supranational orders less hegemonic; it depends also on persons and groups refraining from seeking higher recognition or enforcement of their internal norms. It could be unfortunate for humanity if each and every human norm were subject to state validation, even if most were to be granted legal validity. We might not wish every promise, every private family and friendship event or duty, to be registered and overseen even by a benevolent State. In like manner, we might not wish every national law to be subject to the approval of even the most well-intentioned international or supranational legal regime.

V. THE EXAMPLE OF SAME-SEX MARRIAGE

A useful illustration of the ambiguous benefit of state validation and yet the complexity of resistance can be seen in the topic of same-sex marriage. Departing from its draconian treatment of bigamy (and, historically, miscegenation), current American law nowhere “prohibits” extralegal same-sex marriage in the sense of punishing it; people of the same sex have full behavioral liberty to engage in wedding ceremonies, to bind themselves morally and even by legal contract regarding property and the like, and to live together thereafter.Footnote ²⁷ But, except in MassachusettsFootnote ²⁸ and California,Footnote ²⁹ American law does not validate such marriages qua marriages. By and large, the law just ignores them—leaves them alone.

The current debate over same-sex marriage may turn precisely on the distinction between invalidation and nonvalidation. Some may view the current absence of validation of same-sex marriage as a relatively tolerant system of nonvalidation, one that leaves gay and lesbian cultural norms intact and unregulated. Others may view the current legal regime as closer to relatively hostile invalidation, which declares such unions to be without practical effect.

Perhaps the status of same-sex bonds depends upon the purpose of marriage for the gay and lesbian community or upon the purpose of each individual same-sex marriage for those entering into it. If the point of same-sex marriage is to gain the tax and retirement benefits hitherto open only to heterosexuals or to obtain a supposed moral imprimatur from the law, refusal of legal recognition amounts in effect to invalidation, whatever its intent may be. In order to obtain these goods, validation would have to be sought, even at the price of regulation (e.g., application of bigamy and divorce law). By contrast, if those entering such a marriage were content to bind themselves by a mutual promise of lifelong fidelity or content to secure nonlegal forms of moral recognition by the community, legal nonrecognition would do little or nothing to undo those effects and could be more appropriately characterized as nonvalidation.

Failure to perceive the difference between invalidation and nonvalidation may lead lesbian and gay communities to think that state validation is a prerequisite to all normative stature, not realizing that nonvalidation could actually be a means to protect the moral and social autonomy of same-sex unions. By comparing only invalidation and validation, they may fail to consider nonvalidation as a possible way to achieve normative stature while avoiding state intrusion into private life.