1. Key Kantian Concepts and Terminology

I understand ‘moral laws’ to be, by definition, ‘unconditional practical laws’ (MM, 6: 221).Footnote ¹ A law is ‘practical’ insofar as ‘it makes certain actions duties’ (MM, 6: 224). A practical law is ‘unconditional’ in the sense that it obligates all rational beings regardless of their sensible inclinations (G, 4: 420). For imperfectly rational beings like us, who may have sensible inclinations to the contrary, ‘moral laws are imperatives (commands or prohibitions) and indeed categorical (unconditional) imperatives’ (MM, 6: 221). To summarize, ‘a morally practical law’ is a ‘categorical imperative, because it asserts an obligation with respect to certain actions’ (MM, 6: 222–3). Kant also sometimes refers to moral laws as ‘obligatory laws’ (verbindenden Gesetze) (MM, 6: 224).

There are two types of moral law: ethical and juridical (MM, 6: 214). ‘Ethical’ laws are those moral laws that command us to set ends for ourselves, and they can therefore only be given internally: legislated by the individual agent herself (MM, 6: 219, 239). By contrast, ‘juridical’ laws are those moral laws ‘for which there can be external lawgiving’ because they command us only with respect to our external (physical) conduct (MM, 6: 219, 224). All juridical laws are either ‘natural laws’, which ‘can be recognized as obligatory a priori by reason even without external lawgiving’, or ‘positive laws’, which ‘do not bind without external lawgiving (and so without it would not be laws)’ (MM, 6: 224).

Unfortunately, Kant is not perfectly consistent in his use of the term ‘law’ (Gesetz) and its permutations in his legal and political writing. At times, he is clearly using the term to refer to a juridical law as described above – a type of moral law. For example, Kant writes: ‘[a]ny transgression of the law (des Gesetzes) can and must be explained only as arising from the maxim of the criminal (to make such a crime his rule)’ (MM, 6: 321 footnote). Elsewhere, however, Kant has used the same terminology to refer to legislative enactments that – he seems to be claiming – do not obligate us. For example, he writes, ‘My external (rightful) freedom is, instead, to be defined as follows: it is a warrant to obey no other external laws than those to which I could have given my consent’ (TPP, 8: 349 footnote). Difficult interpretive questions sometimes arise when Kant refers to a ‘law’ in the absence of context that unambiguously indicates the presence of absence of an obligation to obey. To forestall confusion, I will avoid this ambiguous usage myself. Instead, I will use the term ‘legislative enactment’ to refer simply to the concept of a procedurally adequate enactment by the legislative body of a state that holds the legislative, executive and judicial authorities.

The terms ‘duty’, ‘obligation’ and ‘ground of obligation’ also warrant brief clarification. According to Kant, both ‘duty and obligation are concepts that express the objective practical necessity of certain actions’ (MM, 6: 224). Duty, Kant writes, ‘is that action to which someone is bound. It is therefore the matter of obligation’ (MM, 6: 222). Obligation, in turn, is ‘the necessity of a free action under a categorical imperative of reason’ (MM, 6: 222). Following Jens Timmermann, I use the term ‘duties’ to refer to ‘general prescriptive laws or rules that provide the matter of what ought to be done’, and I use the term ‘obligations’ to refer to individual ‘token’ actions that specific agents are under a duty to perform (Timmermann Reference Timmermann2013: 43). I understand juridical laws to give rise to obligations directly (CPrR, 5: 159). This is possible because our juridical duties are ‘perfect’ (i.e. ‘narrow’) duties: unconditional duties to undertake token actions (MM, 6: 240, 411). By contrast, ‘ground of obligation’ refers to an ethical law that puts us under a duty to adopt a morally good end but does not directly yield any external obligations. Instead, our ethical duties to advance our obligatory ends are ‘conditional’ in the sense that they cannot be fulfilled by any action that violates a juridical duty (Timmermann Reference Timmermann2013: 45). Our juridical duties are thus lexically prior to our ethical duties.

2. The Three Authorities

Kant writes that the state’s three authorities (Gewalten) – legislative, executive, and judicial – are ‘the three relations of the united will of the people, which is derived a priori from reason’ (MM, 6: 338). I take Kant to mean that these three authorities (i.e. rightful capacities) are the subject matter of the idea of the original contract (MM, 6: 315). These authorities make possible our external freedom – understood as ‘independence from being constrained by another’s choice’ – by eliminating the state of nature’s three defects (MM, 6: 237, 312). Because we cannot unilaterally impose obligations on others not to interfere with our projects, we need a source of universally binding laws that enable our acquisition and enjoyment of property, contract and status rights (MM, 6: 263). The capacity to generate these laws is the legislative authority. Because we are not independent of the choices of others if they are free to violate our rights, we require assurance that those subject to the law will obey it (MM, 6: 307). The capacity to rightfully coerce all subjects to obey the law is the executive authority. Because we as individuals have no objective way of deciding who is right when disputes arise about the indefinite boundaries of rights, we require an independent source of judgements in specific cases (MM, 6: 312). The capacity to rightfully issue such judgements is the judicial authority.

Kant describes the three authorities as ‘coordinated’ in that ‘each complements the others to complete the constitution of the state’ (MM, 6: 316). Simultaneously, each authority is also ‘subordinate’ to the others in that ‘one of them, in assisting the other, cannot also usurp its function; instead, each has its own principle, that is, it indeed commands in its capacity as a particular person, but still under the condition of the will of a superior’ (MM, 6: 316). I understand Kant to be referring to the omnilateral will as the ‘will of a superior’ in the text above, the contents of which are the three authorities considered as a unity. Kant is thus saying that any one authority can be exercised only on the condition that all three exist. As I will show, each authority provides logically necessary preconditions for the exercise of the others. Kant is also claiming that no authority can ‘usurp’ any other authority. I will demonstrate that the metaphysically distinct natures of the authorities preclude any such usurpation.

First, consider the ways in which each authority creates the necessary conditions for the exercise of the others. The exercise of the executive authority, for example, depends on the existence of both the legislative authority and the judicial authority. The legislative authority is the authority to create juridical laws, including the laws that determine which empirical agent holds the state’s executive authority. In the absence of such a law, the executive authority cannot be exercised by anyone. Moreover, because laws are indefinite in their application to particular individuals and events in the world, the executive cannot coerce us in accordance with them unless they are previously made definite through an act of judgement, which requires the exercise of the judicial authority. The judicial authority similarly cannot be exercised in the absence of the legislative authority: it is not possible to apply laws to specific individuals and events without any laws to apply.

The legislative authority, in turn, cannot be exercised in the absence of the executive authority, because the latter is presupposed by every act of juridical lawgiving. All lawgiving has a two-part structure: ‘first, a law, which represents an action that is to be done as objectively necessary, that is, which makes the action a duty; and second, an incentive, which connects a ground for determining choice to this action subjectively with the representation of the law’ (MM, 6: 218). In a juridical lawgiving, the incentive provided must be ‘pathological’ and thus physical in nature (MM, 6: 219). An incentive is a required part of any juridical lawgiving because our rightful honour prevents us from seeing ourselves as obligated to obey the law in the absence of independent assurance that others are similarly constrained (MM, 6: 255–6; Newhouse Reference Newhouse, Demiray and Walla2019: 114). Because only the executive authority can threaten to coerce us in accordance with law, its existence is a necessary condition for any exercise of the legislative authority.Footnote ²

This insight can make sense of an otherwise puzzling textual detail: Kant reliably names the executive authority before (rather than after) the judicial authority when he lists the three authorities (MM, 6: 313, 316, 318). This ordering can seem counterintuitive because the application of the law to the empirical world seems to logically precede its coercive enforcement. Kant’s comparison of the three authorities to the propositions in a practical syllogism provides an important clue to the puzzle (Williams Reference Williams1983: 171). The legislative, executive and judicial authorities correspond to, respectively:

Whilst the legislative authority is the capacity to generate a law, it is the executive authority that ‘contains the command to behave in accordance’ with any such law. A juridical lawgiving thus relies upon the existence of executive authority to supply the normativity of a law. In the absence of an external incentive, a legislative enactment is ‘a merely theoretical cognition of a possible determination of choice, that is, of practical rules’ (MM, 6: 218).

There is an important sense in which the legislative authority is the first among equals in this tripartite group: the legislative authority includes the capacity to make constitutional laws that allocate the authorities among institutions within the empirical state, and to appoint delegates to occupy state offices (MM, 6: 317). An empirical agent who exercises the legislative authority of a state is thus the ‘sovereign’ (MM, 6: 317). By contrast, the highest executive officer in a state (the ‘ruler’) is subject to the laws promulgated by the sovereign and is charged with administering them as ‘the agent of the state’ (MM, 6: 316). The sovereign can replace an executive officer in his role and can even reform the constitution to alter the institution that exercises the executive authority (MM, 6: 317). However, the sovereign lacks the capacity to alienate the legislative authority itself by means of legislation (MM, 6: 321–2).

No authority can ‘usurp’ the function of any other because they function in different metaphysical domains. As I will explain in section 4, the legislative authority is a noumenal capacity: the capacity to give laws. The executive authority, by contrast, is a physical and therefore phenomenal capacity. The judicial authority functions in the distinctive space between the noumenal and phenomenal worlds that is bridged by means of the exercise of judgement. A noumenal capacity cannot in principle be exerted in the phenomenal domain or vice versa, and neither the legislative nor the executive authority can be exercised in the interstices of their domains, where judgement holds dominion. For example, Kant writes: ‘a people’s sovereign (legislator) cannot also be its ruler, since the ruler is subject to the law and so put under an obligation through the law by another, namely the sovereign’ (MM, 6: 317). Kant is here referring to the roles of sovereign and ruler. A single natural person can in principle occupy both roles (MM, 6: 339). However, the roles themselves must be undertaken independently, at least in the minds of those who attempt to exercise the authorities. A person who attempted make and enforce law as a single, undifferentiated act would not be successfully exercising either authority. This is why autocracy, which concentrates the three authorities in a single natural person, ‘is conducive to despotism’ (i.e. lawless coercion by the state) (MM, 6: 339).

Because each authority occupies its own metaphysical domain, each must have the final word within its sphere in order to make distributive justice possible: ‘the will of the legislator (legislatoris) with regard to what is externally mine or yours is irreproachable … the executive power of the supreme ruler (summi rectoris) is irresistible; [and] the verdict of the highest judge (supremi iudicis) is irreversible (cannot be appealed)’ (MM, 6: 316). As subjects, we are obligated to respect the finality of each authority, because the three together constitute the ‘supreme means for maintaining the right of human beings’ (RPR, 19: 529; MM, 6: 320).

3. How States Emerge

In order to know what our legal obligations are, we must know which (if any) individuals or institutions hold the three authorities. Because it is impossible to make legally binding contracts in a state of nature, this will depend on whether an empirical agent with the capacity to exercise all three authorities has achieved a critical threshold of de facto coercive power in the relevant territory:

Any empirical agent that gains the de facto power to exercise all three of the authorities within a territory will thus ‘find himself in possession of supreme commanding and legislative authority over a people’ (MM, 6: 372; RPR, 19: 593). Note that the deed of ‘seizing supreme power’ does not merely attract the endorsement of a pre-existing omnilateral will (RPR, 19: 503). Rather, a deed of this type constitutes the omnilateral will: its juridical effect is to unite the wills of individuals present within the controlled territory. The sovereign who represents the omnilateral will must therefore exist prior to and as a precondition of the omnilateral will itself.Footnote ³

To illustrate, suppose that an order of wealthy but housebound nuns in a hilltop abbey decided to legislate. They established parliamentary procedures and began voting on bills. Meanwhile, a gang of foreign pirates who could not speak the local language terrorized surrounding villages, seizing valuables and killing objectors. Neither group has any state authority on its own, because they each lack the ability to exercise all three of the authorities. The nuns cannot travel into the villages to adjudicate disputes or enforce their laws. The pirates cannot communicate legal rules or understand the testimony necessary to make legal judgements. Suppose, however, that the nuns hired the pirates to execute their laws (the nuns speak both relevant languages) and appointed local schoolteachers as judges. The nuns would thus become a sovereign: their deed would unite the several wills of the local inhabitants, and they would thereby gain the authority necessary to maintain a civil condition.

Kant’s remarks about the relationship between the omnilateral will and the people and institutions that constitute the empirical state are brief but illuminating. He describes three basic forms of sovereignty: autocratic, aristocratic and democratic (MM, 6: 338; TP, 8: 352). In an autocratic state, a single natural person holds all of the state’s authority. In an aristocratic state (like the one the nuns might have formed) the ‘nobility’ must be united into a group agent, which then gains sovereignty. In a democracy the united will of ‘citizens’ (i.e. those among the people who have the right to vote) forms a commonwealth, which holds the state’s authority (MM, 6: 339).

In each case, sovereignty is initially held by the natural person or group agent who came to power – the nuns, in the example above. Such sovereigns may empower delegates, such as the pirates and schoolteachers, to carry out state business on their behalf, but so long as they are sovereign, they retain the authority to ‘nullify all institutions by their absolute choice’ (MM, 6: 341). For this reason, a civil condition cannot be fully rightful as long as a specific individual or group directly holds sovereignty (MM, 6: 341). Such a sovereign therefore has a duty to establish a form of government – a self-sustaining system of laws governing the exercise of state authority by officials – that promotes mutual independence and proper coordination in the exercise of the three authorities (MM, 6: 340).

Such legal arrangements should approximate a republic: a form of government characterized by the separation of the institutions within it that exercise the legislative, executive and judicial authorities (MM, 6: 341; TP, 8: 352). To illustrate, the sovereign nuns in the example above might pass constitutional legislation establishing internal governance procedures for the executive and judicial branches of the new government, a set of laws governing the appointment of officials to those positions, a legal mechanism by which the nuns could replenish their own membership when vacancies arose, a set of checks and balances to preserve the independence of the branches of government and a process by which these constitutional arrangements can be amended.

Kant characterizes ‘any true republic’ as a representative system of the people: ‘ein repräsentatives System des Volks’ (MM, 6: 341). Kant then ambiguously claims that ‘as soon as the person who is head of state [the sovereign] also lets itself be represented, then the united people (das vereinigte Volk) does not merely represent the sovereign: it is the sovereign itself’ (MM, 6: 341). In most Kantian contexts, ‘the united people’ would refer to the omnilateral will simpliciter. However, on that interpretation this sentence would incoherently suggest that it is possible for the omnilateral will to represent a sovereign as such. On the other hand, if ‘the united people’ is a reference to the previously mentioned representative system of the people, then Kant is making a play on words: the representative system (i.e. the constitutional order) does not merely represent the [former] sovereign; it is the [current] sovereign itself.Footnote ⁴ In other words, the self-sustaining constitutional order itself becomes the sovereign of the state and thus represents the omnilateral will (into which the particular will of the former sovereign has been incorporated). I take this to be the way in which ‘law itself rules and depends on no particular person’ (MM, 6: 341).

Kant’s references to ‘a true republic’ as ‘the final end of all public right’ might suggest that the transfer of sovereignty described is merely an aspirational ideal. However, Kant undermines that inference in the next sentence by offering a real-world example of such a transfer: Louis XIV’s abdication of authority to the French National Assembly, upon which ‘the monarch’s sovereignty wholly disappeared (it was not merely suspended) and passed to the people, to whose legislative will the belongings of every subject became subjected’ (MM, 6: 341). In the analysis that follows, I will suppose that at least some of our familiar current constitutional democracies have likewise replaced flesh-and-blood sovereigns with self-sustaining constitutional orders that directly represent the omnilateral will.

4. Juridical Lawgiving

In section 2, I claimed that the legislative authority is (unlike the other authorities) a noumenal capacity: the capacity to give us laws, which are noumenal objects. This claim has textual support in many passages in which Kant characterizes juridical lawgiving as an act of the omnilateral will: ‘a public law that determines for everyone what is to be rightfully permitted for forbidden is the act of a public will’ (TP, 8: 294). In a nearby footnote, he mentions in passing that ‘the sovereign, which gives laws is, as it were invisible; it is the personified law itself, not its agent’ (TP, 8: 294 footnote). I understand Kant to be referring to the omnilateral will as the invisible lawgiving sovereign, in this passage, and to a legislative body as the sovereign’s ‘agent’. In his Reflections, Kant similarly characterizes the juridical law as ‘ex voluntante communi {proceeding from the common will}’ (RPR, 19: 500). In the Metaphysics of Morals, Kant suggests that only the omnilateral will can give juridical laws:

He similarly writes in Theory and Practice that ‘no particular will can be legislative for the commonwealth’ (TP, 8: 295).

This account of juridical lawgiving as an act of the omnilateral will can make sense of another, otherwise puzzling passage:

With respect to positive laws, I understand Kant to be referring to the omnilateral will as the ‘lawgiver’ in this passage, while the ‘author of the law’ in such cases is a legislative body. Legislative bodies of existing states thus generate the contents (i.e. matter) of positive laws, but only the omnilateral will ‘gives’ those laws to us by making them obligatory.

This is another example of the deep symmetry between internal and external law in Kant’s normative system. Kantian autonomy implies self-legislation of the moral law (G, 4: 431–3). In the ethical domain, our individual lawgiving will performs this act, whilst in the juridical domain, it is the omnilateral will that does so. Kant cannot, of course, mean that the omnilateral will – a noumenal entity – is a physical cause of state action in the phenomenal world, any more than our individual wills can be physical causes of the actions of our own bodies. Instead, it makes sense to say that the omnilateral will has given us a positive law just in case a legislative enactment can be imputed Footnote ⁵ to it.

In section 3, I offered an account of the Kantian state according to which it is comprised of three layers of agency connected by relations of supervenience.Footnote ⁶ As we have seen above, the constitution itself is the sovereign in a mature constitutional order. That constitution supervenes on state officials, all of whom are merely ‘delegates’ (Abgeordneten) of the state and carry out constitutionally assigned functions (MM, 6: 341). The omnilateral will, in turn, supervenes on the constitution. Our existing states have this vertical structure. Within the middle layer, our modern constitutions often establish complex sets of nested institutions that exercise the state’s authorities in different contexts, as well as complex systems of checks and balances between them. These constitutions may be conducive to despotism if they confer multiple authorities on single institutional agents, but that wrinkle does not bear on the account of juridical lawgiving that I am developing here. Therefore, Figure 1 illustrates the state’s metaphysical structure by depicting an unrealistically simple empirical constitution for the fictional state of Republica. It establishes only three state institutions – the Legislature, the Ruler and the Court – each of which is the highest authority within its sphere, and which are coordinate and mutually subordinate as described in section 2.

Figure 1. The Metaphysical Structure of Republica

Supervenience is a unidirectional dependence relation: A supervenes on B just in case there can be no change in A without a change in B. To see how these relationships operate, suppose that every member of Republica’s Legislature converged on the headquarters of the Republica Tribune (a daily newspaper) and burnt it to the ground. If they coordinated their actions sufficiently, we might be able to say that some sort of empirical group agent – a criminal gang that happens to be composed of all lawmakers – burned down the building. But we could not with technical accuracy say that the Legislature had done so, because the Legislature is constituted by its legal procedures, and informally agreeing to burn down a newspaper office (let us assume) does not amount to a use of those procedures. On the other hand, suppose that the Legislature followed its procedures to enact the Real News Act:

It would make sense to say that the Legislature acted in this case. Whether or not the Real News Act can be imputed to Republica’s omnilateral will as a juridical lawgiving is – in principle – a further question. The omnilateral will that supervenes on Republica’s constitutional order is its sole source of legislative authority (MM, 6: 313; RPR, 19: 478). The nature of the supervenience relation does not entail that every Act passed by the Legislature can be imputed to the omnilateral will, any more than it entails that everything that lawmakers do can be imputed to the Legislature.

This claim is not incompatible with the legitimacy of the state of Republica or its institutions. The Legislature holds Republica’s legislative authority at all relevant times: it alone has the capacity to author the contents of positive laws. Korsgaard draws a helpful distinction between an agent’s ‘capacity for action’ and her ‘success in action’ that can clarify this point (Korsgaard Reference Korsgaard, Vargas and Yaffe2014: 191). To hold the three authorities (i.e. for the supervenience relation to obtain) an empirical state must have the ability to exercise them. In a constitutional state, the constitution may allocate the different authorities to different institutions (e.g. Republica’s Legislature or Court). Because their exercise must be ‘coordinate’, these institutions must themselves be united by constitutional laws that regulate their relations. The constitution must also supervene on adequate human and physical resources. It is only under these conditions that the omnilateral will comes into existence and therefore has the ‘capacity to act’.

Whether the omnilateral will has achieved ‘success in action’ through the Real News Act will depend on the constitutive features of the legislative authority itself. In Kant’s words, ‘the social contract is the condition of legislative power’ (RPR, 19: 478). Korsgaard argues that human beings have the capacity to act in the phenomenal world if they supervene on human animals, but they achieve ‘success in action’ only insofar as the human animals on which they supervene adopt maxims that are consistent with the Categorical Imperative (Korsgaard Reference Korsgaard, Vargas and Yaffe2014: 192). When a human animal acts on a defective maxim, the person that supervenes on her has failed to act. Though I have reason to believe Korsgaard thinks otherwise, I will try to show in section 5 that an analogous dynamic operates in the context of state action (cf. Korsgaard Reference Korsgaard2008: 246–7). Republica’s Legislature can act so long as it follows its procedures, but a legislative enactment cannot be imputed to the omnilateral will as the contents of a juridical lawgiving unless that enactment is within the limits of the legislative authority.

5. The Limits of the Legislative Authority

The legislative authority is the capacity that the state must in principle have in order to eliminate a specific problem in the state of nature: individuals’ inability to unilaterally create new external obligations for others without infringing on their external freedom (MM, 6: 263). It is thus the capacity to give juridical laws, which by definition establish universal external obligations omnilaterally (MM, 6: 218–19). The concepts of all three authorities are derived a priori: ‘they arise necessarily from the idea of the state as such’ (MM, 6: 315). The nature and limits of the legislative authority will therefore be identical regardless of the features of the empirical sovereign on which it supervenes (Walla Reference Walla, Krasnoff, Madrid and Satne2018: 41). A despotic sovereign enjoys no more and no less legislative authority than an ideal republic.

I argued in section 4 that a legislative enactment is only a juridical law if its contents can be imputed to the omnilateral will. Kant explains why this must be the case:

This claim – that only the omnilateral will can give us juridical laws because it cannot wrong us – makes sense when you consider that it is conceptually impossible for a juridical law to wrong us. Recall that a juridical law is a type of moral law: an unconditional rational requirement (MM, 6: 214). It is therefore ‘inconceivable’ (conceptually impossible) for any juridical law to conflict with our duty of rightful honour, which obligates us not to surrender our external freedom (MM, 6: 224, 236).

The legislative authority thus does not include the capacity to create juridical laws that are conceptually incompatible with our external freedom, and thus with its logical possibility. As Kant writes, ‘What a people cannot decree for itself, a legislator also cannot decree for a people’ and ‘What a people (the entire mass of subjects) cannot decide with regard to itself and its fellows, the sovereign can also not decide with regard to it’ (TP, 8: 304; MM, 6: 329). This is a different standard than that of whether the people would agree to enact some positive law (i.e. ‘hypothetical’ consent). Kant is clear that we are bound by positive laws to which we would refuse our consent if consulted (TP, 8: 297).

Our hypothetical consent cannot be a condition for lawgiving because the legislative authority cannot secure our freedom in the absence of empirical judgements about the arrangements that will ensure the well-being of the state. On matters of judgement, it is impossible in principle for a single conclusion to be solely objectively correct, which is why the sovereign, acting through its delegates, must have the final word (TP, 8: 299–300). Thus, ‘A head of state must be authorized to judge for himself and alone whether such laws [promoting prosperity or population increase] pertain to the commonwealth’s flourishing which is required to secure its strength and stability’ (TP, 8: 298).

Legislatures therefore have many opportunities to make laws that diminish our happiness – and even endanger our freedom – by imprudently frittering away blood and treasure (TP, 8: 299). For example, lawmakers may establish a knightly order by statute if they believe it will enhance the security of the state, and that statute can be repealed if they decide that it has outlived its usefulness, regardless of the soundness of either judgement (MM, 6: 324). They can establish international trade barriers if they think that doing so will promote prosperity (TP, 8: 299 footnote). Legislation can authorize a war if it is judged – rightly or wrongly – to be necessary to the preservation of the state, and it can draft soldiers to serve in that war (MM, 6: 346; FL, 27: 1386). Kant elaborates:

Indeed, we must regard the state’s judgements as authoritative even when lawmakers are biased by personal motives, such as ‘a lively interest in positions for themselves and their families, in the army, the navy, and the civil service, that depend on the minister’ (MM, 6: 319-20). In such cases, lawmakers’ actions are morally bad (i.e. they exhibit a lack of virtue), but they do not wrong us.Footnote ⁷

When it comes to questions of conceptual compatibility, however, the distinctive faculty of judgement (exercised when principles are applied to particulars) is not called upon – there cannot be two different but equally formally correct answers to a question about the presence or absence of such a conflict. A legislative enactment that is inconsistent with the concept of our external freedom – and thus with its mere logical possibility – therefore cannot be the subject matter of a juridical lawgiving. This is the sense in which ‘the statutory laws obtaining in [a civil] condition cannot infringe upon natural right, (i.e. that right which can be derived from a priori principles for a civil constitution)’ (MM, 6: 256). On my proposed account, Kant is saying that such laws are impossible, not merely that legislative bodies should not enact them.

I will refer to such enactments as Logically Incoherent Enactments (LIEs).Footnote ⁸ Kant provides many examples of LIEs throughout his normative philosophical writings, several of which concern a state’s constitutional arrangements. For example, a provision that authorized the people to overthrow the empirical legislature of a state is inconsistent with the possibility of legal finality, and thus with a civil condition (MM, 6: 320; RPR, 19: 575). For the same reason, a provision that purports to authorize anyone to coerce the person or institution who holds the state’s highest executive authority would be a LIE (MM, 6: 319). The empirical sovereign of a state can reallocate the state’s executive authority to a different person or institution by law, but an enactment that purported to alienate the legislative authority itself to another person or institution would be a LIE (MM, 6: 321–2, 340).

Other LIEs purport to limit the rights of individual subjects in ways that are inconsistent with the logical possibility of their external freedom. For example, the state has no authority to prohibit subjects from acquiring property under the lex permissiva (MM, 6: 256). Also, a statute prohibiting emigration would be a LIE, since ‘the state could not hold [a subject] back as its property’ (MM, 6: 337). Most famously, Kant insisted that the legislative authority does not include the capacity to regulate religious worship: ‘Religion (in appearance) … can neither be imposed upon a people nor taken away from them by any civil authority’ (MM, 6: 368, 327; RPR, 19: 479, 519, 579; FL, 27: 1386). A law requiring witnesses to affirm the existence of God before testifying in legal proceedings is also a LIE (MM, 6: 304–5).Footnote ⁹

A further set of LIEs violate the principle of equality inherent in the idea of external freedom under laws – the principle that one cannot be bound by obligations not applicable to others – by making baseless distinctions when allocating legal benefits or burdens (MM, 6: 237–8). For example, creating a hereditary nobility is beyond the capacity of the legislative authority even if it might advance a state’s prudential interests (MM, 6: 329; TP, 8: 297). A statute that prevented passive citizens (i.e. those who lack the right to vote) from ‘working their way up’ to the status of active citizens would also be a LIE (MM, 6: 315). In both cases, subjects could not possibly will their permanent subordination to rulers they have no prospect of joining. Legislation that removed individuals from public office arbitrarily, rather than for cause, would also be a LIE, as would legislation that denied public offices to practitioners of a disfavoured religion (MM, 6: 328, 368).

Similarly, an enactment may be a LIE if it speciously imposes unequal tax burdens. For example, a war tax imposed only on a disfavoured group is a LIE, since ‘a whole people could not agree to a law of this kind’ (TP, 8: 297 footnote). Similarly, ‘A law can thus not possibly be just if a despot places taxes on merchants but exempts his favorites from them. It is not necessary that he judges whether the people in this case would have made such a law but whether they could have made such a law’ (FL, 27: 1382). Such legislation could not possibly be willed by disfavoured subjects, and therefore it cannot be the subject of a juridical lawgiving by the omnilateral will, which by definition includes the wills of ‘the entire mass of subjects’ (MM, 6: 329; TP, 8: 297).

I have argued that a LIE – a legislative enactment that is conceptually incompatible with the idea of our external freedom – is outside the limits of the legislative authority and therefore cannot supply the contents of a juridical lawgiving. In practice, this amounts to a legal disability (i.e. an incapacity) for legislative bodies that is analogous to certain legal disabilities that individuals have in private law. For example, a contract purporting to sell the signer into slavery is a legal nullity because it is inconsistent with the signer’s rightful honour (MM, 6: 330; RPR, 19: 547). In Korsgaard’s parlance, the human animal has acted in such a case: she has signed her name on a piece of paper that purports to throw away her freedom (Korsgaard Reference Korsgaard, Vargas and Yaffe2014: 192).Footnote ¹⁰ However, the human being has not achieved success in action, which is why such contracts have no juridical effect (TP, 8: 304). Analogously, when a legislative body enacts a LIE, the omnilateral will has not achieved success in action. Kant analogizes between individual and state action explicitly when discussing a LIE establishing hereditary nobility: ‘Since we cannot admit that any human being would throw away his freedom, it is impossible for the general will of the people to assent to such a groundless prerogative, and therefore for the sovereign to validate it’ (MM, 6: 329).

The legislative authority nonetheless remains supreme in its sphere: ‘the will of the legislator (legislatoris) with regard to what is mine and yours is irreproachable’ (MM, 6: 316). Not all legislative enactments can supply the contents of juridical lawgivings, but no competing empirical agent can ever do so. As Kant writes, ‘the authority which already exists, under which you live, is already in possession of legislative authority, and though you can indeed reason publicly about its legislation, you cannot set yourself up as an opposing legislator’ (MM, 6: 372).

6. Legal Finality

I have argued that LIEs cannot serve as the subject matter of juridical lawgivings. It follows that individuals have no juridical (and thus perfect) duty to obey the terms of LIEs. This conclusion may give rise to grave concerns about the structural integrity of Kant’s system of right: can the system provide legal finality in the absence of such a duty? Kant refers to legal finality as ‘distributive justice’: an authoritative judgement about rights in a particular case (MM, 6: 306). The possibility of such final judgements is the sine qua non of a civil condition (MM, 6: 306). I hope to show that my account of the nature and limits of the legislative authority poses no threat to the possibility of legal finality, because legal finality is not achieved by means of the legislative authority alone. Indeed, it is the ultimately the executive authority that makes legal finality possible.

The executive authority ‘is the supreme capacity to exercise coercion in conformity with the law’ (MM, 6:317). Unlike the sovereign, the individual or group agent holding the state’s executive authority (whom I will refer to as the ‘ruler’) can wrong the people by ‘proceed[ing] contrary to law’ (MM, 6: 319; RPR, 19: 507). Kant’s examples of such abuse include discriminatory taxation and recruitment practices and a decision not to punish a convicted criminal (MM, 6: 319, 337). More saliently, the ruler may ‘proceed quite violently (tyrannically)’ (TP, 8: 299–300). When the ruler acts lawlessly, ‘subjects may indeed oppose this injustice by complaints (gravamina) but not by resistance’ (MM, 6: 319). Kant specifies that the resistance forbidden to us by the idea of the original contract is ‘active resistance’, understood as coercion of the government (MM, 6: 322; TP, 8: 300). The very nature of the executive authority as the supreme capacity to exercise lawful coercion entails that no one can have the lawful capacity to coerce the ruler.

The executive authority must have this feature, because the administration of the law necessarily requires the exercise of judgement: the application of the law to events and agents in the world. Questions that require the exercise of judgement can by definition have multiple different but equally correct answers. For this reason, legal finality requires that ‘the people’s judgment to determine how the constitution should be administered is no longer valid’, nor can any other state institution usurp the ruler’s administrative judgements (TP, 8: 300).Footnote ¹¹ If it were otherwise, then a violent conflict between two rival executives could arise to which there could not in principle be a definitive legal resolution (TP, 8: 291, 294). Even ‘unbearable’ abuses of the executive authority thus have only one possible remedy: the legislature’s removal of the ruler from office in favour of a replacement (MM, 6: 320, 317).

I will use the example of the Real News Act to illustrate the compatibility of my account with the requirement of legal finality. When the journalists of the Republica Tribune learn that the Act has passed, they must initially decide whether or not to comply with its terms, which would require them to stop publishing the Tribune, and indeed to stop publishing their views anywhere. Suppose that a handful of journalists quit because they believe that they are obligated to obey the Act. The remaining group concludes that the Act is conceptually incompatible with the concept of their external freedom, because the Act’s terms are inconsistent with their innate right to communicate their thoughts to others, making it formally defective even if it were universally binding (MM, 6: 238). Moreover, they conclude that the Act violates the principle of equality by depriving only the owners and employees of the Tribune of their right to communicate freely (MM, 6: 237). Some of these journalists quit anyway in order to avoid the possibility of imprisonment, but others resolve to continue publishing the Tribune on the grounds that the Real News Act cannot be the subject matter of a juridical law.

The Ruler of Republica must then decide whether or not to arrest the remaining Tribune journalists, and let us suppose that she does so. Since we have stipulated that the Act is a LIE, the Ruler has wronged the journalists (MM, 6: 319; RPR, 19: 507). The journalists are nonetheless absolutely forbidden from resisting arrest by the idea of the original contract itself, which invests the current Ruler with Republica’s supreme coercive authority (MM, 6: 317, 319). The Court would then decide whether or not to convict the journalists (MM, 6: 317; RPR, 19: 507). Since there is no dispute about the facts, the Court need only determine whether the journalists have acted unlawfully. To determine the content of Republica’s laws, the Court will need to decide whether the Act is a law, and the proper method for answering this question is – as I have argued – logical reasoning. But after the Court has made that determination, it will have to apply the law of Republica (as a unity) to the conduct of the journalists order to convict or acquit them. In other words, the actual disposition of the case will require the exercise of judgement, and the constitutional imperative of legal finality makes the verdict of the highest judge irreversible. If the journalists are convicted by the Court, let us suppose that they are imprisoned by the Ruler for the duration of their sentence. Because the Ruler is the supreme coercive authority in Republica, the journalists are, in Kant’s words, permitted to ‘oppose this injustice by complaints (gravamina) but not by resistance’ (MM, 6: 319). This is the sense in which, as Kant writes, the ‘freedom of the pen … is the sole palladium of the people’s rights’ (TP, 8: 304). The people’s rights do not require a palladium if the terms of a LIE are never enforced.