The moral benefits of coercion: A defense of ideal statism

Abstract

This paper contributes to recent discussions on ideal anarchism vs. ideal statism. I argue, contra ideal anarchists, that coercive state institutions would be justified even in a society populated by morally perfect individuals. My defense of ideal statism is novel in that it highlights the moral benefits of state coercion. Rather than the practical effects on individual compliance or the distributive outcomes that follow therefrom, coercive state institutions are justified through the moral benefits they provide. The state is morally beneficial because it a) lessens the demands on the will that fall on agents under ideal anarchism, and b) counters the structural domination that follows from differences in natural endowments. By shifting the focus of the debate from feasibility to desirability, the paper exposes the flaws of ideal anarchism and provides new insights into the moral value of the state.

Would the ideally just society need, or allow, the governing institutions characteristic of the state? This question, which has become the topic of a recent debate, divides political philosophers into two camps. Ideal anarchists argue that the assumption of strict compliance made in ideal theory leaves no room for the state. In their view, state institutions respond to noncompliance with moral demands and can only be justified in its presence (Brennan and Freiman, 2022; Freiman, 2017; Levy, 2016). Ideal statists argue, to the contrary, that the state has a role to fulfill even among willing compliers. In particular, we need the state to solve coordination problems and settle disputes in cases where morality offers no clear guidance (Carroll, 2021; Estlund, 2016; Kavka, 1995; Rawls, 1999).

The current debate finds predecessors in anarchist political thought, in John Locke's social contract theory, and in the Marxist idea of the withering away of the state. Important to note, however, is that the disagreement between ideal anarchists and ideal statists primarily concerns the implications of making certain assumptions in political theory. Ideal anarchists need thus not be committed neither to ideal theory as a theoretical approach, nor to anarchism as a political order. Their claim is simply that if we tailor our theory to ideal conditions – including that of perfect moral motivations – we cannot justify the governing institutions of the state. Whether this shows that anarchism is morally superior to statism or that ideal theory is fundamentally mistaken remains an open question. Why, then, should we care about this debate? First, because we are interested in normative principles as such and not only their application. The fact that it may be difficult to explain what the value of principles that lack practical application consists of constitutes no reason to think that there is no such value. After all, non-practicable knowledge is a common feature in disciplines like mathematics and astronomy. If we hold that this kind of research is valuable, we should adopt the same attitude vis-à-vis political theory (Estlund, 2020: 310). Second, because principled discussions of this kind have effects on how we perceive of the social arrangements we currently live under. When we try to assess the justice of existing states, it makes a difference whether the state constitutes a necessary response to nonideal conditions or an integral part of the ideally just society. Recalling John Rawls’ view that political theory aims, at least in part, at reconciliation, we may consider the present debate as an attempt to settle just how far our reconciliation with the state should go (Rawls, 2001: 38; Stemplowska, 2008: 337).

This paper has two aims. The first is to rebut previous defenses of ideal statism that, I will argue, all fail to acknowledge what moral perfection entails and hence mistakenly incorporate nonideal concerns into their ideal theories. The second aim is to develop a defense of ideal statism that avoids those pitfalls and rebuts ideal anarchism. In contrast to previous defenses, my argument does not appeal to the effects of state institutions on individual compliance, or the distributive outcomes that follow therefrom. I agree with ideal anarchists that among morally perfect individuals, no state will be needed to achieve and maintain a just social scheme. Instead, I argue that the ideal state is superior to the anarchic ideal society because it a) lessens demands on the will of complying agents, and b) counters structural subordination that follows from differences in natural endowments. By highlighting not the practical, but the moral benefits of state institutions, my argument exposes the flaws of ideal anarchism and provides new insights into the moral value of the state.

Before getting to work, a few clarifications are needed. First, this paper is concerned with the justifiability of coercive state institutions. For the sake of argument, I will grant to ideal anarchists that the presence of formal mechanisms for coordination and dispute settlement does not indicate the presence of a state, as long as those mechanisms are not backed up by coercion. In line with previous contributions, I will understand the state as a subset of society which claims monopoly on the legitimate use of coercion and possesses sufficient power to uphold this monopoly (Brennan and Freiman, 2022: 5; Kavka, 1995: 2). To this minimal definition, I will add the assumption that the power of the state is constitutionally constrained. The rule of the state differs from that of a despot or a mafia clan in that it is a rule of law (Hardin, 1999: 8). I do not claim this is the best, or “real”, understanding of what a state is. Neither do I suggest that the state as a historical construct is morally superior to every other imaginable social order. What I do argue is that a social order that includes such a subset is superior to an anarchic one. For simplicity, I will call this social order a state.

Second, I will assume that the members of the ideal society are morally perfect. Following Gregory Kavka, I will understand a morally perfect agent as one that a) has a system of moral beliefs that contains all true (or objectively justified) moral beliefs, and no false (or objectively falsified) moral beliefs, b) always acts conscientiously on his moral beliefs, and c) is like ordinary people in all other respects (Kavka, 1995: 2–3). This goes beyond the understanding of ideal theory advanced by Rawls, which assumes strict compliance only with the principles of justice that, on his account, cater to the “facts of moral psychology” (Rawls, 1999: 125–6). By assuming moral perfection, I do not wish to deny that justice is the first virtue of social institutions, or that some parts of morality are fully private and lack political implications. I include this assumption in order to grant as much as possible to ideal anarchists, who are the primary targets of my argument. I also willingly concede that insofar as the state is justified due to moral imperfections of some sort, there is at least one sense in which our theory is less than fully ideal (see Levy, 2016: 319).

The paper is structured as follows. The first three sections present and rebut previous defenses of ideal statism. I argue that the appeals to assurance, reciprocity, and coordination and dispute settlement all fail to establish the need for coercive institutions among morally perfect agents. In the following two sections I develop my own defense of ideal statism based on demands on the will and structural subordination. After that, I respond to some possible objections to my argument. The last section summarizes my conclusions.

Assurance

The most prominent defender of ideal statism is Rawls, who, in his groundbreaking work A Theory of Justice , sets out to identify the principles that would guide the fully just society. Following a contractualist tradition, Rawls outlines an idealized choice situation – the Original Position – and argues that the principles that would be chosen in this situation constitute principles of justice.

Rawls’ account proceeds on the assumption of strict compliance, that is, he asks us to assume that whatever principles are agreed upon in the Original Position will be fully complied with. The members of what Rawls calls the “well-ordered society” will be guided by a sense of justice, and this fact will be public knowledge among them (Rawls, 1999: 125). Even so, he argues, the fully just society will need the coercive institutions of the state. Though the choosers have agreed to comply with the principles of justice, the agreement remains to be “tied down” through the establishment of effective enforcement mechanisms. Only when such mechanisms are in place can members of the fully just society be sufficiently assured that their fellow citizens will comply. Rawls explains:

The sense of justice leads us to promote just schemes and to do our share in them when we believe that others, or sufficiently many of them, will do theirs. But in normal circumstances a reasonable assurance in this regard can only be given if there is a binding rule effectively enforced (Rawls, 1999: 236).

Throughout A Theory of Justice, Rawls repeatedly returns to this idea. He makes clear that he holds enforcement mechanisms to be necessary even in ideal circumstances, despite the fact that they will not, in such circumstances, be put to practical use. The reason why we will still need them is that we cannot expect members of the ideal society to rely on the good will of others. If no assurance is provided, the social contract is upheld only by the bond of mutual trust. This, Rawls seems to hold, cannot provide the firm ground necessary for maintaining justice. The following passages clarify the details of his view:

Some collective arrangement is necessary and everyone wants assurance that it will be adhered to if he is willingly to do his part. In a large community the degree of mutual confidence in one another's integrity that renders enforcement superfluous is not to be expected. In a well-ordered society the required sanctions are no doubt mild and they may never be applied. Still, the existence of such devices is a normal condition of human life even in this case (Rawls, 1999: 237).

But since even with a sense of justice men's compliance with a cooperative venture is predicated on the belief that others will do their part, citizens may be tempted to avoid making a contribution when they believe, or with reason suspect, that others are not making theirs (Rawls, 1999: 296).

While mutual trust may be sufficient for upholding agreements in smaller settings, Rawls argues that members of large-scale societies will lack the necessary confidence in each other. In want of effective enforcement mechanisms, they will be prone to suspect that others will not contribute and, as a consequence, withhold their own contribution. We may reconstruct Rawls’ argument as follows:

P1 The ideal society is maintained only if all members contribute their fair share.

P2 Members of the ideal society will contribute their fair share only if reasonably assured that others will contribute theirs.

P3 Members of the ideal society can be reasonably assured that others will contribute their fair share only if effective enforcement mechanisms are in place.

C Therefore, maintaining the ideal society will require effective enforcement mechanisms.

So stated, the argument is valid. We must, however, ask with what right the members of the ideal society premise their compliance on that of others. According to premise 2, the members of the ideal society will contribute their fair share only if others do the same. Yet, that others commit wrongful acts does not generally give me a right to do the same (Estlund, 2014: 221). Rawls describes the predicament of the members of the just society in terms like “willingly do his part” and “tempted to avoid making a contribution.” This seems to indicate that what will fail absent assurance is the motivation to comply, in which case his argument appeals to the members’ weakness of will. But weakness of will is exactly what is ruled out by the assumption that morally perfect agents “always act conscientiously on their moral beliefs.” The need for assurance thus seems to import into Rawls’ ideal theory a distinctively nonideal element (Carroll, 2021: 1134). This lends support to ideal anarchists, who argue that ideal statism suffers from incoherency (Freiman, 2017: 302; Levy, 2016: 314).

But even if Rawls’ argument for the state relies on lacking moral motivations and hence removes us from the domain of ideal theory, there might be better ways to defend the need for assurance. One promising argument appeals to the reciprocal nature of justice. In brief, this states that only when assurance is provided is justice truly reciprocal, and only if justice is reciprocal are the demands it places on us fair. The coming section examines this proposal in further detail.

Reciprocity

I have argued that members of the ideal society will not need assurance that others contribute their fair share in order to contribute theirs. To qualify one's compliance on that of others may be common practice in the real world, but this is due to the imperfect moral motivations of real-world agents. But perhaps the problem is not that agents will not comply unless assured that others do the same, but that they have no duty to do so? Insofar as there are moral reasons that fall short of duties, we may doubt that the concept of duty fully captures the motivational span of morally perfect agents. If we assume that such agents will not only comply with all their duties, but also perform all supererogatory acts that they are capable of, the conditions for duties to arise may be of small importance. For the sake of argument, however, I will grant that the moral reasons relevant for maintaining the ideal society translate into duties, and hence that “to act conscientiously on one's moral beliefs” is equivalent to complying with one's moral duties. To formalize this proposal, we may modify Rawls’ argument in the following way:

P1 The ideal society is maintained only if all members contribute their fair share.

P2* Members of the ideal society have a duty to contribute their fair share only if reasonably assured that others will contribute theirs.

P3 Members of the ideal society can be reasonably assured that others will contribute their fair share only if effective enforcement mechanisms are in place.

P4 Members of the ideal society (only) do what they have a duty to do.

C Therefore, maintaining the ideal society will require effective enforcement mechanisms.

To defend premise 2*, we must find a way to explain why the lack of assurance is requirement-blocking. The most promising explanation starts from the idea of reciprocity. Many authors acknowledge the reciprocal nature of justice and its connection to coercive state institutions (Ci, 2006; Lister, 2013; Motchoulski, 2022; Sangiovanni, 2007). While some accounts stress reciprocity as a duty, our current concern is with reciprocity as a constraint on duties. The general idea is that at least some duties only arise if the duty-bearer could reasonably expect the rights-holder to reciprocate if she were in the duty-bearers position (Lister, 2013: 87). Andrew Lister argues that the rationale for a reciprocity condition is that duties are owed to people, and I cannot owe it to you to do something that you would never do for me. If I were to unconditionally accept a duty towards you, I would treat myself as a mere means to your ends. To refuse to unconditionally contribute to the well-being of non-reciprocators can thus be regarded as an affirmation of one's dignity and, in the extension, that of others (Lister, 2020: 216; White, 2003: 65–6).

Given that the reciprocity constraint applies to the duty to contribute one's fair share, we seem to have reason to accept premise 2*. To defend the need for assurance, however, all premises must be true. Premise 3 states that only when effective enforcement mechanisms are in place could members of the ideally just society reasonably expect others to contribute their fair share. This is hard to see given that those others share the same perfect moral motivations. Since they are surrounded by willing compliers, experience should tell the members of the ideal society that all others comply with moral demands virtually all the time (Brennan and Freiman, 2022: 11). Of course, unless they are mind-readers, they can have no guarantee that this trend will continue. But to suspect with reason that others will defect requires at least some indication that this will be the case. Since the members of the fully just society lack experience of previous noncompliance, the only indication I can think of is that their fellow members express their clear intention not to comply (though they will actually do so, since they – as we know – are willing compliers). But this would render the members of the ideal society liars, in which case their morality appears less than perfect.

One may object that this argument misses that there is an epistemic side to the problem of reasonable expectations. With Kavka, I have assumed that morally perfect agents hold all true moral beliefs, and always act conscientiously on those beliefs. We must not, however, forget the third premise, which states that morally perfect agents are like ordinary people in all other respects. Like all of us, members of the anarchic ideal society will suffer from cognitive limitations and imperfect information. This means that they will sometimes fail to comply with moral demands while believing that they so comply, and that they will sometimes believe that others fail to comply even when they actually do comply. Imperfections like these may create uncertainties as to whether others have reciprocated, and, hence, about their future willingness to reciprocate, even among morally perfect agents.

I will return to the problem of cognitive limitations in the subsequent section. For now, it suffices to note that the problem such limitations pose to the reciprocity condition cannot be solved through the establishment of a state. Since the state is ultimately run by individuals, whose cognitive capacities and access to information will be imperfect, too, its rulings will be subject to the same kind of doubt as individuals’ actions (Brennan and Freiman, 2022: 424–5; Freiman, 2017: 302). If members of the anarchic ideal society (correctly or not) suspect that others will fail to comply with their obligations, they will also (correctly or not) suspect that the ideal state will fail to accurately distribute, monitor and enforce such obligations. Recall that I here identify the state with coercive institutions. While coercion may be an effective response to weakness of will, it offers no solution to the problems of cognitive limitations or imperfect information. If the reciprocity condition, due to such limitations, is not fulfilled in the anarchic ideal society, nothing will be gained from shifting to the ideal state. ¹

Coordination and dispute settlement

Several authors have tried to ground the need for a state on its role for coordination and dispute settlement in large-scale societies. A first such attempt takes reasonable disagreement on the principles of justice as its departure point. Even if morally perfect agents will comply with the demands of justice, we may think that there is no guarantee that all members of society will have the same perception of justice. State institutions are thus needed to ensure conformity with the right principles of justice and resolve disputes that arise due to different understandings of what justice demands.

As is easily recognized, this worry rests on a misunderstanding of what strict compliance entails. As Jacob Levy and later Jeffery Carroll point out, the assumption of strict compliance entails an assumption of consensus on the principles of justice. If you and I hold different conceptions of justice, and each of us acts in accordance with our own conception, then at least one of us will fail to comply with the actual demands of justice. In this case, we no longer find ourselves in conditions of strict compliance (Carroll, 2021: 1135; Levy, 2016: 315). The same point can be made by reference to Kavka's definition of morally perfect agents. A morally perfect agent, we recall, is one that a) has a system of moral beliefs that contains all true (or justified) and no false (or falsified) moral beliefs, and b) always acts conscientiously on those beliefs. While the disagreeing agents both fulfill the second of these criteria, it cannot be true that both fulfill the first.

That morally perfect agents agree on the principles of justice may, however, not be enough to overcome moral disagreement among them. Principles of justice are stated in general terms and therefore underdetermine what justice demands in specific settings (O’Neill, 1996: 75; Valentini, 2009: 341). This is not simply an omission, but a response to the fact that which institutional arrangements best realize justice depends on a number of contextual factors (Carroll, 2021: 1135; Estlund, 2016: 295). The general character of moral principles may give rise to secondary moral disagreements of two different kinds. First, we may disagree about which out of various social schemes that all realize justice equally well to opt for. Due to the multitude of ways in which justice can be realized, moral perfection will not guide us in this case (Estlund, 2016: 301). Second, since the relevant factors may be more or less easy to access, and since human cognitive abilities are limited, we get what Carroll calls “epistemic pluralism”. Due to differences in experience, knowledge, reasoning and judgment, disagreement about which arrangements best realize justice may arise even among morally perfect agents who affirm the same principles of justice (Carroll, 2021: 1142; Estlund, 2016: 307).

The possibility of secondary moral disagreements reveals that morally perfect agents, too, will need mechanisms for dispute settlement. It is often assumed that such mechanisms can only be provided by the state. Locke famously characterizes the state of nature as a condition where men live together according to reason, without any authority to judge between them. Following his reasoning, many have taken the establishment of an authorized arbiter to indicate the presence of a state (but see Simmons, 1998: 453). In addition to dispute settlement, defenders of ideal statism have highlighted the coordinative role of the state in large-scale societies. Kavka argues that without the state's access to centralized information and high degree of specialization, any attempt to coordinate large populations will face insurmountable practical problems (Kavka, 1995: 14). Yet, to affirm the need for mechanisms of coordination and dispute settlement is not enough to conclude that we need a state. Recall that what is under scrutiny here are coercive state institutions. If the problems related to coordination and dispute settlement can be solved without coercion, the ideal society is still anarchic in my sense of the term.

To see clearly what is at stake here, it is helpful to distinguish between three functions state institutions could fulfill: selection, coordination, and enforcement. The selective function singles out one out of several possible social schemes that realize justice. The coordinative function accords each member of society a role in this social scheme and communicates this to everyone concerned. The enforcing function makes sure that everyone complies with the duties accorded to them, ultimately through coercive means. It seems beyond dispute that mechanisms of coordination and dispute settlement could fulfill the first two tasks without engaging in the third. We can, for instance, imagine a deliberative process, a voting system, an expert council or a lottery performing the selective function. And we can imagine various centralized and decentralized procedures that perform the coordinative task (Long, 2008: 136–7). As for the state's unique capacities to this end, Christopher Freiman argues, contra Kavka, that private actors could offer centralized information and specialization on terms equal to those of the state (Freiman, 2017: 16–7). But even if it turns out that coordination requires information and specialization of a kind that can only be arranged through state-like institutions, these would not count as states in my understanding as long as they are non-coercive.

While it is easy to imagine mechanisms of coordination and dispute settlement operating on a voluntary basis, we must ask whether we have reason to assume that morally perfect individuals would abide by their decisions. Though we may expect morally perfect agents – in virtue of being reasonable – to seek peaceful dispute settlement, this aim may run into conflict with their substantive moral commitments. As Kavka notes, only if all members of society give reasonability lexical priority over all other moral values can we rely on secondary moral disagreements to be settled on a voluntary basis. As long as some rank other values higher, the ideal society will be vulnerable to subversion by members who refuse to compromise (Kavka, 1995: 9–10). Kavka takes this to show that even morally perfect agents need a state.

Would morally perfect agents grant reasonability lexical priority? Put in other terms, would it be morally objectionable for members of the ideal society to refuse to compromise? By allowing for the possibility of morally perfect uncompromisers, ideal statists like Kavka and Carroll implicitly denies that it would (Carroll, 2021: 1143; Kavka, 1995: 11–2). I will, to the contrary, argue that compromise is a moral requirement for members of the ideal society.

My argument piggybacks on Thomas Christiano's defense of democracy. Christiano grounds the moral imperative of democracy in the prevalence of what he calls “the facts of judgment”: diversity, disagreement, cognitive bias and fallibility. In such conditions, he argues, disagreement will arise not because there are no true moral principles, but because our understanding of such principles and their application will diverge. Since no one can claim privileged access to the truth, what matters under these conditions is that each is able to see that he is treated as an equal (Christiano, 2008: 56). Christiano argues that democratic institutions meet this requirement, but other mechanisms, such as coin-flipping, may do equally well (Estlund, 2008: 66).

The conditions Christiano appeals to differ from those of the ideal society in that there will be no disagreement about the principles of justice in the latter. As we have seen, however, there may be secondary disagreement about the application of principles in specific empirical settings. When disagreement arises, awareness of our cognitive limitations should lead us to seek solutions that allow each to see that he is treated as an equal. Crucially, to fail to seek such solutions will constitute a moral failure since cognitive biases of different kinds tend to make us favor arrangements that benefit ourselves even if we have no intention to do so (Christiano, 2008: 60, see also Kavka, 1995: 6). To acknowledge the effects of our cognitive limitations is to acknowledge that a refusal to compromise accords more weight to our own interests than to those of others. This is morally objectionable since it fails to recognize their equal moral standing. The same cognitive limitations that give rise to disagreement thus render compromise a moral imperative. I conclude that the members of the ideal society should enter into voluntary schemes aimed at coordination and dispute settlement, and follow the decisions resulting from these.

I have argued that the members of the ideal society are under an obligation to compromise, since a failure to do so would – due to the facts of judgment – fail to recognize others’ equal moral standing. But what if they fail to recognize the facts of judgment? As the term indicates, diversity, disagreement, cognitive bias and fallibility are empirical facts. Unlike moral beliefs, then, we have no guarantee that morally perfect agents will be aware of them. We may therefore worry that the moral imperative to compromise applies not to the angelic members of the ideal society but only to a society of “archangels,” in possession of perfect cognitive abilities, flawless judgment, and aware of all existing facts (Kavka, 1995: 9). Paradoxically, archangels will not need to compromise, since their perfect knowledge in combination with their perfect motivation rules out moral disagreements of all sorts.

It is true that my argument assumes that morally perfect agents are aware of some empirical facts. Yet, the threshold is far below that of perfect information or perfect cognitive abilities. Rather, we should assume that morally perfect agents are aware of something like Rawls’ “general facts", a set of well-known facts about society and humankind that are beyond reasonable dispute (Rawls, 1999: 119, 137). By this, I do not wish to take a stand on the complicated issue of whether moral principles are fact-sensitive or not (Cohen, 2008; Jubb, 2009; Lippert-Rasmussen, 2019; Miller, 2013). Leaving this problem aside, I think most can agree that we need at least some factual knowledge to be able to act on our moral beliefs in a meaningful way. Little seems for example to be gained by acknowledging all human beings’ equal moral standing if the ignorant angels cannot distinguish a human being from a chicken or a watering can. Though the line between general and reasonably disputed facts remains to be drawn, it should be safe to assume that members of the ideal society are aware of basic facts such as the cognitive limitations characteristic of all human beings.

Further, we may worry that even if the members of the ideal society adopt a strategy of compromise, nothing ensures that they will coordinate on the same social scheme. Though coordination on any arrangement is almost always better than no coordination, multiple coordination can be as bad as no coordination at all (Hardin, 1999: 3, 13). I think the worry is unwarranted for precisely this reason. Assuming that the problem with multiple coordination is a loss in efficiency and therefore well-being, we should expect morally perfect agents to coordinate on a single scheme insofar as their moral commitments include a concern for well-being. Truly, on moral outlooks that attribute no value to well-being in absolute terms – such as strict egalitarianism – coordination on a single or multiple schemes will be a matter of taste. But this is an effect of the substantive moral theory at work, not of any shortcoming in the compromising strategy.

We have now seen that the defenses of ideal statism based on the need for assurance, reciprocity, coordination, and dispute settlement all fail to establish the need for coercive enforcement mechanisms among morally perfect agents. Even so, I believe that we shall resist the conclusion that the ideal society is an anarchic one. The coming sections outline two independent arguments for ideal statism, grounded in what I call the moral benefits of coercion. I will argue that the state's coercive institutions are morally beneficial since they a) lift demands on the will of complying agents, and b) counters the structural subordination that follows from differences in natural endowments. This, rather than any practical achievement, is what justifies the state in ideal conditions.

Lifting demands on the will

The disagreement between ideal anarchists and ideal statists concerns the justifiability of coercive state institutions in a society populated by morally perfect individuals. Previous defenses for ideal statism all tried to show that even in such conditions, a just social scheme cannot be achieved and maintained absent a state. Contrary to them, my own argument for ideal statism does not attempt to show that individual compliance, or the distributive outcomes resulting therefrom, cannot be secured without the coercive institutions of the state. I agree with ideal anarchists that voluntary arrangements will suffice for this purpose in a society populated by morally perfect individuals. Instead, I will argue that even if justice is in this sense realized, the anarchic ideal society is morally inferior to the ideal state. It is so because coercive enforcement mechanisms bring with them two benefits that the anarchic ideal society is lacking.

The first of these benefits is the alleviation of what Liam Murphy calls “demands on the will.” While morality by necessity places demands of different kinds on agents, demands on the will correspond to the cost of actively choosing to give up what one could otherwise (unjustly) have kept for oneself (Murphy, 2000: 49). By relieving us of demands on the will, state institutions reduce the burdens justice places on us. Thomas Nagel explains this idea as follows:

Most people will tolerate a universal system of compulsory taxation without feeling entitled to complain, whereas they would feel justified in refusing an appeal that they contribute the same amount voluntarily. This is partly due to lack of assurance that others would do likewise and fear of relative disadvantage; but it is also a sensible rejection of excessive demands on the will, which can be more irksome than automatic demands on the purse (Nagel, 1975: 145–6, my emphasis).

Nagel's conclusion that we are generally required to contribute more to a compulsory system of taxation than we would have to give up on a voluntary basis follows from the transformation of active demands into passive. What makes the difference is that we need not actively do anything in order to comply with the taxation system: the money is simply withdrawn from our gross salary and hence our duty to contribute places no demands on us to actively give away what has (unjustly) fallen into our lap. I think most share the intuition that actively giving away something places higher demands on the will than the passive withholding of the same. Yet, if this were the only way in which state institutions lifted demands on the will, the argument would provide no support for ideal statism. Clearly, the morally perfect individuals populating the ideal society could arrange for automatic withdrawals without backing them up by enforcement mechanisms. To the extent that they were interested in avoiding demands on the will, the members of society would voluntarily sign up for automatic transfers.

There is, however, another way in which state institutions lessen the demands on the will. This derives not from the automatization of transfers, but from the cost-benefit calculus that determines the weight of the demands on the will placed on us by a particular duty. Such demands arise because agents have something to gain from noncompliance with moral requirements. Even if I would be worse off in circumstances where no one complied, I would generally be better off as a single noncomplier among willing compliers. As long as I face no threat of punishment, the gains of noncompliance will often be high enough to generate considerable demands on the will. Yet, the introduction of enforcement mechanisms radically alters this cost-benefit calculus. If noncompliance comes at a risk of sanctions, agents typically have more to gain from compliance than from noncompliance (Murphy, 2000: 47). Since compliance under such conditions is not only morally required but also lies in the rational self-interest of individuals, enforcement mechanisms relieve agents of the demands on the will they face under ideal anarchism. If we think that such demands are to be avoided, then, the ideal state is morally superior to the anarchic ideal society.

Note that I am not suggesting that, unless relieved of the demands on the will, members of the ideal society would fall for the temptation to free-ride. To do so would display weakness of will, which violates the assumption of moral perfection and hence removes us from the domain of ideal theory as here understood. My point is only that enforcement mechanisms offer something morally valuable since they relieve complying agents of burdens that they face in their absence. Perhaps someone will object that demands on the will are not themselves morally objectionable. Indeed, some may even hold that such demands carry moral weight in the sense that an act gains its moral value from the efforts it takes to perform it. While this may be true for some moral acts, such as the sacrifices a loyal friend would make for you, it seems far less plausible when it comes to the duties members of a society have vis-à-vis one another. The contribution to common goods does not express a morally valuable bond between agent and patient, but simply sees to it that everyone gets what she has a rightful claim to. A society which can achieve this without placing demands on the will on its members seems clearly preferable to one that does place such demands.

Perhaps some will doubt that morally perfect agents face demands on the will in the first place. If we are genuinely motivated by moral principles, why would complying with those principles be burdensome for us? Indeed, if complying is what we want to do, noncompliance seems to place demands on our will! In order to respond to this challenge, we must make a convincing case that morally perfect agents can have immoral inclinations. Since demands on the will track the efforts of overcoming such inclinations, there would be no gain from state institutions if moral perfection ruled them out. The absence of immoral inclinations deviates from Kavka's definition of morally perfect agents as those that hold a flawless system of moral beliefs and always act conscientiously on those beliefs. If what is assessed according to moral standards are beliefs (or commitments) on the one hand, and acts on the other, the presence of certain inclinations will not taint the conscience of the members of the ideal society.

But what if Kavka's definition falls short of moral perfection? Talbot Brewer convincingly argues that some inclinations constitute a moral failure even if not acted upon. If members of the ideal society regularly had to overcome their temptation to kill or maim each other, we would arguably have reason to doubt that their moral motivations were perfect. This is because malicious inclinations of this kind signal that the agent does not fully appreciate the equal moral status of others (Brewer, 2002: 114). But even if malicious inclinations are incompatible with moral perfection, the temptation to free-ride that is under scrutiny here does not fall into this category. While the free-rider can rightfully be accused of reaping unfair benefits from others’ contributions and thereby showing them disrespect, the conclusion that she fails to recognize their equal moral status seems exaggerated. Contrary to the inclination to cause someone harm, the temptation to free-ride displays no malice, only a desire to pursue what lies in one's rational self-interest. To get what one wants or needs, Brewer argues, is generally a good thing and not itself morally objectionable. To illustrate this idea, he imagines a group of shipwreck survivors who mutually agree on a fair division of their small food supplies. If one of the survivors desires to take more than his fair share, yet does not act on his desire, this would display no moral shortcoming on his part. Rather than the objectionable desire to have the foods of others, his desire is the impeccable one of having more food (Brewer, 2002: 121). Similarly, my desire to free-ride does not generally entail a desire that others pay for my ride. If no one had to pay for it, all the better. As long as agents do not act on their temptation to free-ride, then, the mere inclination does not indicate that we have left the domain of ideal theory.

From this we may conclude that morally perfect agents can, and sometimes will, face demands on the will under ideal anarchism that they will not face under ideal statism. Whether this shows that the state is justified in ideal conditions depends on whether we think lifting the demands on the will is a concern important enough to justify coercive institutions. Given that coercion constitutes a pro tanto wrong, the moral costs attached to such institutions must be outweighed by the expected moral benefits (Brennan and Freiman, 2022: 10; Kushner, 2019: 461). In Carroll's wording, the proposed state institutions must meet the “better-than-anarchy-condition” (Carroll, 2021: 1134). I will return to this issue below. Let us just here note that even if we consider demands on the will a relatively minor problem, what needs justification is only the establishment of coercive institutions for incentivizing purposes. As we all know, no use of force will actually be needed among morally perfect agents. Rather than offering deterrence, the role played by ideal state institutions resembles that of my firm commitment not to leave my office before having graded the essays I promised to my colleague. Just like the morally perfect members of the ideal society have independent reasons for complying with justice, I have reasons for grading the essays that are independent of the prospect of getting home to my cozy bed. Yet, by raising the stakes, I ensure that my rational self-interest aligns with my obligations. ² Similarly, the coercive institutions of the ideal state serve a self-incentivizing purpose.

Countering structural subordination

I have argued that coercive state institutions are morally beneficial in the sense that they relieve members of the ideal society of demands on the will that fall on them under ideal anarchism. This section explores another moral benefit that follows from the establishment of a state. State institutions counter, so I will argue, the structural subordination that follows from differences in natural endowments among members of the ideal society.

My argument builds on Philip Pettit's defense of neorepublican freedom. Against the liberal conception of freedom as non-interference, Pettit advocates a conception of freedom as non-domination, or “antipower.” What matters for the sake of freedom, he argues, is not interference as such, but the capacity for arbitrary interference with others. It is this capacity that Pettit calls power. In unequal power relations, the dominant party constrains the freedom of the subjugated party even if no interference occurs. A benevolent slave-owner who does not interfere with his slaves’ doings constrains their freedom simply in virtue of his ownership status. Likewise, a loving husband in a traditional marriage or a fair-minded employer on a capitalist market constrain the freedom of their wife/employees due to the dependency of the latter on the former (Pettit, 1996: 581). Important to note is that it is the positioning of the powerful vis-à-vis the powerless that obstruct the freedom of the latter, not their acts or expected acts. In this sense, subordination is structural rather than agential. Even if no interference occurs, the mere fact that dominant parties can interfere on an arbitrary basis means that “the powerless are at the mercy of the powerful and not on equal terms” (Pettit, 1996: 583).

Unequal power relations are a crucial concern also for thinkers outside the neorepublican paradigm. Notably, relational egalitarians like Elizabeth Anderson argue that the proper aim of egalitarian justice is to “create a community in which people stand in relations of equality to others”, seeking “to abolish oppression – that is, forms of social relationship by which some people dominate … others” (Anderson, 1999: 289). And luck-egalitarian Gerald Cohen argues that some inequalities that are permitted by his preferred principle of justice must nonetheless be abolished due to the differences in power they give rise to (Cohen, 2009: 36). Though we may doubt that countering domination is the sole aim of egalitarian justice, few would probably deny that it is one of its aims, perhaps even the most important one.

According to Pettit, what is required to put an end to domination is the establishment of coercive institutions that ensure that arbitrary interference will not go unchecked. Provided that institutions are subject to constitutional constraints, they counter structural subordination without themselves exercising domination:

[I]f the fact of X's having power over Y is undermined by the presence of an ideal sort of constitutional authority which can punish X for any interference, X's power over Y disappears without any power over X being given to that authority (Pettit, 1996: 588).

If domination is unjust, and this injustice can be avoided only through the establishment of coercive institutions, we seem to have good reasons to favor ideal statism over ideal anarchism. Yet, in order to arrive at this conclusion, we must first show that domination occurs not only in nonideal circumstances, but also among the morally perfect members of the ideal society.

To determine this requires us to take a closer look at the causes of structural subordination. As sources of power, Pettit identifies a number of capacities and resources that are unevenly distributed in real-world societies. Some of these, such as differences in social standing, political authority and financial clout, are the effects of an unfair social order and will therefore not arise in the ideal society. Yet other, such as differences in physical strength, intelligence, boldness and charm are (at least partly) the result of natural endowments that will persist even in ideal conditions. Some of these differences will affect individuals’ capacity for free-riding or in other ways shirking their moral obligations. The strong person could scare someone to contribute in her place, whereas the charming person could talk someone into doing the same. The intelligent person could find a way to make it look like she contributed while she in fact did not, etc. Surely, morally perfect agents will not use their endowments to secure unfair advantages for themselves in this way. Still, their mere capacity to do so renders the less endowed dependent on their good will. This is so because in the anarchic ideal society, the only thing that stops the naturally endowed from shirking their moral obligations is their own moral commitments.

Does dependency on others’ good will lead to subordination even among morally perfect agents? Pettit touches upon this issue in his discussion of whether “benign or saintly” power bearers give rise to dominance to the same extent as malicious or indifferent ones. He argues that if power bearers recognize standards of conduct according to which they may be held accountable, this reduces the domination they represent compared to a scenario where they recognize no such standards. Since morally perfect agents recognize moral standards of conduct, this should lead us to expect domination in the ideal society to be lesser than in nonideal conditions. Contrary to the subjugated parties Pettit considers, the less endowed members of the ideal society will rest assured that society's shared moral standards can be used to “quote the powerful against themselves” (Pettit, 1996: 586). Note, however, that the recourse to accountability is superfluous among morally perfect agents. As I previously argued, these have no reason to expect that others will fail to comply with moral standards. Since their experience tell them that all others comply virtually all the time, they will need no assurance of the kind provided by the appeal to standards of conduct. If domination follows from the fear that others will misuse the power they possess, then, domination will not occur in the ideal society.

Apart from the subjective dimension of fear and its consequences, however, there is another, objective, dimension to domination. Domination, we recall, is structural rather than agential: it arises when persons occupy certain positions vis-à-vis each other. The observation that I am at the mercy of others is an objective fact and does not change because those others happen to be well-intentioned. Indeed, even if they were to promise not to make inappropriate use of their natural endowments, and I wholeheartedly trusted that they would honor this promise for the rest of our common life, I would still be at their mercy in the sense that whether or not I get what is rightfully mine depends on their good will. ³ To see why this is a problem, consider the situation of the wife of a loving husband who enjoys legal protection from violence and sexual abuse to that of a wife of an equally loving husband who does not. Due to their husbands’ loving character, neither of the two will ever be subjected to violence or abuse. Yet, the legal protection only the former enjoys puts them in very different positions vis-à-vis their husbands, and society at large. Only where recourse to some mechanism of enforcement – a mechanism for empowering the powerless – is available are the less endowed (in the relevant ways) no longer at the mercy of the more endowed.

So far, so good. But why, we may ask, must the enforcement mechanism come in the shape of coercive institutions? Would it not suffice with some – formal or informal – procedure for subjecting potential noncompliers to criticism or public shaming? While I here identify coercive institutions with the state, we can imagine institutions that put pressure on the noncomplier and still fall short of coercion. Since the members of the ideal society want to do what is right, making them aware of the fact that they have misbehaved would arguably suffice both to change their future behavior and pay reparations to victims. A procedure for public criticism constitutes a reliable tool which the less endowed could use to address potential abuses by the more endowed, and hence there is no need for coercion.

In response, let us distinguish between three mechanisms through which criticism or public shaming may produce the intended effect of changing noncompliers’ behavior. First, the criticism informs noncompliers that they have failed to comply. Clearly, awareness of such failure should lead morally perfect agents to change their behavior. Second, through their criticism, the compliers express discontent with the noncompliers. Insofar as seeking the approval of others equally motivated is part of morally perfect agents’ disposition, this, too, should lead noncompliers to comply. Note, however, that the success of both these mechanisms depend on the noncompliers’ moral commitments. Since the more endowed members’ susceptibility to criticism is an effect of their wish to comply with moral demands, the less endowed will – where these mechanisms are at work – still be dependent on their good will.

There is, however, a third mechanism through which criticism and public shaming could produce the desired effect. Compliers could withhold social benefits from those subjected to criticism and thereby attach costs to noncompliance. This could take various forms, but assume that compliers were to exclude noncompliers from the neighborhood party, or forbid their kids to play with the noncompliers’ kids. If this leads noncompliers to change their behavior, their decision would be grounded not in moral commitment but in their interest in accessing the social benefits. Where this mechanism is put in place, the less endowed are no longer dependent on the good will of the more endowed.

I will not dispute that ideal anarchism could counter structural subordination in this way. For the sake of argument, I will also grant that the withholding of social benefits falls short of coercion and hence that its institutionalization does not entail the establishment of a state. Instead, I argue that such institutions would deprive ideal anarchism of its relative advantage over ideal statism. Ideal anarchists argue that the anarchic ideal society is prima facie superior to ideal statism, since coercion constitutes a pro tanto wrong. If the anarchic ideal society can deliver justice without imposing the moral cost of coercion, we should prefer this to the ideal state. Coercion, however, is not the only thing that is morally costly. Social ostracism also constitutes a pro tanto wrong, as does mutual surveillance, another foreseeable effect of institutionalizing criticism and public shaming. Since the attitudes fostered in this environment risk to undermine community and social trust, we have good reason to believe that the wrongs they impose on society outweigh the wrongs of coercive state institutions. If ideal anarchism can only counter structural subordination through reliance on such mechanisms, the ideal state still fulfills the “better-than-anarchy-condition”.

Response to objections

I have argued that coercive institutions are morally beneficial in two ways: They lift the demands on the will that fall on agents under ideal anarchism, and they counter the structural subordination that follows from differences in natural endowments. If the establishment of coercive institutions were morally costless, this would suffice to conclude that the ideal state is preferable to ideal anarchism. Yet, coercion is not morally costless, and hence the cost of coercion must be outweighed by a greater benefit. The moral gains of lifting demands on the will and countering structural subordination must, in other words, exceed the moral costs of living under coercive institutions. Apart from costs directly related to coercion, i.e., violence or threats thereof, this category includes costs in the form of reduced well-being resulting from necessary expenditures. To establish and uphold state institutions come at considerable monetary costs for society, expenses that should be factored in as opportunity costs when comparing the ideal state to the anarchic ideal society (see e.g., Farrelly, 2007: 851). Given this, some may doubt that the benefits I have pointed to outweigh the costs of establishing and upholding state institutions. This leads to the objection that once the opportunity costs are taken into account, the ideal state does not fulfill the better-than-anarchy-condition.

I agree that the cost-benefit calculous must be carefully conducted. Note, however, that the opportunity costs imposed by state institutions may be smaller than they seem. First, the voluntary mechanisms for coordination and dispute settlement of the anarchic ideal society will, too, be costly to uphold. We should thus compare the costs of state institutions not to null, but to the costs of establishing and upholding non-coercive institutions. Second, under ideal conditions, state institutions could be significantly less extensive, and thus less costly, than today. Rather than a multitude of bodies working solely for this purpose, we should imagine enforcement as a dormant add-on to already existing mechanisms for coordination and dispute settlement. Truly, there will still be costs to account for, such as the cost of educating legal professionals (who will add this competence to that of their everyday work). Still, I think we have good reasons to believe that these relatively small costs will be outweighed by the moral benefits of the state.

Importantly, my defense of ideal statism also withstands another common objection raised by ideal anarchists. According to the so-called asymmetry objection, ideal statists arrive at their conclusion by comparing the ideal state to nonideal individual behavior. While individuals are held to be fallible, ignorant, immoral etc., state institutions are assumed to work flawlessly. Yet, if we assume that individuals will fail to comply with moral demands, we must acknowledge that states will do so, too, since states are, in the end, run by individuals (Brennan, 2014: 72; Freiman, 2017: 302; Long, 2008: 142). I will make no attempt to assess whether this argument defeats previous defenses of ideal statism. It suffices to note that it leaves my defense intact. Since I assume that individuals’ behavior will not deviate from moral standards, I am allowed to assume that state behavior, too, will be morally flawless. This shifts the burden of proof over to ideal anarchists. Given the minimal risks involved, it is now up to them to show that the moral costs imposed by the ideal state outweigh the moral benefits I have here pointed to.

Conclusion

This paper has developed a defense of ideal statism grounded in the moral benefits of coercion. Rather than a necessary tool for securing individual compliance or the distributive outcomes that follow therefrom, I have argued that the state is justified because it provides moral benefits to members of the ideal society. State institutions are morally beneficial because they a) lessen the demands on the will that fall on agents under ideal anarchism, and b) counter the structural subordination that follows from differences in natural endowments. The list may not end here. Rather than providing an exhaustive account of the moral benefits of the state, the approach outlined in this paper shifts the terms of the debate on ideal statism from feasibility to desirability. To see why the ideal society needs coercive institutions, we must treat the state not as a response to technical or motivational feasibility constraints, but as a framework for agency that carries moral weight of its own. By exploring the moral benefits of coercion, the argument outlined on these pages has exposed the flaws of ideal anarchism and provided new insights on the moral value of the state.